Taotlus




Study Protocol
P4-C2-017
DARWIN EU® - Time to onset of thromboembolic events in adults with selected types of cancer


10/10/2025
Version 1.0
Authors: Melissa Leung, Cesar Barboza, Ionna Nika, Anton Barchuk, Talita Duarte-Salles
Confidential


CONTENTS
LIST OF ABBREVIATIONS 5
1. TITLE 7
2. DESCRIPTION OF THE STUDY TEAM 7
3. ABSTRACT 8
4. AMENDMENTS AND UPDATES 11
5. MILESTONES 11
6. RATIONALE AND BACKGROUND 11
7. RESEARCH QUESTION AND OBJECTIVES 11
8. RESEARCH METHODS 11
8.1. Study design 11
Figure 1. Graphical depiction of the study design. 12
8.2. Follow-up 12
8.3. Study population with inclusion and exclusion criteria 12
8.4. Study setting and data sources 13
Table 1. Data sources. 13
8.5. Study period 14
8.6. Variables 14
8.6.1. Exposure 14
8.6.2. Outcome 14
8.6.3. Intercurrent events (only for causal studies) 14
8.6.4. Covariates, including confounders, effect modifiers, and other variables 14
8.7. Study size 15
8.8. Analysis 15
8.8.1. Federated network analyses 15
8.8.2. Data privacy protection 15
8.8.3. Statistical model specification and assumptions of the analytical approach considered 15
8.8.4. Output 16
Table 1. Attrition of study participants. 17
Figure 1. Cumulative probability of not having thromboembolic event after the first cancer diagnosis accounting for a competing risk of death. 19
Table 2. Median time in days (95% CI) to thromboembolic event after first cancer diagnosis. 20
9. STRENGTHS AND LIMITATIONS 20
10. REFERENCES 21
11. ANNEXES 22
ANNEX I. Description of data sources 22
ANNEX II. Fitness for use assessment 33
ANNEX III. Operational and reporting considerations 37
ANNEX IV. List of stand-alone documents 39
Table S1. List of concepts used to define deep vein thrombosis (DVT). 39
Table S2. List of concepts used to define pulmonary embolism (PE). 43
Table S3. List of concepts used to define venous thromboembolism (VTE). 43
Table S4. List of concepts used to define pelvic vein thrombosis (PVT) (concept sets included all descendants of listed concepts). 48
Table S5. List of concepts used to splanchnic vein thrombosis (SVT). 49
Table S6. List of concepts used to define retinal vein thrombosis (RVT). 49
Table S7. List of concepts used to define disseminated intravascular coagulation (DIC). 50
ANNEX V. ENCePP checklist for study protocols 52
ANNEX VI. Glossary 58

Study title
DARWIN EU® - Time to onset of thromboembolic events in adults with selected types of cancer
Protocol version
V1.0
Date
10/10/2025
EUPAS number
Study not registered yet
Active substance
None
Medicinal product
None
Research question and objectives
The aim of this study is to estimate time to onset of venous thromboembolic events in adults with each type of selected cancer.
The specific objectives of the study are:
1. To estimate the probability of not having thromboembolic events at 6-month intervals within 5 years in adults with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
2. To estimate median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
Countries of study
Belgium, Denmark, Estonia, Finland, Germany, The Netherlands, Spain, United Kingdom
Authors
Melissa Leung ([email protected])
Cesar Barboza ([email protected])
Ionna Nika ([email protected])
Anton Barchuk ([email protected])
Talita Duarte-Salles ([email protected])
This is a routinely repeated study from P3-C3-005 with EUPAS1000000440 (https://catalogues.ema.europa.eu/node/4341).
LIST OF ABBREVIATIONS
Acronyms/terms
Description
ADHD
Attention deficit hyperactivity disorder
AJCC/UICC
American Joint Committee on Cancer and the International Union Against Cancer
ATC
Anatomical Therapeutic Chemical
CDM 
Common Data Model 
CI
Confidence interval
CPRD 
Clinical Practice Research Datalink 
DARWIN EU® 
Data Analysis and Real World Interrogation Network 
DK-DHR
Danish Data Health Registries
DOI 
Declaration Of Interests 
DQD 
Data Quality Dashboard 
DRE
Digital Research Environment
DVT
Deep Venous Thrombosis
DIC
Disseminated Intravascular Coagulation
EHR 
Electronic Health Record 
EMA 
European Medicines Agency 
EBB 
Estonian Biobank 
EGCUT 
Estonian Genome Center at the University of Tartu 
ENCePP 
European Network of Centres for Pharmacoepidemiology and Pharmacovigilance 
EU
European Union
EUPAS
EU Post-Authorisation Studies Register
GDPR
General Data Protection Regulation
GP 
General Practitioner 
HIV
Human Immunodeficiency Virus
ICD-O-3
International Classification of Diseases for Oncology, 3rd Edition
ICD-10
International Classification of Diseases, 10th revision
ICPC-1
International Classification of Primary Care
IP
Inpatient
IPCI 
Integrated Primary Care Information Project 
IR
Incidence rate
IRB
Institutional Review Board
LPD 
Longitudinal Patient Database 
OHDSI
Observational Health Data Sciences and Informatics
OMOP 
Observational Medical Outcomes Partnership 
OP
Outpatient
PE
Pulmonary Embolism
PVT
Pelvic Venous Thrombosis
PY
Person-years
RVT
Retinal vein thrombosis
SNOMED 
Systematized Nomenclature of Medicine 
SVT
Splanchnic Vein Thrombosis
UKBB
UK Biobank
VTE
Venous Thromboembolism

1. TITLE
DARWIN EU® - Time to onset of thromboembolic events in adults with selected types of cancer
2. DESCRIPTION OF THE STUDY TEAM
Study team role
Names
Organisation
Principal Investigator
Melissa Leung
Anton Barchuk
Talita Duarte-Salles
Erasmus MC
Data Scientist
Cesar Barboza
Ioanna Nika
Erasmus MC
Clinical Domain Expert
Anton Barchuk
Erasmus MC
Study Manager
Natasha Yefimenko
Erasmus MC
Data source
Names
Data Partner Organisation*
IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium)
IQVIA Disease Analyzer Germany (IQVIA DA Germany)
Gargi Jadhav   
Isabella Kacmarczyl  
Akram Mendez   
Hanne van Ballegooijen  
Dina Vojinovic  
IQVIA
Danish Data Health Registries (DK-DHR)
Elvira Bräuner 
Susanne Bruun 
Danish Medicines Agency
Estonian Biobank (EBB)
Marek Oja
Raivo Kolde
Ami Sild
Estonian Biobank, Estonia
Finnish Care Register for Health Care (FinOMOP-THL)
Anna Hammais
Gustav Klingstedt
Finnish Care Register for Health Care, Finland
Integrated Primary Care Information (IPCI)
Katia Verhamme
Integrated Primary Care Information, Netherlands
The Information System for the Development of Research in Primary Care (SIDIAP)
Anna Palomar-Cros
Irene López-Sánchez
Agustina Giuliodori
IDIAPJGol
Clinical Practice Research Datalink GOLD (CPRD GOLD) and UK BioBank (UKBB)
Antonella Delmestri
University of Oxford 
*Data partners do not have an investigator role. Data partners execute code at their data source, review, and approve their results.
3. ABSTRACT
Title
DARWIN EU® - Time to onset of thromboembolic events in adults with selected types of cancer
Rationale and background
Thromboembolic events are a common complication for individuals with cancer, with risk varying according to the cancer site, suggesting cancer-specific mechanisms playing a role in the occurrence of these events. Haematological malignancies and lung, pancreas, stomach, bowel, and brain cancers are generally associated with a high risk of clot formation, whilst prostate and breast cancers are associated with low risk of thrombosis.
When a safety signal of a thromboembolic event appears in cancer populations, it can be challenging to assess a potential association with the oncologic treatment without reliable information on the background risk. This study is intended to address this knowledge gap by generating evidence on the time to onset of different venous thromboembolic events among adults with selected cancer types.
Research question and objectives
Research question
What was the time to onset of venous thromboembolic events in adults newly diagnosed with each type of selected cancer during the period 2016–2022?
Objectives
The aim of this study is to estimate time to onset of venous thromboembolic events in adults with each type of selected cancer.
The specific objectives of the study are:
1. To estimate the probability of not having thromboembolic events at 6-month intervals within 5 years in adults with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
2. To estimate median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
Methods
Study design
Population-based cohort study. The index date, i.e., date of cohort entry, will be the date of the first cancer diagnosis. Individuals are followed up until the earliest of occurrence of the outcome, loss to follow-up, end of data availability, end of the study period, or death.
Population
The study population will be the population that was included in the study EUPAS1000000440, of which this is a routinely repeated study. This study population will include all individuals aged 18 years and above with a primary diagnosis of one of the selected cancers (bone, brain, breast, colorectal, corpus uteri, kidney, leukaemia and lymphoma, liver, lung, melanoma, oesophageal, ovary, pancreas, prostate, stomach) during the inclusion period (from 01/01/2016 to 31/12/2022). Only individuals with an incident cancer diagnosis (excluding non-melanoma skin cancer), defined as a first cancer diagnosis after ≥365 days cancer-free history, will be included. Cancer cases and thromboembolic events will be identified based on appropriate computable phenotyping algorithms. Conditions in the OMOP CDM use the Systematised Nomenclature of Medicine (SNOMED) as the standard vocabulary for diagnosis codes. The International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) will also be considered for cancer diagnoses. Other eligibility criteria will include at least 365 days of database history prior to index date and at least 365 days between index date and end of data availability in the data source.
Variables
Exposure:
Not applicable.
Outcome:
The outcomes will include thromboembolic events, specifically: deep vein thrombosis (DVT), pulmonary embolism (PE), venous thromboembolism (VTE, composite of DVT and PE), pelvic venous thrombosis (PVT), splanchnic vein thrombosis (SVT, including hepatic and extra-hepatic vein thrombosis), retinal vein thrombosis (RVT, including retinal central vein thrombosis), and disseminated intravascular coagulation (DIC).
Relevant covariates:
The following covariates will be assessed at index date: age group in years (18–34, 35–44, 45–54, 55–64, 65–74, 75–84, and ≥85), sex, and study subperiod (2016–2019 and 2020–2022). These variables will be used to stratify the results.
Data sources
1. Belgium: IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium)
2. Denmark: Danish Data Health Registries (DK-DHR)
3. Estonia: Estonian Biobank (EBB)
4. Finland: Finnish Care Register for Health Care (FinOMOP-THL)
5. Germany: IQVIA Disease Analyzer Germany (IQVIA DA Germany)
6. Netherlands: Integrated Primary Care Information (IPCI)
7. Spain: The Information System for Research on Primary Care (SIDIAP)
8. United Kingdom: Clinical Practice Research Datalink GOLD (CPRD GOLD)
9. United Kingdom: UK BioBank (UKBB)
Study size
No sample size will be calculated, as this is an exploratory study which will not test a specific hypothesis. Based on the results of the study EUPAS1000000440, the expected number of person counts will be the lowest for DIC (during 1-year follow-up: 5 in FinOMOP-THL – 171 in SIDIAP, with 0 counts in CPRD GOLD, EBB, IPCI, IQVIA DA Germany, and IQVIA LPD Belgium) and highest for VTE (during 1-year follow-up: 27 in IQVIA LPD Belgium – 4,597 in FinOMOP-THL).
Statistical analysis
Analyses will be conducted separately for each data source and carried out in a federated manner, allowing analyses to be run locally without sharing individual-level data.
Objective 1
The probabilities of not having thromboembolic events at 6-month intervals within 5 years in adults with each type of selected cancer will be assessed using the R package CohortSurvival, accounting for a competing risk of death.
Objective 2
The median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer will be assessed using the R package CohortSurvival.
The R package CohortSurvival is designed to work with data in the OMOP CDM format to extract and summarise survival data applying the Kaplan-Meier method. The analyses will be conducted for the overall cohorts as well as by strata of age group, sex, and study subperiod.
Absence of diagnosis codes will be interpreted as a lack of the conditions themselves. A minimum cell count of 5 will be used when reporting results, with any smaller count reported as “<5” and zero counts as “0”.

4. AMENDMENTS AND UPDATES
None
5. MILESTONES
Study milestones and deliverables
Planned dates
Final Study Protocol
To be confirmed by EMA
Creation of Analytical code
31 October 2025
Execution of Analytical Code on the data
14 November 2025
Draft Study Report
Depending on IRB approvals
Final Study Report
Depending on IRB approvals
*Planned dates are dependent on obtaining approvals from the internal review boards of the data sources.
6. RATIONALE AND BACKGROUND
This study is a routine repeated study of a previous DARWIN EU® study (EUPAS1000000440) focused on estimating the incidence rates of venous thromboembolic events in adults newly diagnosed with any of the selected cancers (bone, brain, breast, colorectal, corpus uteri, kidney, leukaemia and lymphoma, liver, lung, melanoma, oesophageal, ovary, pancreas, prostate, stomach) during the period 2016–2022 and describing the individuals’ characteristics at the time of cancer diagnosis. This study is now being repeated with the same study population and outcome, but a different objective, i.e., to estimate the time to onset of thromboembolic events in adults with selected types of cancer.
7. RESEARCH QUESTION AND OBJECTIVES
Research question
The aim of this study is to estimate time to onset of venous thromboembolic events in adults with each type of selected cancer.
Research objectives
The specific objectives of the study are:
1. To estimate the probability of not having thromboembolic events at 6-month intervals within 5 years in adults with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
2. To estimate median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
8. RESEARCH METHODS
8.1. Study design
A cohort study will be conducted. The study will comprise:
• a characterisation study to address objective 1, assessing the probability of not having thromboembolic events at 6-month intervals within 5 years in adults with each type of selected cancer, overall and stratified by age group, sex, and study subperiod
• a characterisation study to address objective 2, assessing median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
The study design to address objective 1, including assessment windows, is visualised in Figure 1. For objective 2, we will subset the cohort of adults with cancer with thromboembolic events.

Figure 1. Graphical depiction of the study design.
a. The censor date will be the earliest of occurrence of the outcome, loss to follow-up, end of data availability, or death.
8.2. Follow-up
For both objectives, follow-up in the survival analysis will start on the date of cancer diagnosis (index date) and end on the earliest of occurrence of the outcome (thromboembolic event), loss to follow-up, end of data availability, or death.
8.3. Study population with inclusion and exclusion criteria
Objective 1
Inclusion criteria
• First diagnosis of a selected cancer (index date) between 01/01/2016 and 31/12/2022
• Age ≥18 years at cancer diagnosis
• Minimum 365 days of available history before the cancer diagnosis date
• Cancer diagnosis date ≥365 days prior to end of data availability of the data source
Exclusion criteria
• History of any cancer diagnosis ever before the selected cancer diagnosis date
• Outcome during the year before the cancer diagnosis date
Objective 2
Inclusion criteria
• Included in the study population of objective 1
• Occurrence of the outcome (thromboembolic event) during follow-up.


8.4. Study setting and data sources
This study will be conducted using routinely collected data from different health care settings from 9 data sources in the DARWIN EU® network of data partners from 8 countries across Europe, of which 7 EU member states (Table 1). All data were a priori mapped to the OMOP CDM.
Table 1. Data sources.
Country
Name of Data source
Health Care setting
Type of Data
Number of active individuals
Calendar period covered by each data source
Contributing to
BE
IQVIA LPD Belgium
Primary care
EHRs
189k
2015–2025
All objectives
DK
DK-DHR
All settings
EHRs, registries, claims
5.98M
1995–2025
All objectives
EE
EBB
Primary care, hospital care (IP and OP)
EHRs, claims, registries, biobank
212k
2004–2025
All objectives
FI
FinOMOP-THL
Hospital care (IP and OP)
EHRs, registries
5.7M
2011–2025
All objectives
DE
IQVIA DA Germany
Primary care
EHRs
4.48M
1992–2025
All objectives
NL
IPCI
Primary care
EHRs
1.33M
2006–2025
All objectives
ES
SIDIAP
Primary care
EHRs
5.95M
2006–2025
All objectives
GB
CPRD GOLD
Primary care, hospital care (OP)
EHRs
2.83M
1987–2025
All objectives
GB
UKBB
Primary care (up to 2017), hospital care (IP and OP, up to November 2022)
EHRs, registries, biobank
500k
1940–2025
All objectives
Countries: BE=Belgium, DE=Germany, DK=Denmark, EE=Estonia, ES=Spain, FI=Finland, GB=United Kingdom of Great Britain and Northern Ireland, NL=The Netherlands
Data sources: IQVIA LPD=IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium); DK-DHR=Danish Data Health Registries; EBB=Estonian Biobank; FinOMOP-THL=Finnish Care Register for Health Care; IQVIA DA=IQVIA Disease Analyzer Germany (IQVIA DA Germany); IPCI=Integrated Primary Care Information; SIDIAP=The Information System for Research in Primary Care; CPRD GOLD=Clinical Practice Research Datalink GOLD; UKBB=UK BioBank
Types of data: EHR=electronic health record, IP=inpatient, OP=outpatient
Number of active subjects: k=thousands, M=millions

Data sources selection
These data sources fulfil the criteria required in terms of data quality, completeness, timeliness, and representativeness for the cohort study while covering different regions of Europe (Annex II).



8.5. Study period
The study period is from 01/01/2016 to the most recent data available for each contributing data source.
8.6. Variables
8.6.1. Exposure
None.
8.6.2. Outcome
All objectives
The thromboembolic event outcomes in this study are identical to those in EUPAS1000000440, of which this is a routinely repeated study:
• Deep vein thrombosis (DVT)
• Pulmonary embolism (PE)
• Venous thromboembolism (VTE, composite of DVT and PE)
• Pelvic venous thrombosis (PVT)
• Splanchnic vein thrombosis (SVT), including hepatic and extra-hepatic vein thrombosis
• Retinal vein thrombosis (RVT), including retinal central vein thrombosis
• Disseminated intravascular coagulation (DIC)
Each of the specific thromboembolic events will be a primary, binary outcome. The outcome will be assessed at any diagnosis position in the electronic health record and from any of the care settings in each data source.
The list of concepts, from EUPAS1000000440 and based on SNOMED codes and aligned with previous studies that used OMOP CDM and VTE as an outcome (Burn et al., 2022), is provided in Annex IV.
8.6.3. Intercurrent events (only for causal studies)
Not applicable.
8.6.4. Covariates, including confounders, effect modifiers, and other variables
All Objectives
• Sex
◦ Female/male
• Age groups (years) at cancer diagnosis date:
◦ 18–34
◦ 35–44
◦ 45–54
◦ 55–64
◦ 65–74
◦ 75–84
◦ ≥85
Age at cancer diagnosis date will be calculated using January 1st of the year of birth as proxy for the actual birthday. Date/month is either not present or cannot be made available for governance reasons. If available, date is often set to first of the month for personal privacy.
• Study subperiod:
◦ 2016–2019
◦ 2020–2022
The status of each covariate will be assessed at the cancer diagnosis date. Each covariate will be used to stratify the results.
8.7. Study size
No sample size will be calculated, as this is a characterisation study which will not test a specific hypothesis. In addition, we will use data from the study population that was included in EUPAS1000000440, of which this is a routinely repeated study. Thus, the sample size is driven by the availability of data for adults with both cancer and a thromboembolic event. Based on the results of the study EUPAS1000000440, the expected number of person counts is lowest for DIC (during 1-year follow-up: 5 in FinOMOP-THL – 171 in SIDIAP, with 0 counts in CPRD GOLD, EBB, IPCI, IQVIA DA Germany, and IQVIA LPD Belgium) and highest for VTE (during 1-year follow-up: 27 in IQVIA LPD Belgium – 4,597 in FinOMOP-THL).
8.8. Analysis
8.8.1. Federated network analyses
All analyses will be conducted separately for each data source, and will be carried out in a federated manner, allowing analyses to be run locally without sharing individuals’ data.
Before sharing the study package, test runs of the analytics will be performed on a subset of the data sources and quality control checks will be performed. After all the tests are passed (see Annex III. Operational and reporting considerations), the final package will be released in a version-controlled study repository for execution against all the participating data sources.
8.8.2. Data privacy protection
The data partners will locally execute the analytics against the OMOP CDM in R Studio and review and approve the default aggregated results. They will then be made available to the Principal Investigators and study team in secure online repository of DTZ (Data Transfer Zone). All results will be locked and timestamped for reproducibility and transparency. The study results of all data sources will be checked, after which they are made available to the team, and the Study Dissemination Phase can start. All analyses will be conducted separately for each database, and will be carried out in a federated manner, allowing analyses to be run locally without sharing individual-level data. Cell counts <5 will be suppressed when reporting results to comply with the data source’s privacy protection regulations.
8.8.3. Statistical model specification and assumptions of the analytical approach considered
Objective 1
The probability of not having thromboembolic events at 6-month intervals within 5 years (6, 12, 18, 24, 30, 36, 42, 48, 54, and 60 months) in adults with each type of selected cancer will be calculated based on OMOP CDM mapped data using the R package CohortSurvival, developed by DARWIN EU®. The key measure of not having events at 6-month intervals within 5 years in adults with each type of selected cancer will be the probability, i.e., the proportion of individuals in the cohort who have not yet experienced a thromboembolic event after each interval, accounting for the competing risk of death. Individuals will be censored at the date of loss to follow-up, the end of data availability, or the end of the study period. The estimateCompetingRiskSurvival() function will be used in the analysis.
Objective 2
The median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer will be calculated based on OMOP CDM mapped data using the R package CohortSurvival, developed by DARWIN EU®. The key measure of median time to onset of venous thromboembolic events will be median survival time obtained using the estimateSingleEventSurvival() function.
All objectives
The R package CohortSurvival is designed to work with data in the OMOP CDM format to extract and summarise survival data applying the Kaplan-Meier method. For objective 1, death will be accounted for as a competing risk. The analyses will be conducted for the overall cohorts as well as by strata of age group, sex, and study subperiod.
The absence of diagnosis codes will be interpreted as the absence of the conditions themselves. A minimum cell count of 5 will be used when reporting results, with any smaller count reported as “<5” and zero counts as “0”.
Sensitivity analysis
Not applicable.
8.8.4. Output
Output will include a PDF report including an executive summary, and tables and figures. Mock versions of the intended tables and figures are listed below.
• Table 1. Attrition of study participants (objectives 1 and 2).
• Figure 1. Probability of not having thromboembolic event after the first cancer diagnosis (objective 1).
This figure will be plotted for each combination of cancer type and outcome, faceted by data source. The main report will include 15 such figures: one for each type of cancer. In the meta-analysis results of EUPAS1000000440, of which this is a routinely repeated study, the most common outcome was VTE across all cancer types, except for liver cancer. In liver cancer, the most common outcome was SVT. Therefore, Figure 1 in the main report will be plotted for SVT in liver cancer and for VTE in all other cancer types. The corresponding figures for all other outcomes across all cancer types will be available in Shiny.
• Table 2. Median time to thromboembolic event after first cancer diagnosis in those with occurrence of thromboembolic event (objective 2).
The main report will include 15 such tables: one for each cancer type.
An interactive dashboard (Shiny) will be generated by incorporating all the results (tables and figures) included in the PDF report mentioned above. Specifically, the Shiny will contain:
• Overall results:
◦ Objective 1: 15 cancer types * 7 outcomes = 105 figures as presented in Figure 1.
◦ Objective 2: 15 tables as presented in Table 2.
• Stratified results:
◦ By age group:
▪ Objective 1: 7 age groups * 105 figures = 735 figures
▪ Objective 2: 7 age groups * 15 tables = 105 tables
◦ By sex:
▪ Objective 1: 2 sexes * 105 figures = 210 figures
▪ Objective 2: 2 sexes * 15 tables = 30 tables
◦ By study subperiod:
▪ Objective 1: 2 subperiods * 105 figures = 210 figures
▪ Objective 2: 2 subperiods * 15 tables = 30 tables.

Table 1. Attrition of study participants.


IQVIA LPD
DK-DHR
EBB
FinOMOP-THL
IQVIA DA
IPCI
SIDIAP
CPRD GOLD
UKBB
Bone cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Brain cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Breast cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Colorectal cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Corpus uteri cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Oesophageal cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Kidney cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Liver cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Lung cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Lymphoma and Leukaemia
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Melanoma
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Ovarian cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Pancreatic cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Prostate cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Stomach cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









N=the number of individuals meeting each criterion.
IQVIA LPD=IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium); DK-DHR=Danish Data Health Registries; EBB=Estonian Biobank; FinOMOP-THL=Finnish Care Register for Health Care; IQVIA DA=IQVIA Disease Analyzer Germany (IQVIA DA Germany); IPCI=Integrated Primary Care Information; SIDIAP=The Information System for Research in Primary Care; CPRD GOLD=Clinical Practice Research Datalink GOLD; UKBB=UK BioBank.


Figure 1. Cumulative probability of not having thromboembolic event after the first cancer diagnosis accounting for a competing risk of death.

Table 2. Median time in days (95% CI) to thromboembolic event after first cancer diagnosis.

IQVIA LPD
DK-DHR
EBB
IQVIA DA
IPCI
SIDIAP
CPRD GOLD
UKBB
FinOMOP-THL
Deep vein thrombosis (DVT)









PE and DVT combined (VTE)









Pelvic vein thrombosis (PVT)









Pulmonary embolism (PE)









Retinal vein thrombosis (RVT)









Splanchnic extrahepatic vein thrombosis (SVT)









IQVIA LPD=IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium); DK-DHR=Danish Data Health Registries; EBB=Estonian Biobank; FinOMOP-THL=Finnish Care Register for Health Care; IQVIA DA=IQVIA Disease Analyzer Germany (IQVIA DA Germany); IPCI=Integrated Primary Care Information; SIDIAP=The Information System for Research in Primary Care; CPRD GOLD=Clinical Practice Research Datalink GOLD; UKBB=UK BioBank; NC=Not calculated (fewer than 5 events).

9. STRENGTHS AND LIMITATIONS
The study will be informed by routinely collected healthcare data, and therefore, data quality issues must be considered. In particular, the identification of individuals with cancer and thromboembolic events may vary across data sources. While relatively few false positives are expected, false negatives may be more likely, especially for primary care data sources that lack patient-level linkage to secondary care data. We expect misclassification to be minimal in registry data sources and in the primary care data sources where cancer diagnoses have been previously validated. Underestimation of thromboembolic events is also possible, particularly for rare events with complex diagnoses, such as RVT and SVT.
Given the large number and diverse nature of participating data sources, it is important to note that differences in patient representations might have resulted from disparate coding practices and specifics of data capture. The granularity or detail of concepts representing clinical facts can vary across source terminologies (e.g., ICD-10, Read codes), influencing how information is later transformed into standardised vocabularies [1]. The preliminary code lists created to identify individuals with cancer include codes from standard vocabularies used in cancer registries, such as ICD-O-3 codes. However, most data sources capture information on cancer diagnoses using SNOMED codes, which may not be granular enough to cover all the topology and morphology of cancer [2]. ICD-O-3 codes are only available at DK-DHR, EBB, and UKBB.

10. REFERENCES
1. Ostropolets A, Reich C, Ryan P, Weng C, Molinaro A, DeFalco F, et al. Characterizing database granularity using SNOMED-CT hierarchy. AMIA Annu Symp Proc. 2021 Jan 25;2020:983–92.
2. Campbell WS, Campbell JR, West WW, McClay JC, Hinrichs SH. Semantic analysis of SNOMED CT for a post-coordinated database of histopathology findings. J Am Med Inform Assoc. 2014 Sept 1;21(5):885–92.

11. ANNEXES
ANNEX I. Description of data sources
DATA SOURCES DESCRIPTION
IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium)
#
Section
Description
1
Database Identification and country
IQVIA LPD Belgium (IQVIA Longitudinal Patient Database Belgium) Belgium
2
Data partner information section
IQVIA IQVIA Europe
3
Coverage and timespan
Data collection since: 2005 Extent: Nation-wide. Panel of 300 GPs in Belgium. The panel is maintained as a representative sample of the primary care physician population in Belgium, according to three criteria known to influence prescribing: age, sex, and geographical distribution.
4
Healthcare setting / type of data
Primary care – gps. Ambulatory visits, with diagnosis, prescriptions, procedures, and laboratory tests.
5
Data collection process
Outpatient electronic health records. Records are entered by GPs at the healthcare encounter.
6
General representativeness
The panel of contributing physicians (a stable 300 GPs) is maintained as a representative sample of the primary care physician population in Belgium, according to three criteria known to influence prescribing: age, sex, and geographical distribution. The panel consists of a stable 300 GPs that are geographically well spread. The total number of active GPs in Belgium is 15,602. The regional geographical spread of physicians in the LPD data is also representative of the distribution across the country: 57% GPs in the North (compared to 54% nationally), 31% in the South (33% nationally), and 12% in Brussels (13%).The provider of the data has more than 2,250 GPs under contract so in case of a drop out a replacement is easily found.
7
Data content /source coding
No information on source coding.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. The patient ID is per practice. So a patient can have different IDs in the DB, one per practice. In Belgium, patients are typically registered at only one GP practice, so duplication should be minimal.
9
Quality control (database specific)
No QC. Integrity constraints only.
10
Linkage
No linkage.
11
Vital status
Death information is derived from healthcare events.
12
Limitations
No database-specific limitations documented. General limitations for the data type applicable.
13
Main references
No main reference provided.
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111116 Website: https://iqvia.com



Danish Data Health Registries (DK-DHR)
#
Section
Description
1
Database Identification and country
DK-DHR (Danish Data Health Registries) Denmark
2
Data partner information section
Danish Medicines Agency (DKMA), Data Analytics Centre (DAC)
3
Coverage and timespan
Data collection since: 1995, Extent: Nationwide. The data is representative of the entire Danish population.
4
Healthcare setting / type of data
Community pharmacists, and secondary care – specialists (ambulatory or hospital outpatient care), and hospital inpatient care. The following data elements are collected: diagnosis (including rare diseases and pregnancy data), hospital admissions, discharge and ICU data, Cause of death, Drug prescription retrievals, vaccination and contraception, Procedures, and Sociodemographic information (sex and age but no information on income, education, occupation).
5
Data collection process
Outpatient electronic health records, and Inpatient hospital electronic health records, and Registries, and Other. All causes of deaths, all retrieved drug prescriptions, all records of vaccinations, all hospital inpatient and outpatients contacts including disease diagnoses and hospital surgical and non-surgical procedures, histologically confirmed incident cancers, laboratory test results for the entire Danish population from 1/1/1995 onwards.
6
General representativeness
The data is representative of the entire Danish population. Healthcare is free in Denmark, so we do not expect any bias in data collection based on socio-economic status.
7
Data content /source coding
Diagnoses and causes of death are collected using the ICD-10 vocabulary. ATC and RxNorm are used for Drugs. SNOMED codes are used for Procedures.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC, ICDO3, cancer Modifier). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. .
9
Quality control (database specific)
The data we have received relating to nationwide Danish Health Data registries offer an opportunity for large-scale, population-based studies with several advantages 1) Their large size improves the precision of estimates and enables the study of rare exposures and outcomes with long-term latency, 2) Inclusion of nearly all individuals in the target population ensures that the data reflect routine clinical care and all clinical segments of the source population, 3) Data are collected independently of each research study, thus minimising certain types of bias, e.g., non-response, and the influence from attention to the research question on the diagnostic process. Before the source data is sent to us, the Danish Health Data Authority does running and comprehensive checks of the registry table data validity of the variables, breaks in data, changes in variable coding, missingness, etc. We perform checks of missingness/completeness in relation to requested variables. In essence, we are receiving a dump of a mirror of the data that is controlled by the SDS. The documentation performed by SDS is available online, in Danish primarily https://www.esundhed.dk/Dokumentation (all variables), but also in English https://sundhedsdatastyrelsen.dk/da/english/health_data_and_registers/national_health_registers
10
Linkage
There is no linkage in this data source.
11
Vital status
The Cause of Death registry (DAR) is used, the cause of death is collected using ICD-10 codes.
12
Limitations
There are no clinical measurements in the data. DK-DHR has the following limitations, which may be relevant confounders for certain complex Darwin EU studies:
• We lack information on key socio-economic status (SES) factors, such as occupation, education, and income. These variables may be important for analysis in some studies.
• We only have complete data on lifestyle factors (such as smoking status and weight) for pregnant women.
• We have no information on patient contacts in primary care (visits to the GP). Consequently, the incidence of chronic diseases like Type 2 Diab
13
Main references
Schmidt M, Schmidt SAJ,Adelborg K,Sundbøll J,Laugesen K,Ehrenstein V,Sørensen HT "The Danish health care system and epidemiological research: from health care contacts to database records." Clinical epidemiology (2019): 31372058
14
Link to HMA-EMA catalogue and database webpage
Website: https://sundhedsdatastyrelsen.dk/da/english/health_data_and_registers/healthdatadenmark HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111217

Estonian Biobank (EBB)
#
Section
Description
1
Database Identification and country
EBB (Estonian Biobank) Estonia
2
Data partner information section
University of Tartu Institute of Computer Science
3
Coverage and timespan
Data collection since: 2004 Extent: Nation-wide. EBB is a nation-wide database containing records from 2004 onwards. Estonian population-based cohort size of 211,800 participants (01/01/2024) aged 18 years and older recruited at GP offices, private practices, and hospitals or in the recruitment offices of the Estonian Genome Center.
4
Healthcare setting / type of data
Primary care – GPs, and community pharmacists, and primary care specialists (e.g. paediatricians), and secondary care – specialists (ambulatory or hospital outpatient care), and hospital inpatient care. Registry which collects electronic records from the biobank and cohort study.
5
Data collection process
Data is retrieved by Estonian Biobank once a year from national registries. The insurance claims are requested from Estonian Health Insurance Fund. The inpatient and outpatient electronic health records are requested from National Health Information System. The cancer registry and cause of death registry information is requested from The National Institute for Health Development. The data is sent to the national registry by the healthcare providers.

6
General representativeness
The age, sex, and geographical distribution closely reflect those of the Estonian adult population and encompass close to 5approximately 20% of adult population. Female participants are over-represented in EBB. Overall, 3.4% of Estonian men and 5.5% of Estonian women are represented in EBB. Older people tend to participate less frequently, however, all age groups are well represented.
7
Data content /source coding
All participants have undergone a standardized health assessment, including provision of blood samples for purification of DNA, white blood cells, and plasma, and completed a questionnaire covering various health-related topics, such as lifestyle, diet, and clinical diagnoses. Diseases and health problems are recorded as ICD-10 codes and prescribed medicine according to the ATC classification and local package codes. Procedures and services are coded with NOMESCO classifier and local service codes.  
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. There is one national identifier that allows linking together all encounters across databases.
9
Quality control (database specific)
The quality control procedures in the Estonian Biobank aim to remove the most obvious mistakes in the data, misspellings, impossible dates, duplicates. Before performing the ETL, several problems are fixed on the source data. Since the ETL procedures are used for a number of different datasets (from the same national sources), we have a growing number of pre-processing steps that correspond to the issues we have discovered previously in the data, such as checking for the presence of critical values, harmonizing date and unit of measurement formats, checking the validity of certain entries against classifiers, etc.
10
Linkage
Follow‐up data are available via linkage with national health‐related registries and via re‐examination of participants. Furthermore, electronic health records are updated for phenotypic outcome information every year. The EBB database is regularly linked with national registries, hospital databases, and the databases of the Estonian Health Insurance Fund (EHIF) and the National Health Information System (NHIS)
11
Vital status
Vital status (death date and causes of death) are obtained from the Causes of Death Registry.
12
Limitations
Participation in EBB cohort is voluntary, therefore the biobank does not represent a random sample and could be subject to recruitment bias. Although recruitment was open to everyone, there is a disproportion between ethnic Estonians and ethnic Russians in the biobank, with Estonians being overrepresented.

13
Main references
Milani, L., Alver, M., Laur, S. et al. The Estonian Biobank’s journey from biobanking to personalized medicine. Nat Commun 16, 3270 (2025). https://doi.org/10.1038/s41467-025-58465-3
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111114 Website: https://genomics.ut.ee/en/content/estonian-biobank

Finnish Care Register for Health Care (FinOMOP-THL)
#
Section
Description
1
Database Identification and country
FinOMOP-THL (Finnish Care Register for Health Care) Finland
2
Data partner information section
Finnish Institute for Health and Welfare (THL) Department of Knowledge Brokers
3
Coverage and timespan
Data collection since: 1998 Extent: Nation-wide. The current CDM population comprises all persons having been alive and residing in Finland since the beginning of 2011.
4
Healthcare setting / type of data
Primary care – gps, and primary care specialists (e.g. paediatricians), and secondary care – specialists (ambulatory or hospital outpatient care), and hospital inpatient care. The THL database covers both public and private, primary, and specialised inpatient and outpatient health care encounters in Finland, starting from 2011. The entire public sector and private inpatient encounters have been included since 2011, while private outpatient encounters, including occupational care, are included since 2020. Since 1998, the register has covered both public outpatient and inpatient specialized care and private inpatient care (TerveysHilmo). Since 2009, the Finnish National Vaccination Register is covered (complete since 2020). The vaccination register covers all vaccinations from the public sector and from a large part of private vaccination providers, with the data coverage from both sections being very good from 2020 onwards. Since 2011, the register has covered public primary care (AvoHilmo). Since 2020, the register has covered private outpatient care and occupational care. In addition, the CDM also contains positive COVID-19 test results from the Finnish National Infectious Diseases Register, which is maintained by THL.
5
Data collection process
Outpatient electronic health records, and Inpatient hospital electronic health records, and Registries. Data is entered by clinicians upon healthcare contact and processed by THL.
6
General representativeness
The THL data has national coverage and is therefore well representative of the Finnish population. Using the complete population as a basis for the person table also serves to facilitate calculations on a population level, e.g. incidence rates.
7
Data content /source coding
The following coding systems have been OMOP-mapped, typically to a good level of completeness: ICD10fi Finnish Extension, ATC, Toimenpideluokitus (procedure classification adapted from the Nordic Classification of Surgical Procedures (NCSP)), Terveydenhuollon erikoisalat (Hilmo specific provider speciality), Rokotustapa (AR/YDIN National classification for vaccine administration), Tupakointistatus (AR/YDIN National classification for smoking status). Vaccinations are identified on product level based on batch number, trade name, vaccine title, and ATC-code. This is mapped on brand and type in the OMOP CDM.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. Each patient in THL has a unique identifier.
9
Quality control (database specific)
The source data collection undergoes a structural and semantic validation before entry into the source database. Additionally, some coded variables undergo quality assessment against the respective code systems post entry into the database. The source registers are also assessed for completeness and coverage, with the aim of improving future collection in the areas where data is lacking.
10
Linkage
THL is already a linkage of multiple Finnish registries (see above).
11
Vital status
The National Population registry data forms the basis for forming the patient population. This ensures an up-to-date location (municipality of residence) of patients, as well as complete death occurrences (although not the cause of death).
12
Limitations
No database-specific limitations documented. General limitations for the data type applicable.
13
Main references
Häkkinen, Pirjo; Mölläri, Kaisa; Saukkonen, Sanna-Mari; Väyrynen, Riikka; Mielikäinen, Lasse; Järvelin, Jutta "Hilmo - Sosiaali- ja terveydenhuollon hoitoilmoitus 2020 : Määrittelyt ja ohjeistus : Voimassa 1.1.2020 alkaen" Terveyden ja hyvinvoinnin laitos (2019):
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111187 Website: https://thl.fi/fi/tilastot-ja-data/ohjeet-tietojen-toimittamiseen/hoitoilmoitusjarjestelma-hilmo

IQVIA Disease Analyzer Germany (IQVIA DA Germany)
#
Section
Description
1
Database Identification and country
IQVIA DA Germany (IQVIA Disease Analyzer Germany) Germany
2
Data partner information section
IQVIA
3
Coverage and timespan
Data collection since: 1989 Extent: Nation-wide. GP and specialists in Germany using specific patient management software.
4
Healthcare setting / type of data
Primary care – gps, and primary care specialists (e.g. paediatricians). Diagnoses, medication, and procedures from an ambulatory setting. Medications are recorded as prescriptions of marketed products.
5
Data collection process
Outpatient electronic health records. By clinicians at healthcare contact.
6
General representativeness
No specific details on general representativeness given.
7
Data content /source coding
Prescription is on product code level (German PZN), ICD10, NFC, Local lab coding.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. There can be patients registered under different ID numbers, because there is no linkage between different GPs.
9
Quality control (database specific)
Data is quality checked on plausibility.
10
Linkage
No.
11
Vital status
Death information is derived from medical events.
12
Limitations
No database-specific limitations documented. General limitations for the data type applicable.
13
Main references
No main reference provided.
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/104282 Website: https://www.iqvia.com/

Integrated Primary Care Information (IPCI)
#
Section
Description
1
Database Identification and country
IPCI (Integrated Primary Care Information) Netherlands
2
Data partner information section
Erasmus University Medical Center Department of Medical Informatics
3
Coverage and timespan
Data collection since: 2006 Extent: Nation-wide. IPCI is a Dutch database that contains patient records from 2006 onwards. However, it mainly covers the central part of the country, including the most densely populated area (the ‘Randstad’) and non-urban areas. IPCI contains information on all patients registered with GPs responsible for non-emergency care and referrals. A patient is registered at birth or at first encounter with the GP.
4
Healthcare setting / type of data
Primary care – gps. Data is collected from primary care EHR. This includes demographic information, complaints and symptoms, diagnoses, laboratory test results, lifestyle factors (in limited amount), and correspondence with secondary care, such as referral and discharge letters.
5
Data collection process
Outpatient electronic health records. Data is entered into the EHR system by the GPs, during or after the visit. Data is aggregated by Erasmus MC data managers and combined in one harmonized database. Several checks are done on this database to ensure correct data processing. Persons are mostly uniquely identified, with the exception of when persons change GP practice (when the same individual can receive several different identifiers).
6
General representativeness
More than 99% of the Dutch population has health insurance, and almost all citizens are registered with a general practitioner. Over 12 months, around 78% of the population has at least one contact with their GP. IPCI included around 350 GP practices out of around 5000 in the country (~ 7%). The demographic composition of the IPCI population mirrors that of the general Dutch population in terms of age and sex.
7
Data content /source coding
Dutch GPs use mainly Dutch standard codes, like ICPC-1 and Diagnostische Bepalingen maintained by NHG. And for therapy the G-Standard is used, maintained by ZIndex.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. Patients can be registered under different IDs. However, in the Netherlands, patients typically have one GP and changing practice is uncommon.
9
Quality control (database specific)
Prior to each data release, extensive quality control steps are performed, e.g., comparison of patient characteristics between practices, and checks to identify abnormal temporal data patterns in practices. For each practice, around 200 quality indicators are obtained. Of these indicators, a quarter refer to population characteristics, e.g. number of birth and mortalities relative to practice size, temporal consistency. The other indicators are based on medical data, e.g. distribution of measurement values, frequencies of diagnoses and procedures relative to age, completeness of data. The indicators are combined in a couple of quality scores for each practice. For these scores, cut-off values for acceptable quality have been defined. Practices with a score below a cut-off are excluded for research. This approach has shown to be very important, for example to check if data from practices that just joined the database are at an acceptable level of quality. The details of the approach, like the cut-off values for acceptance, are based on years of experience. In addition, trends are compared with the previous database release. Extensive quality control steps are performed before each data release. These include comparing patient characteristics between practices and checks to identify abnormal temporal data patterns in practices. Additional checks include over 200 indicators related to population characteristics (e.g., reliability of birth and mortality rates) and medical data (e.g., availability of durations of prescriptions and completeness of laboratory results). Records of low quality are excluded from the database.
10
Linkage
Linkage requires additional approval steps and needs to be assessed on a case-by-case basis. IPCI is not routinely linked with other databases.
11
Vital status
Vital status (death date and cause) is collected based on GP records.
12
Limitations
The main limitation comes with the fact that IPCI is limited to GP records, and although it contains information on referrals and discharge letters, it may not fully capture specific hospital information. IPCI does not include coded/detailed data about medications/procedures/test results from the hospital or other care-providers.
13
Main references
de Ridder MAJ, de Wilde M,de Ben C,Leyba AR,Mosseveld BMT,Verhamme KMC,van der Lei J,Rijnbeek PR "Data Resource Profile: The Integrated Primary Care Information (IPCI) database, The Netherlands." International journal of epidemiology (2022): 35182143
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/42618 Website: http://www.ipci.nl

The Information System for the Development of Research on Primary Care (SIDIAP)
#
Section
Description
1
Database Identification and country
SIDIAP (The Information System for the Development of Research in Primary Care) Catalunya, Spain
2
Data partner information section
IDIAPJGol
3
Coverage and timespan
Data collection since: 2006 Extent: Regional. The SIDIAP database contains records of around 6 million people residing in Catalonia, estimated to be representing around 76% of the Catalan population.
4
Healthcare setting / type of data
Primary care – gps, and hospital inpatient care. SIDIAP captured data includes routine visits, sociodemographic information, diagnoses, laboratory tests, drugs (prescribed and dispensed), referrals, and lifestyle information.
5
Data collection process
Outpatient electronic health records, and Inpatient hospital electronic health records, and Other. Data is entered by primary care physicians upon healthcare contact, supplemented with hospital discharge records. The Institut Catala de la Salut is the owner of the data and acts as the data controller.
6
General representativeness
It was previously shown that the captured SIDIAP population is highly representative of the entire Catalan region in terms of geographic, age, and sex distributions.
7
Data content /source coding
SIDIAP data covers all services that occur at the Primary Care Centres, as well as support services, such as sexual and reproductive health or home end-of-life care. Drugs are coded in ATC-WHO terminology in the source data. Health outcomes are captured in ICD-10CM codes. The SIDIAP contains all laboratory tests and results performed in primary health centres. Demographics, geographical, as well as socio-economic factors are recorded for each patient.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. No.
9
Quality control (database specific)
Internal and external validation processes are carried out to determine the data quality of the SIDIAP information at each data update. These include stratifying the data by geographical regions and year in order to identify differences in data collection that need to be harmonized (e.g. recording of specific information under different codes). The measurement units of variables measuring one characteristic are also homogenized (e.g. transformation of the data from every laboratory that measures haemoglobin to grams per decilitre). Visual inspection of all data included in the database by week is also conducted, allowing one to see temporal patterns in the registry of a certain variable. With this information, the SIDIAP team can issue recommendations to researchers about the most common variable(s) where certain information is recorded (e.g., there are several variables with information concerning the women’s menopausal status and with these visual inspection tools the SIDIAP team can inform the researchers about which related variables have the largest number of records and could be more helpful to capture menopause). Data availability (longitudinally and reliability), plausibility (range checks and unusual values), and consistency are inspected through visualisation tools. In addition, before accessing the data for a requested project, research teams have access to a quality-control report. This document contains counts, years, percentiles, maximums and minimums, incidences, and prevalence of the data requested for the project, allowing detection of inconsistencies in the data extraction prior to data delivery. External validation processes of the SIDIAP database mainly include assessing the data recorded in SIDIAP through linkage to external gold standard data sources, by analysing free text, or by sending questionnaires to health professionals.
10
Linkage
SIDIAP is linked to a hospital discharge database, pharmacy dispensation, and primary care laboratories. It can also be linked to other registries in Catalonia on a project-by-project basis.
11
Vital status
Mortality is fully captured in SIDIAP. The cause of death is not available but can be linked to the Spanish death registry on a project-by-project basis.
12
Limitations
The SIDIAP data is not representative of individuals not using public primary care, and conditions that are usually followed by specialist care might not be properly captured. In addition, there is limited information on lifestyle variables (not always requested in primary care visit and, therefore, the information is missing in many cases). Patients are followed until Death or when transferring to another primary health care centre that does not contribute to SIDIAP.
13
Main references
Recalde M, Rodríguez C,Burn E,Far M,García D,Carrere-Molina J,Benítez M,Moleras A,Pistillo A,Bolíbar B,Aragón M,Duarte-Salles T "Data Resource Profile: The Information System for Research in Primary Care (SIDIAP)." International journal of epidemiology (2022): 35415748
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/50190 Website: https://www.sidiap.org/index.php/en

Clinical Practice Research Datalink GOLD (Oxford) (CPRD GOLD)
#
Section
Description
1
Database Identification and country
CPRD GOLD (Clinical Practice Research Datalink GOLD ) United Kingdom
2
Data partner information section
University of Oxford NDORMS
3
Coverage and timespan
Data collection since: 1987 Extent: Nation-wide. CPRD GOLD consists of patients in contributing practices using Vision software. Historically this covered the whole of the UK, but the number of contributing practices in the England is dropping. In January 2025 only 3 practices from England were a part of CPRD GOLD, while historical patient data were from the whole of the UK, and will continue to be so. In the future, no practices from England will be present, only practices from Scotland, Wales, and Northern Ireland.
4
Healthcare setting / type of data
Primary care – gps, and primary care specialists (e.g. paediatricians), and secondary care – specialists (ambulatory or hospital outpatient care), and hospital inpatient care. CPRD GOLD data include patient demographics, biological measurements, clinical symptoms and diagnoses, referrals to specialist/hospital and their outcome, laboratory tests/results, and prescribed medications.
5
Data collection process
Outpatient electronic health records. Data is entered by clinicians into the EHR. Data is processed by CPRD and provides data releases for research.
6
General representativeness
CPRD GOLD has been assessed and found to be broadly representative of the UK general population in terms of age, gender, and ethnicity. In CPRD GOLD in January 2025 there were 2,730,707 current acceptable patients (i.e. registered at currently contributing practices that use Vision software, excluding transferred out, deceased patients, and those flagged by CPRD as not acceptable for clinical research for data quality issues). This equals to 4.07%, based on the UK population estimates of 67,026,300 from the Office of National Statistics (mid-2023). Current patients are only from Scotland, Wales, and Northern Ireland. Historically, GOLD does contain data from England as well.
7
Data content /source coding
Gemscript, Read, dm+d
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. In GOLD, a patient can be registered under different ID numbers upon changing practice or re-registration. Researchers are not able to identify these patients, as the data are anonymised. However, GOLD covers less than 5% of the current UK GP practices and it is unlikely that an individual who does change GP practice ends up in another GP practice which uses the Vision software and accepts the CPRD data collection agreement. The very small number of duplicated IDs will have different observation periods and should not have an impact on the data analyses.
9
Quality control (database specific)
CPRD GOLD only includes practices whose data quality is assessed to be up-to-standard (uts). Each practice is associated to an uts date set when the data quality standards become satisfactory, and CPRD recommend using only longitudinal data starting from this uts date. Every time CPRD collect the EHR from a practice, checks are run for the data quality standards and if they are not adequate, the EHR is not accepted. When the data quality becomes acceptable again, CPRD updates the practice uts date. CPRD also check data quality standards at the patient level and associate each patient to a flag, reporting if its data is acceptable for clinical research. Only patients with acceptable data quality are included in the population to be mapped to CDM.
10
Linkage
CPRD GOLD can be linked to several sources, however our Oxford OMOP CDM is only linked to the CPRD GOLD Ethnicity Record and to the CPRD Townsend Deprivation Index at Practice Level
11
Vital status
Vital status is retrieved from the GP records. Population registry (ONS) data can be requested on a study-by-study basis and linked. This data only covers England and is planned to be mapped to OMOP in the future. The cause of death is not captured.
12
Limitations
The main limitation is due to the fact that CPRD GOLD is limited to GP records, and although it contains information on referrals and discharge letters, it may not fully capture specific hospital information. Events from hospital and specialist care are not covered.
13
Main references
Sanchez-Santos MT, Axson EL,Dedman D,Delmestri A "Data Resource Profile Update: CPRD GOLD." International journal of epidemiology (2025): 40499193
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111113 Website: https://cprd.com

UK BioBank (UKBB)
#
Section
Description
1
Database Identification and country
UKBB (UK BioBank) United Kingdom
2
Data partner information section
Oxford University NDORMS
3
Coverage and timespan
Data collection since: 2006 Extent: Nation-wide. People recruited from whole of the UK.
4
Healthcare setting / type of data
Primary care – gps, and primary care specialists (e.g. paediatricians), and secondary care – specialists (ambulatory or hospital outpatient care), and hospital inpatient care, and other (specify). UK Biobank is made by a rich variety of data sources, which include genetic data, primary care data, hospital inpatient data, death data, and cancer registry.
5
Data collection process
Inpatient hospital electronic health records, and Registries, and Biobank. The baseline assessment is consisting of both patient reported data (questionnaire) and physical measurements. GP and hospital data, as well as death and cancer registry records, are linked afterwards and are subject to data validation: https://biobank.ctsu.ox.ac.uk/~bbdatan/Data_cleaning_overall_doc_showcase_v1.pdf
6
General representativeness
The database population consists of volunteers aged 40-69years. We can expect that volunteers might have been more willing to participate if living closer to one of the 22 recruitment centres or if more interested in health issues compared to the general population. These aspects might have introduced an unavoidable bias in the cohort.
7
Data content /source coding
READ2, READ3, DM+D, ICD9, ICD10, OPCS3, OPCS4, ICD-O-3 are used.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. No.
9
Quality control (database specific)
All UK BioBank data are provided already curated and each of the many datasets have specific curation algorithms and procedures. As always, primary care and hospital data, which come from real-world setting, need special attention regarding data quality. Please refer to the link below for specific details https://biobank.ndph.ox.ac.uk/showcase/showcase/docs/primary_care_data.pdf https://biobank.ndph.ox.ac.uk/showcase/showcase/docs/HospitalEpisodeStatistics.pdf
10
Linkage
The database contains liked data from death and cancer registry, GP and hospital for the participants.
11
Vital status
Linked to the national death registry.
12
Limitations
The UKBB source data will not be updated anymore and have no new records after December 2022. There is no day's supply information captured in the source. GP prescription data are available for 45% of the cohort. There is no information on dispensed medicines. GP laboratory tests and results are not available.
13
Main references
Hewitt J, Walters M,Padmanabhan S,Dawson J "Cohort profile of the UK Biobank: diagnosis and characteristics of cerebrovascular disease." BMJ open (2016): 27006341
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111233 Website: https://www.ukbiobank.ac.uk/

 
ANNEX II. Fitness for use assessment
Data source justification for inclusion and key characteristics
The selected data sources met the criteria required to capture outcomes of interest and relevant data, enabling a patient-level characterisation of newly diagnosed individuals with cancer across different European settings and regions. The main criterion was a meaningful number of person counts for the population of interest (individuals with cancer) and outcomes (thromboembolic events) assessed at the feasibility stage for all data sources included in the study. Data sources were also selected based on European representativeness. Not all data sources had records of all outcomes of interest.
Additional criteria reflect other data quality domains assessed at the DARWIN EU® data partners onboarding stage. With every new release of the data partners OMOP CDM, the DARWIN EU® coordination centre also receives new results of the CdmOnboarding, DashboardExport, and DataQualityDashboard packages and assesses the quality of the data. No open quality issues related to the study population and outcomes were present for any of the data sources selected for the study.
Relevance was also assessed based on the previous research related to the study population of the outcome. Previously, IPCI, EBB, SIDIAP, CPRD, IQVIA DA Germany, UKBB, and EBB were used in studies with thromboembolic events as an outcome. (Ali et al., 2020, Mercadé-Besora et al., 2024, Li et al., 2022, Voss et al, 2023). CPRD, FinOMOP-THL, UKBB, SIDIAP, IPCI, and UKBB were used in studies repeated on individuals with cancer (Hagberg et al., 2023; Corby et al., 2024, Leinonen et al., 2017, Smith et al., 2024; Recalde et al, 2019; van Soest 2008; Chen et al., 2024). 
In addition to that, DK-DHR is a nationwide, fully representative data source that includes information from the National Patient Registry and the National Cancer Register.

Design elements
Operational definition
Data elements for valid capture
Criticality of the quality of the element, including justification where relevant
Study population
Objective 1
Inclusion criteria
• First diagnosis of a selected cancer (index date) between 01/01/2016 and 31/12/2022
• Age ≥18 years at index date
• Minimum 365 days of available history before index date
• Index date ≥365 days prior to end of data availability of the data source
Exclusion criteria
• Diagnoses of multiple primary tumours at index date
• History of cancer diagnosis ever before index date
• Outcome during the year prior to index date
Objective 2
Inclusion criteria
• Included in the study population of objective 1
• Occurrence of the outcome during follow-up
• First diagnosis of a selected cancer
Diagnosis of:
• Deep vein thrombosis (DVT)
• Pulmonary embolisms (PE)
• Venous thromboembolism (VTE, composite of DVT and PE)
• Pelvic venous thrombosis (PVT)
• Splanchnic vein thrombosis (SVT), including hepatic and extra-hepatic vein thrombosis
• Retinal vein thrombosis (RVT), including retinal central vein thrombosis
• Disseminated intravascular coagulation (DIC)
Low/Medium/High
Treatment/ exposure
Not applicable.

Low/Medium/High
Comparator group (if relevant)
Not applicable.

Low/Medium/High
Outcomes (if relevant)
All objectives
The thromboembolic event outcomes in this study are identical to those in EUPAS1000000440, of which this is a routinely repeated study:
• Deep vein thrombosis (DVT)
• Pulmonary embolisms (PE)
• Venous thromboembolism (VTE, composite of DVT and PE)
• Pelvic venous thrombosis (PVT)
• Splanchnic vein thrombosis (SVT), including hepatic and extra-hepatic vein thrombosis
• Retinal vein thrombosis (RVT), including retinal central vein thrombosis
• Disseminated intravascular coagulation (DIC)
Each of the specific thromboembolic events was a primary, binary outcome. The outcome was assessed at any diagnosis position in the electronic health record and from any of the care settings in each data source.
The list of concepts, from EUPAS1000000440 and based on SNOMED codes and aligned with previous studies that used OMOP CDM and VTE as an outcome (Burn et al., 2022), is provided in Annex IV.
Diagnosis of:
• Deep vein thrombosis (DVT)
• Pulmonary embolisms (PE)
• Venous thromboembolism (VTE, composite of DVT and PE)
• Pelvic venous thrombosis (PVT)
• Splanchnic vein thrombosis (SVT), including hepatic and extra-hepatic vein thrombosis
• Retinal vein thrombosis (RVT), including retinal central vein thrombosis
• Disseminated intravascular coagulation (DIC)
Low/Medium/High
Covariates (including confounders if relevant)
All Objectives
• Sex
◦ Female/male
• Age groups (years) at index date:
◦ 18–34
◦ 35–44
◦ 45–54
◦ 55–64
◦ 65–74
◦ 75–84
◦ ≥85
Age at index date will be calculated using January 1st of the year of birth as proxy for the actual birthday. Date/month is either not present or cannot be made available for governance reasons. If available, date is often set to first of the month for personal privacy.
• Study subperiod:
◦ 2016–2019
◦ 2020-2022
The status of each covariate will be assessed at index date. Each covariate will be used to stratify the results.
• Date of first diagnosis of selected cancer (index date)
• Sex
• Age at index date
Low/Medium/High
Follow-up time (if relevant)
For both objectives, follow-up will start on the date of cancer diagnosis (index date) and end on the earliest of occurrence of the outcome, loss to follow-up, end of data availability, death, or end of the study period (31/12/2024).
• Date of first diagnosis of selected cancer (index date)
• Date of thromboembolic event
• Data of end of data availability
• Death date
Low/Medium/High
EMA Data Quality Framework for EU medicines regulation: application to Real-World Data for more information (https://www.ema.europa.eu/system/files/documents/other/data-quality-framework-eu-medicines-regulation-application-real-world-data_en.pdf)
ANNEX III. Operational and reporting considerations
DATA MANAGEMENT
Data management
All data sources have previously mapped their data to the OMOP common data model. This enables the use of standardised analytics and using DARWIN EU® tools across the network, since the structure of the data and the terminology system is harmonised. The OMOP CDM was developed and maintained by the Observational Health Data Sciences and Informatics (OHDSI) initiative and is described in detail on the wiki page of the CDM: https://ohdsi.github.io/CommonDataModel and in The Book of OHDSI: http://book.ohdsi.org.
The analytic code for this study will be written in R and will use standardized analytics wherever possible. Each data partner will execute the study code against their data source containing patient-level data and then return the results (csv files), which will only contain aggregated data. The results from each of the contributing data sites will then be combined in tables and figures for the study report.
Data storage and protection
For this study, participants from various EU member states will process personal data from individuals that is collected in national/regional electronic health record data sources. Due to the sensitive nature of this personal medical data, it is important to be fully aware of ethical and regulatory aspects and to strive to take all reasonable measures to ensure compliance with ethical and regulatory issues on privacy.
All data sources used in this study are already used for pharmaco-epidemiological research and have a well-developed mechanism to ensure that European and local regulations dealing with ethical use of the data and adequate privacy control are adhered to. In agreement with these regulations, rather than combining person level data and performing only a central analysis, local analyses will be run, which generate non-identifiable aggregate summary results. 
The output files are stored in the DARWIN EU® Remote Research Environment (RRE). These output files do not contain any data that allow identification of subjects included in the study. The RRE implements further security measures to ensure a high level of stored data protection to comply with the local implementation of the General Data Protection Regulation (GDPR) (EU) 679/20161 in the various member states. 
QUALITY CONTROL
Data source quality control
When defining drug cohorts, non-systemic products will be excluded from the list of included codes summarised on the ingredient level.
When defining cohorts for indications, a systematic search of possible codes for inclusion will be identified using the CodelistGenerator R package (https://github.com/darwin-eu/CodelistGenerator). This package allows the user to define a search strategy and will use this to query the vocabulary tables of the OMOP common data model so as to find potentially relevant codes. In addition, the CohortDiagnostics (https://github.com/OHDSI/CohortDiagnostics) and DrugExposureDiagnostics (https://cran.r-project.org/web/packages/DrugExposureDiagnostics/index.html) R packages will be run, if needed, to assess the use of different codes across the data sources contributing to the study and identify any codes potentially omitted in error. The DrugExposureDiagnostics package evaluates ingredient-specific attributes and patterns in drug exposure records.
The study code will be based on DARWIN EU® R packages: IncidencePrevalence to estimate Incidence and Prevalence, DrugUtilisation to characterise the drug use, and CohortCharacteristics to characterise the cohort by indication. These packages will include numerous automated unit tests to ensure the validity of the codes, alongside software peer review and user testing. The R package will be made publicly available via GitHub. 
PLANS FOR DISSEMINATING AND COMMUNICATING STUDY RESULTS
A PDF report including an executive summary, and the specified tables and/or figures will be submitted to EMA by the DARWIN EU® CC upon completion of the study.
An interactive dashboard incorporating all the results (tables and figures) will be provided alongside the PDF report. The full set of underlying aggregated data used in the dashboard will also be made available, if requested.

ANNEX IV. List of stand-alone documents
Concepts to define individuals with cancer (populations of interest) are available in a stand-alone document (DARWIN_EU_P3_C3_005 _Cancer_phenotypes.xlsx). Tables S1 to S7 include concepts used to define outcomes (thromboembolic events).

Table S1. List of concepts used to define deep vein thrombosis (DVT).
Concept ID
Concept Name
Domain
Vocabulary
762047
Acute bilateral thrombosis of subclavian veins
Condition
SNOMED
762148
Acute deep vein thrombosis of bilateral iliac veins
Condition
SNOMED
37169261
Acute deep vein thrombosis of bilateral lower limbs following procedure
Condition
SNOMED
37169249
Acute deep vein thrombosis of bilateral upper limbs following procedure
Condition
SNOMED
35616028
Acute deep vein thrombosis of left iliac vein
Condition
SNOMED
35615035
Acute deep vein thrombosis of left lower limb following procedure
Condition
SNOMED
35615031
Acute deep vein thrombosis of left upper limb following procedure
Condition
SNOMED
43531681
Acute deep vein thrombosis of lower limb
Condition
SNOMED
35616027
Acute deep vein thrombosis of right iliac vein
Condition
SNOMED
35615034
Acute deep vein thrombosis of right lower limb following procedure
Condition
SNOMED
35615030
Acute deep vein thrombosis of right upper limb following procedure
Condition
SNOMED
44782746
Acute deep venous thrombosis
Condition
SNOMED
44782751
Acute deep venous thrombosis of axillary vein
Condition
SNOMED
762008
Acute deep venous thrombosis of bilateral axillary veins
Condition
SNOMED
760875
Acute deep venous thrombosis of bilateral calves
Condition
SNOMED
765155
Acute deep venous thrombosis of bilateral ileofemoral veins
Condition
SNOMED
762017
Acute deep venous thrombosis of bilateral internal jugular veins
Condition
SNOMED
762417
Acute deep venous thrombosis of bilateral legs
Condition
SNOMED
761461
Acute deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
762020
Acute deep venous thrombosis of bilateral popliteal veins
Condition
SNOMED
765546
Acute deep venous thrombosis of bilateral tibial veins
Condition
SNOMED
762004
Acute deep venous thrombosis of both upper extremities
Condition
SNOMED
44782742
Acute deep venous thrombosis of calf
Condition
SNOMED
44782747
Acute deep venous thrombosis of femoral vein
Condition
SNOMED
762015
Acute deep venous thrombosis of ileofemoral vein of left leg
Condition
SNOMED
765541
Acute deep venous thrombosis of ileofemoral vein of right lower extremity
Condition
SNOMED
44782748
Acute deep venous thrombosis of iliofemoral vein
Condition
SNOMED
44782752
Acute deep venous thrombosis of internal jugular vein
Condition
SNOMED
762009
Acute deep venous thrombosis of left axillary vein
Condition
SNOMED
760876
Acute deep venous thrombosis of left calf
Condition
SNOMED
765540
Acute deep venous thrombosis of left femoral vein
Condition
SNOMED
765922
Acute deep venous thrombosis of left internal jugular vein
Condition
SNOMED
762418
Acute deep venous thrombosis of left lower extremity
Condition
SNOMED
761462
Acute deep venous thrombosis of left pelvic vein
Condition
SNOMED
618482
Acute deep venous thrombosis of left peroneal vein
Condition
SNOMED
765537
Acute deep venous thrombosis of left upper extremity
Condition
SNOMED
44782767
Acute deep venous thrombosis of lower extremity as complication of procedure
Condition
SNOMED
44782761
Acute deep venous thrombosis of pelvic vein
Condition
SNOMED
762022
Acute deep venous thrombosis of politeal vein of right leg
Condition
SNOMED
44782743
Acute deep venous thrombosis of popliteal vein
Condition
SNOMED
762021
Acute deep venous thrombosis of popliteal vein of left leg
Condition
SNOMED
762010
Acute deep venous thrombosis of right axillary vein
Condition
SNOMED
760877
Acute deep venous thrombosis of right calf
Condition
SNOMED
762013
Acute deep venous thrombosis of right femoral vein
Condition
SNOMED
762018
Acute deep venous thrombosis of right internal jugular vein
Condition
SNOMED
762419
Acute deep venous thrombosis of right lower extremity
Condition
SNOMED
765229
Acute deep venous thrombosis of right pelvic vein
Condition
SNOMED
618681
Acute deep venous thrombosis of right peroneal vein
Condition
SNOMED
762005
Acute deep venous thrombosis of right upper extremity
Condition
SNOMED
44782745
Acute deep venous thrombosis of thigh
Condition
SNOMED
44782744
Acute deep venous thrombosis of tibial vein
Condition
SNOMED
762026
Acute deep venous thrombosis of tibial vein of left leg
Condition
SNOMED
765156
Acute deep venous thrombosis of tibial vein of right leg
Condition
SNOMED
44782421
Acute deep venous thrombosis of upper extremity
Condition
SNOMED
44782766
Acute deep venous thrombosis of upper extremity as complication of procedure
Condition
SNOMED
37171353
Acute ischemia of colon due to thrombosis of mesenteric vein
Condition
SNOMED
37170675
Acute ischemia of small intestine due to thrombosis of mesenteric vein
Condition
SNOMED
762048
Acute thrombosis of left subclavian vein
Condition
SNOMED
45757410
Acute thrombosis of mesenteric vein
Condition
SNOMED
762049
Acute thrombosis of right subclavian vein
Condition
SNOMED
36712892
Acute thrombosis of splenic vein
Condition
SNOMED
44782762
Acute thrombosis of subclavian vein
Condition
SNOMED
4179911
Axillary vein thrombosis
Condition
SNOMED
37109253
Bilateral acute deep vein thrombosis of femoral veins
Condition
SNOMED
618678
Bilateral acute deep venous thrombosis of peroneal veins
Condition
SNOMED
609003
Bilateral deep femoral vein thrombophlebitis
Condition
SNOMED
3179900
Bilateral deep vein thromboses
Condition
Nebraska Lexicon
40478951
Bilateral deep vein thrombosis of lower extremities
Condition
SNOMED
609002
Bilateral femoral vein thrombophlebitis
Condition
SNOMED
608965
Bilateral iliac vein thrombophlebitis
Condition
SNOMED
1245776
Bilateral popliteal vein thrombophlebitis
Condition
SNOMED
609006
Bilateral tibial vein thrombophlebitis
Condition
SNOMED
4042396
Deep thrombophlebitis
Condition
SNOMED
4046884
Deep vein thrombosis of leg related to air travel
Condition
SNOMED
3655221
Deep vein thrombosis of lower extremity due to intravenous drug use
Condition
SNOMED
4133004
Deep venous thrombosis
Condition
SNOMED
761013
Deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
37163011
Deep venous thrombosis of calf
Condition
SNOMED
45773536
Deep venous thrombosis of femoropopliteal vein
Condition
SNOMED
763942
Deep venous thrombosis of left lower extremity
Condition
SNOMED
1075379
Deep venous thrombosis of left posterior tibial vein
Condition
SNOMED
761980
Deep venous thrombosis of left upper extremity
Condition
SNOMED
443537
Deep venous thrombosis of lower extremity
Condition
SNOMED
4133975
Deep venous thrombosis of pelvic vein
Condition
SNOMED
40480555
Deep venous thrombosis of peroneal vein
Condition
SNOMED
1075377
Deep venous thrombosis of posterior tibial vein
Condition
SNOMED
4322565
Deep venous thrombosis of profunda femoris vein
Condition
SNOMED
763941
Deep venous thrombosis of right lower extremity
Condition
SNOMED
1075378
Deep venous thrombosis of right posterior tibial vein
Condition
SNOMED
761928
Deep venous thrombosis of right upper extremity
Condition
SNOMED
4207899
Deep venous thrombosis of tibial vein
Condition
SNOMED
4028057
Deep venous thrombosis of upper extremity
Condition
SNOMED
193512
Embolism and thrombosis of the renal vein
Condition
SNOMED
435565
Embolism and thrombosis of the vena cava
Condition
SNOMED
4258295
Embolism from thrombosis of vein of distal lower extremity
Condition
SNOMED
40481089
Embolism from thrombosis of vein of lower extremity
Condition
SNOMED
40479840
Embolism from thrombosis of vein of thigh
Condition
SNOMED
4119760
Iliofemoral deep vein thrombosis
Condition
SNOMED
4124856
Inferior mesenteric vein thrombosis
Condition
SNOMED
608964
Left iliac vein thrombophlebitis
Condition
SNOMED
602592
Left peroneal vein thrombophlebitis
Condition
SNOMED
600938
Left subclavian vein thrombophlebitis
Condition
SNOMED
37164448
Lemierre syndrome
Condition
SNOMED
4281689
Phlegmasia alba dolens
Condition
SNOMED
4284538
Phlegmasia cerulea dolens
Condition
SNOMED
3185768
Popliteal vein thrombosis
Condition
Nebraska Lexicon
4309333
Postoperative deep vein thrombosis
Condition
SNOMED
1245858
Postpartum acute deep vein thrombosis
Condition
SNOMED
46285905
Provoked deep vein thrombosis
Condition
SNOMED
608963
Right iliac vein thrombophlebitis
Condition
SNOMED
602583
Right peroneal vein thrombophlebitis
Condition
SNOMED
600939
Right subclavian vein thrombophlebitis
Condition
SNOMED
4033521
Splenic vein thrombosis
Condition
SNOMED
4055089
Superior mesenteric vein thrombosis
Condition
SNOMED
4230403
Thrombophlebitis of axillary vein
Condition
SNOMED
4069561
Thrombophlebitis of deep femoral vein
Condition
SNOMED
761831
Thrombophlebitis of deep vein of bilateral lower limbs
Condition
SNOMED
761830
Thrombophlebitis of deep vein of left lower limb
Condition
SNOMED
761808
Thrombophlebitis of deep vein of left upper limb
Condition
SNOMED
761832
Thrombophlebitis of deep vein of right lower limb
Condition
SNOMED
761809
Thrombophlebitis of deep vein of right upper limb
Condition
SNOMED
4221821
Thrombophlebitis of deep veins of lower extremity
Condition
SNOMED
440750
Thrombophlebitis of deep veins of upper extremities
Condition
SNOMED
4203618
Thrombophlebitis of femoropopliteal vein
Condition
SNOMED
4176614
Thrombophlebitis of iliac vein
Condition
SNOMED
764715
Thrombophlebitis of internal jugular vein
Condition
SNOMED
608904
Thrombophlebitis of left axillary vein
Condition
SNOMED
761821
Thrombophlebitis of left deep femoral vein
Condition
SNOMED
761819
Thrombophlebitis of left femoral vein
Condition
SNOMED
609000
Thrombophlebitis of left popliteal vein
Condition
SNOMED
609005
Thrombophlebitis of left tibial vein
Condition
SNOMED
4318407
Thrombophlebitis of mesenteric vein
Condition
SNOMED
608903
Thrombophlebitis of right axillary vein
Condition
SNOMED
761820
Thrombophlebitis of right deep femoral vein
Condition
SNOMED
761818
Thrombophlebitis of right femoral vein
Condition
SNOMED
609001
Thrombophlebitis of right popliteal vein
Condition
SNOMED
609004
Thrombophlebitis of right tibial vein
Condition
SNOMED
4205652
Thrombophlebitis of subclavian vein
Condition
SNOMED
4110339
Thrombophlebitis of the anterior tibial vein
Condition
SNOMED
4111868
Thrombophlebitis of the common iliac vein
Condition
SNOMED
4110343
Thrombophlebitis of the external iliac vein
Condition
SNOMED
439314
Thrombophlebitis of the femoral vein
Condition
SNOMED
4109877
Thrombophlebitis of the internal iliac vein
Condition
SNOMED
4112171
Thrombophlebitis of the popliteal vein
Condition
SNOMED
4112172
Thrombophlebitis of the posterior tibial vein
Condition
SNOMED
4250765
Thrombophlebitis of tibial vein
Condition
SNOMED
42538533
Thrombosis of iliac vein
Condition
SNOMED
44811347
Thrombosis of internal jugular vein
Condition
SNOMED
765049
Thrombosis of left peroneal vein
Condition
SNOMED
4317289
Thrombosis of mesenteric vein
Condition
SNOMED
4203836
Thrombosis of subclavian vein
Condition
SNOMED
4175649
Thrombosis of the popliteal vein
Condition
SNOMED
4153353
Traumatic thrombosis of axillary vein
Condition
SNOMED
46285904
Unprovoked deep vein thrombosis
Condition
SNOMED
37163265
Venous thromboembolism due to thrombosis of vein of lower limb
Condition
SNOMED

Table S2. List of concepts used to define pulmonary embolism (PE).
Concept ID
Concept Name
Domain
Vocabulary
608954
Acute cor pulmonale due to septic pulmonary embolism
Condition
SNOMED
4120091
Acute massive pulmonary embolism
Condition
SNOMED
45768439
Acute pulmonary embolism
Condition
SNOMED
45768888
Acute pulmonary thromboembolism
Condition
SNOMED
762808
Infarction of lung due to embolus
Condition
SNOMED
40480461
Infarction of lung due to iatrogenic pulmonary embolism
Condition
SNOMED
4108681
Postoperative pulmonary embolus
Condition
SNOMED
37160752
Postoperative pulmonary thromboembolism
Condition
SNOMED
1244882
Pulmonary artery embolism due to foreign body
Condition
SNOMED
440417
Pulmonary embolism
Condition
SNOMED
37109911
Pulmonary embolism due to and following acute myocardial infarction
Condition
SNOMED
37016922
Pulmonary embolism on long-term anticoagulation therapy
Condition
SNOMED
43530605
Pulmonary embolism with pulmonary infarction
Condition
SNOMED
4253796
Pulmonary microemboli
Condition
SNOMED
4121618
Pulmonary thromboembolism
Condition
SNOMED
36713113
Saddle embolus of pulmonary artery
Condition
SNOMED
35615055
Saddle embolus of pulmonary artery with acute cor pulmonale
Condition
SNOMED
40479606
Septic pulmonary embolism
Condition
SNOMED
4119607
Subacute massive pulmonary embolism
Condition
SNOMED

Table S3. List of concepts used to define venous thromboembolism (VTE).
Concept ID
Concept Name
Domain
Vocabulary
762047
Acute bilateral thrombosis of subclavian veins
Condition
SNOMED
608954
Acute cor pulmonale due to septic pulmonary embolism
Condition
SNOMED
762148
Acute deep vein thrombosis of bilateral iliac veins
Condition
SNOMED
37169261
Acute deep vein thrombosis of bilateral lower limbs following procedure
Condition
SNOMED
37169249
Acute deep vein thrombosis of bilateral upper limbs following procedure
Condition
SNOMED
35616028
Acute deep vein thrombosis of left iliac vein
Condition
SNOMED
35615035
Acute deep vein thrombosis of left lower limb following procedure
Condition
SNOMED
35615031
Acute deep vein thrombosis of left upper limb following procedure
Condition
SNOMED
43531681
Acute deep vein thrombosis of lower limb
Condition
SNOMED
35616027
Acute deep vein thrombosis of right iliac vein
Condition
SNOMED
35615034
Acute deep vein thrombosis of right lower limb following procedure
Condition
SNOMED
35615030
Acute deep vein thrombosis of right upper limb following procedure
Condition
SNOMED
44782746
Acute deep venous thrombosis
Condition
SNOMED
44782751
Acute deep venous thrombosis of axillary vein
Condition
SNOMED
762008
Acute deep venous thrombosis of bilateral axillary veins
Condition
SNOMED
760875
Acute deep venous thrombosis of bilateral calves
Condition
SNOMED
765155
Acute deep venous thrombosis of bilateral ileofemoral veins
Condition
SNOMED
762017
Acute deep venous thrombosis of bilateral internal jugular veins
Condition
SNOMED
762417
Acute deep venous thrombosis of bilateral legs
Condition
SNOMED
761461
Acute deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
762020
Acute deep venous thrombosis of bilateral popliteal veins
Condition
SNOMED
765546
Acute deep venous thrombosis of bilateral tibial veins
Condition
SNOMED
762004
Acute deep venous thrombosis of both upper extremities
Condition
SNOMED
44782742
Acute deep venous thrombosis of calf
Condition
SNOMED
44782747
Acute deep venous thrombosis of femoral vein
Condition
SNOMED
762015
Acute deep venous thrombosis of ileofemoral vein of left leg
Condition
SNOMED
765541
Acute deep venous thrombosis of ileofemoral vein of right lower extremity
Condition
SNOMED
44782748
Acute deep venous thrombosis of iliofemoral vein
Condition
SNOMED
44782752
Acute deep venous thrombosis of internal jugular vein
Condition
SNOMED
762009
Acute deep venous thrombosis of left axillary vein
Condition
SNOMED
760876
Acute deep venous thrombosis of left calf
Condition
SNOMED
765540
Acute deep venous thrombosis of left femoral vein
Condition
SNOMED
765922
Acute deep venous thrombosis of left internal jugular vein
Condition
SNOMED
762418
Acute deep venous thrombosis of left lower extremity
Condition
SNOMED
761462
Acute deep venous thrombosis of left pelvic vein
Condition
SNOMED
618482
Acute deep venous thrombosis of left peroneal vein
Condition
SNOMED
765537
Acute deep venous thrombosis of left upper extremity
Condition
SNOMED
44782767
Acute deep venous thrombosis of lower extremity as complication of procedure
Condition
SNOMED
44782761
Acute deep venous thrombosis of pelvic vein
Condition
SNOMED
762022
Acute deep venous thrombosis of politeal vein of right leg
Condition
SNOMED
44782743
Acute deep venous thrombosis of popliteal vein
Condition
SNOMED
762021
Acute deep venous thrombosis of popliteal vein of left leg
Condition
SNOMED
762010
Acute deep venous thrombosis of right axillary vein
Condition
SNOMED
760877
Acute deep venous thrombosis of right calf
Condition
SNOMED
762013
Acute deep venous thrombosis of right femoral vein
Condition
SNOMED
762018
Acute deep venous thrombosis of right internal jugular vein
Condition
SNOMED
762419
Acute deep venous thrombosis of right lower extremity
Condition
SNOMED
765229
Acute deep venous thrombosis of right pelvic vein
Condition
SNOMED
618681
Acute deep venous thrombosis of right peroneal vein
Condition
SNOMED
762005
Acute deep venous thrombosis of right upper extremity
Condition
SNOMED
44782745
Acute deep venous thrombosis of thigh
Condition
SNOMED
44782744
Acute deep venous thrombosis of tibial vein
Condition
SNOMED
762026
Acute deep venous thrombosis of tibial vein of left leg
Condition
SNOMED
765156
Acute deep venous thrombosis of tibial vein of right leg
Condition
SNOMED
44782421
Acute deep venous thrombosis of upper extremity
Condition
SNOMED
44782766
Acute deep venous thrombosis of upper extremity as complication of procedure
Condition
SNOMED
37171353
Acute ischemia of colon due to thrombosis of mesenteric vein
Condition
SNOMED
37170675
Acute ischemia of small intestine due to thrombosis of mesenteric vein
Condition
SNOMED
4120091
Acute massive pulmonary embolism
Condition
SNOMED
45768439
Acute pulmonary embolism
Condition
SNOMED
45768888
Acute pulmonary thromboembolism
Condition
SNOMED
762048
Acute thrombosis of left subclavian vein
Condition
SNOMED
45757410
Acute thrombosis of mesenteric vein
Condition
SNOMED
762049
Acute thrombosis of right subclavian vein
Condition
SNOMED
36712892
Acute thrombosis of splenic vein
Condition
SNOMED
44782762
Acute thrombosis of subclavian vein
Condition
SNOMED
4179911
Axillary vein thrombosis
Condition
SNOMED
37109253
Bilateral acute deep vein thrombosis of femoral veins
Condition
SNOMED
618678
Bilateral acute deep venous thrombosis of peroneal veins
Condition
SNOMED
609003
Bilateral deep femoral vein thrombophlebitis
Condition
SNOMED
3179900
Bilateral deep vein thromboses
Condition
Nebraska Lexicon
40478951
Bilateral deep vein thrombosis of lower extremities
Condition
SNOMED
609002
Bilateral femoral vein thrombophlebitis
Condition
SNOMED
608965
Bilateral iliac vein thrombophlebitis
Condition
SNOMED
1245776
Bilateral popliteal vein thrombophlebitis
Condition
SNOMED
609006
Bilateral tibial vein thrombophlebitis
Condition
SNOMED
44782732
Chronic pulmonary embolism
Condition
SNOMED
45768887
Chronic pulmonary thromboembolism
Condition
SNOMED
45771016
Chronic pulmonary thromboembolism without pulmonary hypertension

Condition
SNOMED
4042396
Deep thrombophlebitis
Condition
SNOMED
4046884
Deep vein thrombosis of leg related to air travel
Condition
SNOMED
3655221
Deep vein thrombosis of lower extremity due to intravenous drug use
Condition
SNOMED
4133004
Deep venous thrombosis
Condition
SNOMED
761013
Deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
37163011
Deep venous thrombosis of calf
Condition
SNOMED
45773536
Deep venous thrombosis of femoropopliteal vein
Condition
SNOMED
763942
Deep venous thrombosis of left lower extremity
Condition
SNOMED
1075379
Deep venous thrombosis of left posterior tibial vein
Condition
SNOMED
761980
Deep venous thrombosis of left upper extremity
Condition
SNOMED
443537
Deep venous thrombosis of lower extremity
Condition
SNOMED
4133975
Deep venous thrombosis of pelvic vein
Condition
SNOMED
40480555
Deep venous thrombosis of peroneal vein
Condition
SNOMED
1075377
Deep venous thrombosis of posterior tibial vein
Condition
SNOMED
4322565
Deep venous thrombosis of profunda femoris vein
Condition
SNOMED
763941
Deep venous thrombosis of right lower extremity
Condition
SNOMED
1075378
Deep venous thrombosis of right posterior tibial vein
Condition
SNOMED
761928
Deep venous thrombosis of right upper extremity
Condition
SNOMED
4207899
Deep venous thrombosis of tibial vein
Condition
SNOMED
4028057
Deep venous thrombosis of upper extremity
Condition
SNOMED
193512
Embolism and thrombosis of the renal vein
Condition
SNOMED
435565
Embolism and thrombosis of the vena cava
Condition
SNOMED
4258295
Embolism from thrombosis of vein of distal lower extremity
Condition
SNOMED
40481089
Embolism from thrombosis of vein of lower extremity
Condition
SNOMED
40479840
Embolism from thrombosis of vein of thigh
Condition
SNOMED
4119760
Iliofemoral deep vein thrombosis
Condition
SNOMED
43530934
Induced termination of pregnancy complicated by pulmonary embolism
Condition
SNOMED
762808
Infarction of lung due to embolus
Condition
SNOMED
40480461
Infarction of lung due to iatrogenic pulmonary embolism
Condition
SNOMED
4124856
Inferior mesenteric vein thrombosis
Condition
SNOMED
608964
Left iliac vein thrombophlebitis
Condition
SNOMED
602592
Left peroneal vein thrombophlebitis
Condition
SNOMED
600938
Left subclavian vein thrombophlebitis
Condition
SNOMED
37164448
Lemierre syndrome
Condition
SNOMED
4281689
Phlegmasia alba dolens
Condition
SNOMED
4284538
Phlegmasia cerulea dolens
Condition
SNOMED
3185768
Popliteal vein thrombosis
Condition
Nebraska Lexicon
4309333
Postoperative deep vein thrombosis
Condition
SNOMED
4108681
Postoperative pulmonary embolus
Condition
SNOMED
37160752
Postoperative pulmonary thromboembolism
Condition
SNOMED
1245858
Postpartum acute deep vein thrombosis
Condition
SNOMED
46285905
Provoked deep vein thrombosis
Condition
SNOMED
1244882
Pulmonary artery embolism due to foreign body
Condition
SNOMED
440417
Pulmonary embolism
Condition
SNOMED
37109911
Pulmonary embolism due to and following acute myocardial infarction
Condition
SNOMED
3655209
Pulmonary embolism due to and following ectopic pregnancy
Condition
SNOMED
3655210
Pulmonary embolism due to and following molar pregnancy
Condition
SNOMED
37016922
Pulmonary embolism on long-term anticoagulation therapy
Condition
SNOMED
43530605
Pulmonary embolism with pulmonary infarction
Condition
SNOMED
4253796
Pulmonary microemboli
Condition
SNOMED
4121618
Pulmonary thromboembolism
Condition
SNOMED
4236271
Recurrent pulmonary embolism
Condition
SNOMED
608963
Right iliac vein thrombophlebitis
Condition
SNOMED
602583
Right peroneal vein thrombophlebitis
Condition
SNOMED
600939
Right subclavian vein thrombophlebitis
Condition
SNOMED
36713113
Saddle embolus of pulmonary artery
Condition
SNOMED
35615055
Saddle embolus of pulmonary artery with acute cor pulmonale
Condition
SNOMED
40479606
Septic pulmonary embolism
Condition
SNOMED
4033521
Splenic vein thrombosis
Condition
SNOMED
4119607
Subacute massive pulmonary embolism
Condition
SNOMED
4055089
Superior mesenteric vein thrombosis
Condition
SNOMED
4230403
Thrombophlebitis of axillary vein
Condition
SNOMED
4069561
Thrombophlebitis of deep femoral vein
Condition
SNOMED
761831
Thrombophlebitis of deep vein of bilateral lower limbs
Condition
SNOMED
761830
Thrombophlebitis of deep vein of left lower limb
Condition
SNOMED
761808
Thrombophlebitis of deep vein of left upper limb
Condition
SNOMED
761832
Thrombophlebitis of deep vein of right lower limb
Condition
SNOMED
761809
Thrombophlebitis of deep vein of right upper limb
Condition
SNOMED
4221821
Thrombophlebitis of deep veins of lower extremity
Condition
SNOMED
440750
Thrombophlebitis of deep veins of upper extremities
Condition
SNOMED
4203618
Thrombophlebitis of femoropopliteal vein
Condition
SNOMED
4176614
Thrombophlebitis of iliac vein
Condition
SNOMED
764715
Thrombophlebitis of internal jugular vein
Condition
SNOMED
608904
Thrombophlebitis of left axillary vein
Condition
SNOMED
761821
Thrombophlebitis of left deep femoral vein
Condition
SNOMED
761819
Thrombophlebitis of left femoral vein
Condition
SNOMED
609000
Thrombophlebitis of left popliteal vein
Condition
SNOMED
609005
Thrombophlebitis of left tibial vein
Condition
SNOMED
4318407
Thrombophlebitis of mesenteric vein
Condition
SNOMED
608903
Thrombophlebitis of right axillary vein
Condition
SNOMED
761820
Thrombophlebitis of right deep femoral vein
Condition
SNOMED
761818
Thrombophlebitis of right femoral vein
Condition
SNOMED
609001
Thrombophlebitis of right popliteal vein
Condition
SNOMED
609004
Thrombophlebitis of right tibial vein
Condition
SNOMED
4205652
Thrombophlebitis of subclavian vein
Condition
SNOMED
4110339
Thrombophlebitis of the anterior tibial vein
Condition
SNOMED
4111868
Thrombophlebitis of the common iliac vein
Condition
SNOMED
4110343
Thrombophlebitis of the external iliac vein
Condition
SNOMED
439314
Thrombophlebitis of the femoral vein
Condition
SNOMED
4109877
Thrombophlebitis of the internal iliac vein
Condition
SNOMED
4112171
Thrombophlebitis of the popliteal vein
Condition
SNOMED
4112172
Thrombophlebitis of the posterior tibial vein
Condition
SNOMED
4250765
Thrombophlebitis of tibial vein
Condition
SNOMED
42538533
Thrombosis of iliac vein
Condition
SNOMED
44811347
Thrombosis of internal jugular vein
Condition
SNOMED
765049
Thrombosis of left peroneal vein
Condition
SNOMED
4317289
Thrombosis of mesenteric vein
Condition
SNOMED
4203836
Thrombosis of subclavian vein
Condition
SNOMED
4175649
Thrombosis of the popliteal vein
Condition
SNOMED
4153353
Traumatic thrombosis of axillary vein
Condition
SNOMED
46285904
Unprovoked deep vein thrombosis
Condition
SNOMED
37163265
Venous thromboembolism due to thrombosis of vein of lower limb
Condition
SNOMED

Table S4. List of concepts used to define pelvic vein thrombosis (PVT) (concept sets included all descendants of listed concepts).
Concept ID
Concept Name
Domain
Vocabulary
762148
Acute deep vein thrombosis of bilateral iliac veins
Condition
SNOMED
35616028
Acute deep vein thrombosis of left iliac vein
Condition
SNOMED
35616027
Acute deep vein thrombosis of right iliac vein
Condition
SNOMED
765155
Acute deep venous thrombosis of bilateral ileofemoral veins
Condition
SNOMED
761461
Acute deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
762015
Acute deep venous thrombosis of ileofemoral vein of left leg
Condition
SNOMED
765541
Acute deep venous thrombosis of ileofemoral vein of right lower extremity
Condition
SNOMED
761462
Acute deep venous thrombosis of left pelvic vein
Condition
SNOMED
44782761
Acute deep venous thrombosis of pelvic vein
Condition
SNOMED
765229
Acute deep venous thrombosis of right pelvic vein
Condition
SNOMED
608965
Bilateral iliac vein thrombophlebitis
Condition
SNOMED
765152
Chronic deep vein thrombosis of bilateral iliac veins
Condition
SNOMED
35616026
Chronic deep vein thrombosis of left iliac vein
Condition
SNOMED
761439
Chronic deep vein thrombosis of left pelvic vein
Condition
SNOMED
46271548
Chronic deep vein thrombosis of pelvic vein
Condition
SNOMED
35616025
Chronic deep vein thrombosis of right iliac vein
Condition
SNOMED
761441
Chronic deep vein thrombosis of right pelvic vein
Condition
SNOMED
765542
Chronic deep venous thrombosis of bilateral ileofemoral veins
Condition
SNOMED
761440
Chronic deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
765543
Chronic deep venous thrombosis of left ileofemoral vein
Condition
SNOMED
762016
Chronic deep venous thrombosis of right ileofemoral vein
Condition
SNOMED
761013
Deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
4133975
Deep venous thrombosis of pelvic vein
Condition
SNOMED
608964
Left iliac vein thrombophlebitis
Condition
SNOMED
4285751
Pelvic thrombophlebitis in puerperium
Condition
SNOMED
608963
Right iliac vein thrombophlebitis
Condition
SNOMED
4176614
Thrombophlebitis of iliac vein
Condition
SNOMED
4317290
Thrombophlebitis of pelvic vein
Condition
SNOMED
4111868
Thrombophlebitis of the common iliac vein
Condition
SNOMED
4110343
Thrombophlebitis of the external iliac vein
Condition
SNOMED
4109877
Thrombophlebitis of the internal iliac vein
Condition
SNOMED
42538533
Thrombosis of iliac vein
Condition
SNOMED
4319327
Thrombosis of pelvic vein
Condition
SNOMED

Table S5. List of concepts used to splanchnic vein thrombosis (SVT).
Concept ID
Concept Name
Domain
Vocabulary
37171353
Acute ischemia of colon due to thrombosis of mesenteric vein
Condition
SNOMED
37170675
Acute ischemia of small intestine due to thrombosis of mesenteric vein
Condition
SNOMED
45757410
Acute thrombosis of mesenteric vein
Condition
SNOMED
36712892
Acute thrombosis of splenic vein
Condition
SNOMED
196715
Budd-Chiari syndrome
Condition
SNOMED
4301208
Hepatic vein thrombosis
Condition
SNOMED
4124856
Inferior mesenteric vein thrombosis
Condition
SNOMED
4092406
Portal thrombophlebitis
Condition
SNOMED
199837
Portal vein thrombosis
Condition
SNOMED
4033521
Splenic vein thrombosis
Condition
SNOMED
4055089
Superior mesenteric vein thrombosis
Condition
SNOMED
4318407
Thrombophlebitis of mesenteric vein
Condition
SNOMED
4317289
Thrombosis of mesenteric vein
Condition
SNOMED

Table S6. List of concepts used to define retinal vein thrombosis (RVT).
Concept ID
Concept Name
Domain
Vocabulary
437544
Arterial retinal branch occlusion
Condition
SNOMED
3657106
Bilateral occlusion of branch retinal arteries
Condition
SNOMED
37310623
Bilateral occlusion of central retinal arteries
Condition
SNOMED
37169454
Bilateral vascular occlusion of retina of eyes
Condition
SNOMED
4336004
Branch macular artery occlusion
Condition
SNOMED
4339013
Branch retinal vein occlusion with macular edema
Condition
SNOMED
4334248
Branch retinal vein occlusion with neovascularization
Condition
SNOMED
4199035
Branch retinal vein occlusion with no neovascularization
Condition
SNOMED
437540
Central retinal artery occlusion
Condition
SNOMED
313761
Central retinal vein occlusion
Condition
SNOMED
4208221
Central retinal vein occlusion - ischemic
Condition
SNOMED
4208222
Central retinal vein occlusion - non-ischemic
Condition
SNOMED
4339010
Central retinal vein occlusion with macular edema
Condition
SNOMED
4334246
Central retinal vein occlusion with neovascularization
Condition
SNOMED
4338905
Cilioretinal artery occlusion
Condition
SNOMED
42535735
Combined occlusion by thrombus of retinal artery and retinal vein
Condition
SNOMED
4102317
Incipient occlusion of retinal vein
Condition
SNOMED
4083482
Macular branch retinal vein occlusion
Condition
SNOMED
37206377
Occlusion of branch of retinal vein of left eye
Condition
SNOMED
37206378
Occlusion of branch of retinal vein of right eye
Condition
SNOMED
37206381
Occlusion of central retinal vein of left eye
Condition
SNOMED
37206380
Occlusion of central retinal vein of right eye
Condition
SNOMED
36713329
Occlusion of left branch retinal artery
Condition
SNOMED
37207955
Occlusion of left central retinal artery
Condition
SNOMED
3657873
Occlusion of left cilioretinal artery
Condition
SNOMED
36713330
Occlusion of right branch retinal artery
Condition
SNOMED
37207895
Occlusion of right central retinal artery
Condition
SNOMED
3657872
Occlusion of right cilioretinal artery
Condition
SNOMED
4334245
Retinal artery occlusion
Condition
SNOMED
4324290
Retinal phlebitis
Condition
SNOMED
440392
Retinal vascular occlusion
Condition
SNOMED
3183076
Right branch retinal artery occlusion
Condition
Nebraska Lexicon
4216561
Thrombophlebitis of retinal vein
Condition
SNOMED
4187790
Thrombosis of retinal vein
Condition
SNOMED
3657847
Vascular occlusion of retina of left eye
Condition
SNOMED
3657848
Vascular occlusion of retina of right eye
Condition
SNOMED
312622
Venous retinal branch occlusion
Condition
SNOMED

Table S7. List of concepts used to define disseminated intravascular coagulation (DIC).
Concept ID
Concept Name
Domain
Vocabulary
37117819
Acquired purpura fulminans
Condition
SNOMED
436093
Disseminated intravascular coagulation
Condition
SNOMED
4028488
Purpura fulminans
Condition
SNOMED

ANNEX V. ENCePP checklist for study protocols
ENCePP Checklist for Study Protocols (Revision 4)
Doc.Ref. EMA/540136/2009
Adopted by the ENCePP Steering Group on 15/10/2018
Study title: DARWIN EU® - Time to onset of thromboembolic events in adults with selected types of cancer


EU PAS Register® number: Study not registered yet
Study reference number (if applicable): P4-C2-017

Section 1: Milestones
Yes
No
N/A
Section Number
1.1 Does the protocol specify timelines for




1.1.1 Start of data collection1
X


8.5
1.1.2 End of data collection2
X


8.5
1.1.3 Progress report(s)
X


5
1.1.4 Interim report(s)
X


5
1.1.5 Registration in the EU PAS Register®
X


5
1.1.6 Final report of study results.
X


5
Comments:


Section 2: Research question
Yes
No
N/A
Section Number
2.1 Does the formulation of the research question and objectives clearly explain:




2.1.1 Why the study is conducted? (e.g. to address an important public health concern, a risk identified in the risk management plan, an emerging safety issue)
X


6
2.1.2 The objective(s) of the study?
X


7
2.1.3 The target population? (i.e. population or subgroup to whom the study results are intended to be generalised)
X


8.3
2.1.4 Which hypothesis(-es) is (are) to be tested?


X

2.1.5 If applicable, that there is no a priori hypothesis?


X

Comments:


Section 3: Study design
Yes
No
N/A
Section Number
3.1 Is the study design described? (e.g. cohort, case-control, cross-sectional, other design)
X


8.1
3.2 Does the protocol specify whether the study is based on primary, secondary or combined data collection?
X


8.4
3.3 Does the protocol specify measures of occurrence? (e.g., rate, risk, prevalence)
X


8.8.3
3.4 Does the protocol specify measure(s) of association? (e.g. risk, odds ratio, excess risk, rate ratio, hazard ratio, risk/rate difference, number needed to harm (NNH))


X

3.5 Does the protocol describe the approach for the collection and reporting of adverse events/adverse reactions? (e.g. adverse events that will not be collected in case of primary data collection)


X

Comments:


Section 4: Source and study populations
Yes
No
N/A
Section Number
4.1 Is the source population described?
X


Annex I
4.2 Is the planned study population defined in terms of:




4.2.1 Study time period
X


8.5
4.2.2 Age and sex
X


8.3
4.2.3 Country of origin
X


8.4
4.2.4 Disease/indication
X


8.3
4.2.5 Duration of follow-up
X


8.2
4.3 Does the protocol define how the study population will be sampled from the source population? (e.g. event or inclusion/exclusion criteria)
X


8.3
Comments:


Section 5: Exposure definition and measurement
Yes
No
N/A
Section Number
5.1 Does the protocol describe how the study exposure is defined and measured? (e.g. operational details for defining and categorising exposure, measurement of dose and duration of drug exposure)


X

5.2 Does the protocol address the validity of the exposure measurement? (e.g. precision, accuracy, use of validation sub-study)


X

5.3 Is exposure categorised according to time windows?


X

5.4 Is intensity of exposure addressed?
(e.g. dose, duration)


X

5.5 Is exposure categorised based on biological mechanism of action and taking into account the pharmacokinetics and pharmacodynamics of the drug?


X

5.6 Is (are) (an) appropriate comparator(s) identified?


X

Comments:


Section 6: Outcome definition and measurement
Yes
No
N/A
Section Number
6.1 Does the protocol specify the primary and secondary (if applicable) outcome(s) to be investigated?
X


8.6.2
6.2 Does the protocol describe how the outcomes are defined and measured?
X


8.6.2
6.3 Does the protocol address the validity of outcome measurement? (e.g. precision, accuracy, sensitivity, specificity, positive predictive value, use of validation sub-study)
X


9
6.4 Does the protocol describe specific outcomes relevant for Health Technology Assessment? (e.g. HRQoL, QALYs, DALYS, health care services utilisation, burden of disease or treatment, compliance, disease management)


X

Comments:


Section 7: Bias
Yes
No
N/A
Section Number
7.1 Does the protocol address ways to measure confounding? (e.g. confounding by indication)


X

7.2 Does the protocol address selection bias? (e.g. healthy user/adherer bias)
X


9
7.3 Does the protocol address information bias? (e.g. misclassification of exposure and outcomes, time-related bias)
X


9
Comments:


Section 8: Effect measure modification
Yes
No
N/A
Section Number
8.1 Does the protocol address effect modifiers? (e.g. collection of data on known effect modifiers, sub-group analyses, anticipated direction of effect)


X

Comments:


Section 9: Data sources
Yes
No
N/A
Section Number
9.1 Does the protocol describe the data source(s) used in the study for the ascertainment of:




9.1.1 Exposure? (e.g. pharmacy dispensing, general practice prescribing, claims data, self-report, face-to-face interview)


X

9.1.2 Outcomes? (e.g. clinical records, laboratory markers or values, claims data, self-report, patient interview including scales and questionnaires, vital statistics)
X


Annex I
9.1.3 Covariates and other characteristics?
X


Annex I
9.2 Does the protocol describe the information available from the data source(s) on:




9.2.1 Exposure? (e.g. date of dispensing, drug quantity, dose, number of days of supply prescription, daily dosage, prescriber)


X

9.2.2 Outcomes? (e.g. date of occurrence, multiple event, severity measures related to event)
X


Annex I
9.2.3 Covariates and other characteristics? (e.g. age, sex, clinical and drug use history, co-morbidity, co-medications, lifestyle)
X


Annex I
9.3 Is a coding system described for:




9.3.1 Exposure? (e.g. WHO Drug Dictionary, Anatomical Therapeutic Chemical (ATC) Classification System)


X

9.3.2 Outcomes? (e.g. International Classification of Diseases (ICD), Medical Dictionary for Regulatory Activities (MedDRA))
X


Annex I
9.3.3 Covariates and other characteristics?
X


Annex I
9.4 Is a linkage method between data sources described? (e.g. based on a unique identifier or other)


X

Comments:


Section 10: Analysis plan
Yes
No
N/A
Section Number
10.1 Are the statistical methods and the reason for their choice described?
X


8.8.3
10.2 Is study size and/or statistical precision estimated?


X

10.3 Are descriptive analyses included?
X


8.8.3
10.4 Are stratified analyses included?
X


8.8.3
10.5 Does the plan describe methods for analytic control of confounding?


X

10.6 Does the plan describe methods for analytic control of outcome misclassification?


X

10.7 Does the plan describe methods for handling missing data?
X


8.8.3
10.8 Are relevant sensitivity analyses described?


X

Comments:


Section 11: Data management and quality control
Yes
No
N/A
Section Number
11.1 Does the protocol provide information on data storage? (e.g. software and IT environment, database maintenance and anti-fraud protection, archiving)
X


Annex III
11.2 Are methods of quality assurance described?
X


Annex III
11.3 Is there a system in place for independent review of study results?
X


Annex III
Comments:


Section 12: Limitations
Yes
No
N/A
Section Number
12.1 Does the protocol discuss the impact on the study results of:




12.1.1 Selection bias?
X


9
12.1.2 Information bias?
X


9
12.1.3 Residual/unmeasured confounding?
(e.g. anticipated direction and magnitude of such biases, validation sub-study, use of validation and external data, analytical methods).
X


9
12.2 Does the protocol discuss study feasibility? (e.g. study size, anticipated exposure uptake, duration of follow-up in a cohort study, patient recruitment, precision of the estimates)
X


8.7
Comments:


Section 13: Ethical/data protection issues
Yes
No
N/A
Section Number
13.1 Have requirements of Ethics Committee/ Institutional Review Board been described?
X


Annex III
13.2 Has any outcome of an ethical review procedure been addressed?


X

13.3 Have data protection requirements been described?
X


Annex III
Comments:


Section 14: Amendments and deviations
Yes
No
N/A
Section Number
14.1 Does the protocol include a section to document amendments and deviations?
X


4
Comments:


Section 15: Plans for communication of study results
Yes
No
N/A
Section Number
15.1 Are plans described for communicating study results (e.g. to regulatory authorities)?
X


Annex III
15.2 Are plans described for disseminating study results externally, including publication?
X


Annex III
Comments:



ANNEX VI. Glossary
Additional definitions are available in the EMA Glossary of terms https://www.ema.europa.eu/en/about-us/glossaries.
Aggregated Data
Data collected and combined from multiple sources to generate summary information, typically anonymised.
Benefit-Risk Assessment
Evaluation of the positive therapeutic effects of a medicine compared to its risks (e.g., side effects).
Common Data Model (CDM)
A standardized data structure that enables data from multiple sources to be harmonized, making analysis consistent and reproducible. DARWIN EU® utilises the OMOP CDM maintained by the OHDSI community.
Complex Studies (C3)
Studies requiring the development or customisation of specific study designs, protocols, and Statistical Analysis Plans (SAPs), with extensive collection or extraction of data. Examples include etiological studies measuring the strength and determinants of an association between an exposure and the occurrence of a health outcome in a defined population considering sources of bias, potential confounding factors, and effect modifiers.
Coordination Centre (CC)
The central hub responsible for managing and overseeing the activities within DARWIN EU®. It is based at Erasmus University Medical Centre in Rotterdam, the Netherlands.
Data Access
The process of obtaining permission to use specific datasets for regulatory or scientific studies.
Data Quality Framework
A set of standards and procedures to ensure accuracy, completeness, timeliness, and consistency of data used in DARWIN EU®.
Data Source
A database or repository of structured health-related data, such as electronic health records (EHRs), insurance claims, or registries.
DARWIN EU®
The European Medicines Agency's (EMA) federated network of real-world data sources designed to generate evidence to support regulatory decision-making.
EMA (European Medicines Agency)
The regulatory body responsible for the evaluation and supervision of medicinal products in the EU, overseeing DARWIN EU®.
Evidence Generation
The process of analysing real-world data to produce scientific information that can inform healthcare or regulatory decisions.
Federated Network
A data infrastructure where data remain at their original location but can be analysed in a harmonised way across multiple partners using a common model and tools.
GDPR (General Data Protection Regulation)
The EU regulation governing the protection of personal data and privacy, crucial to how DARWIN EU® handles health data.
Health Technology Assessment (HTA)
A systematic evaluation of properties and impacts of health technology, often using DARWIN EU® data to support assessments.
Metadata
Descriptive information about a data source (e.g., its content, quality, and structure), essential for identifying relevant databases in DARWIN EU® studies.
Off-the-Shelf Studies (OTS)
Studies for which a standard protocol per study/analysis type and standardised analytics may be developed and applied or adapted, typically relating to a descriptive research question. This includes studies on disease epidemiology, for example, the estimation of the prevalence or incidence of health outcomes in defined time periods and population groups, or drug utilisation studies at the population or individual level.
OHDSI (Observational Health Data Sciences and Informatics)
An open-science collaborative community that develops tools and standards (including the OMOP CDM) to enable large-scale analytics of observational health data. OHDSI provides the technical and scientific foundation for DARWIN EU®’s analytical ecosystem.
Patient-Level Data
Data related to individuals, de-identified, used for longitudinal or detailed analyses.
OMOP (Observational Medical Outcomes Partnership)
A common data model (CDM) that standardises the structure and content of observational healthcare data, enabling systematic analysis across disparate datasets. DARWIN EU® uses the OMOP CDM to ensure interoperability and consistency in real-world evidence generation.
Real-World Data (RWD)
Data relating to individual health status or healthcare delivery that is collected from routine clinical practice rather than from randomised controlled trials.
Real-World Evidence (RWE)
Clinical evidence derived from the analysis of RWD, used to inform decisions by regulators, payers, or clinicians.
Regulatory Decision-Making
The process by which authorities like EMA assess data to authorise, monitor, or modify the use of medicines in the EU.
Routine Repeated Studies (RR)
Studies that are either Off-the-Shelf or Complex studies repeated on a regular basis, following the same protocol and study code, but with updated data and/or different data partners.
Study Protocol
A detailed plan describing how a specific real-world study will be conducted, including objectives, design, data sources, and analyses.
Very Complex Studies (C4)
Studies which cannot rely only on electronic health care databases, or which would require complex methodological work, for example, due to the occurrence of events that cannot be defined by existing diagnosis codes, including events that do not yet have a diagnosis code, where it may be necessary to combine a diagnosis code with other data such as results of laboratory investigations. These studies might require the collection of data prospectively, or the inclusion of new (not previously onboarded) data sources.
Name of the main author of the protocol:
Melissa Leung
Date: 10/10/2025

Signature:
M. Leung




Study Protocol
P4-C2-017
DARWIN EU® - Time to onset of thromboembolic events in adults with selected types of cancer


10/10/2025
Version 1.0
Authors: Melissa Leung, Cesar Barboza, Ionna Nika, Anton Barchuk, Talita Duarte-Salles
Confidential


CONTENTS
LIST OF ABBREVIATIONS 5
1. TITLE 7
2. DESCRIPTION OF THE STUDY TEAM 7
3. ABSTRACT 8
4. AMENDMENTS AND UPDATES 11
5. MILESTONES 11
6. RATIONALE AND BACKGROUND 11
7. RESEARCH QUESTION AND OBJECTIVES 11
8. RESEARCH METHODS 11
8.1. Study design 11
Figure 1. Graphical depiction of the study design. 12
8.2. Follow-up 12
8.3. Study population with inclusion and exclusion criteria 12
8.4. Study setting and data sources 13
Table 1. Data sources. 13
8.5. Study period 14
8.6. Variables 14
8.6.1. Exposure 14
8.6.2. Outcome 14
8.6.3. Intercurrent events (only for causal studies) 14
8.6.4. Covariates, including confounders, effect modifiers, and other variables 14
8.7. Study size 15
8.8. Analysis 15
8.8.1. Federated network analyses 15
8.8.2. Data privacy protection 15
8.8.3. Statistical model specification and assumptions of the analytical approach considered 15
8.8.4. Output 16
Table 1. Attrition of study participants. 17
Figure 1. Cumulative probability of not having thromboembolic event after the first cancer diagnosis accounting for a competing risk of death. 19
Table 2. Median time in days (95% CI) to thromboembolic event after first cancer diagnosis. 20
9. STRENGTHS AND LIMITATIONS 20
10. REFERENCES 21
11. ANNEXES 22
ANNEX I. Description of data sources 22
ANNEX II. Fitness for use assessment 33
ANNEX III. Operational and reporting considerations 37
ANNEX IV. List of stand-alone documents 39
Table S1. List of concepts used to define deep vein thrombosis (DVT). 39
Table S2. List of concepts used to define pulmonary embolism (PE). 43
Table S3. List of concepts used to define venous thromboembolism (VTE). 43
Table S4. List of concepts used to define pelvic vein thrombosis (PVT) (concept sets included all descendants of listed concepts). 48
Table S5. List of concepts used to splanchnic vein thrombosis (SVT). 49
Table S6. List of concepts used to define retinal vein thrombosis (RVT). 49
Table S7. List of concepts used to define disseminated intravascular coagulation (DIC). 50
ANNEX V. ENCePP checklist for study protocols 52
ANNEX VI. Glossary 58

Study title
DARWIN EU® - Time to onset of thromboembolic events in adults with selected types of cancer
Protocol version
V1.0
Date
10/10/2025
EUPAS number
Study not registered yet
Active substance
None
Medicinal product
None
Research question and objectives
The aim of this study is to estimate time to onset of venous thromboembolic events in adults with each type of selected cancer.
The specific objectives of the study are:
1. To estimate the probability of not having thromboembolic events at 6-month intervals within 5 years in adults with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
2. To estimate median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
Countries of study
Belgium, Denmark, Estonia, Finland, Germany, The Netherlands, Spain, United Kingdom
Authors
Melissa Leung ([email protected])
Cesar Barboza ([email protected])
Ionna Nika ([email protected])
Anton Barchuk ([email protected])
Talita Duarte-Salles ([email protected])
This is a routinely repeated study from P3-C3-005 with EUPAS1000000440 (https://catalogues.ema.europa.eu/node/4341).
LIST OF ABBREVIATIONS
Acronyms/terms
Description
ADHD
Attention deficit hyperactivity disorder
AJCC/UICC
American Joint Committee on Cancer and the International Union Against Cancer
ATC
Anatomical Therapeutic Chemical
CDM 
Common Data Model 
CI
Confidence interval
CPRD 
Clinical Practice Research Datalink 
DARWIN EU® 
Data Analysis and Real World Interrogation Network 
DK-DHR
Danish Data Health Registries
DOI 
Declaration Of Interests 
DQD 
Data Quality Dashboard 
DRE
Digital Research Environment
DVT
Deep Venous Thrombosis
DIC
Disseminated Intravascular Coagulation
EHR 
Electronic Health Record 
EMA 
European Medicines Agency 
EBB 
Estonian Biobank 
EGCUT 
Estonian Genome Center at the University of Tartu 
ENCePP 
European Network of Centres for Pharmacoepidemiology and Pharmacovigilance 
EU
European Union
EUPAS
EU Post-Authorisation Studies Register
GDPR
General Data Protection Regulation
GP 
General Practitioner 
HIV
Human Immunodeficiency Virus
ICD-O-3
International Classification of Diseases for Oncology, 3rd Edition
ICD-10
International Classification of Diseases, 10th revision
ICPC-1
International Classification of Primary Care
IP
Inpatient
IPCI 
Integrated Primary Care Information Project 
IR
Incidence rate
IRB
Institutional Review Board
LPD 
Longitudinal Patient Database 
OHDSI
Observational Health Data Sciences and Informatics
OMOP 
Observational Medical Outcomes Partnership 
OP
Outpatient
PE
Pulmonary Embolism
PVT
Pelvic Venous Thrombosis
PY
Person-years
RVT
Retinal vein thrombosis
SNOMED 
Systematized Nomenclature of Medicine 
SVT
Splanchnic Vein Thrombosis
UKBB
UK Biobank
VTE
Venous Thromboembolism

1. TITLE
DARWIN EU® - Time to onset of thromboembolic events in adults with selected types of cancer
2. DESCRIPTION OF THE STUDY TEAM
Study team role
Names
Organisation
Principal Investigator
Melissa Leung
Anton Barchuk
Talita Duarte-Salles
Erasmus MC
Data Scientist
Cesar Barboza
Ioanna Nika
Erasmus MC
Clinical Domain Expert
Anton Barchuk
Erasmus MC
Study Manager
Natasha Yefimenko
Erasmus MC
Data source
Names
Data Partner Organisation*
IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium)
IQVIA Disease Analyzer Germany (IQVIA DA Germany)
Gargi Jadhav   
Isabella Kacmarczyl  
Akram Mendez   
Hanne van Ballegooijen  
Dina Vojinovic  
IQVIA
Danish Data Health Registries (DK-DHR)
Elvira Bräuner 
Susanne Bruun 
Danish Medicines Agency
Estonian Biobank (EBB)
Marek Oja
Raivo Kolde
Ami Sild
Estonian Biobank, Estonia
Finnish Care Register for Health Care (FinOMOP-THL)
Anna Hammais
Gustav Klingstedt
Finnish Care Register for Health Care, Finland
Integrated Primary Care Information (IPCI)
Katia Verhamme
Integrated Primary Care Information, Netherlands
The Information System for the Development of Research in Primary Care (SIDIAP)
Anna Palomar-Cros
Irene López-Sánchez
Agustina Giuliodori
IDIAPJGol
Clinical Practice Research Datalink GOLD (CPRD GOLD) and UK BioBank (UKBB)
Antonella Delmestri
University of Oxford 
*Data partners do not have an investigator role. Data partners execute code at their data source, review, and approve their results.
3. ABSTRACT
Title
DARWIN EU® - Time to onset of thromboembolic events in adults with selected types of cancer
Rationale and background
Thromboembolic events are a common complication for individuals with cancer, with risk varying according to the cancer site, suggesting cancer-specific mechanisms playing a role in the occurrence of these events. Haematological malignancies and lung, pancreas, stomach, bowel, and brain cancers are generally associated with a high risk of clot formation, whilst prostate and breast cancers are associated with low risk of thrombosis.
When a safety signal of a thromboembolic event appears in cancer populations, it can be challenging to assess a potential association with the oncologic treatment without reliable information on the background risk. This study is intended to address this knowledge gap by generating evidence on the time to onset of different venous thromboembolic events among adults with selected cancer types.
Research question and objectives
Research question
What was the time to onset of venous thromboembolic events in adults newly diagnosed with each type of selected cancer during the period 2016–2022?
Objectives
The aim of this study is to estimate time to onset of venous thromboembolic events in adults with each type of selected cancer.
The specific objectives of the study are:
1. To estimate the probability of not having thromboembolic events at 6-month intervals within 5 years in adults with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
2. To estimate median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
Methods
Study design
Population-based cohort study. The index date, i.e., date of cohort entry, will be the date of the first cancer diagnosis. Individuals are followed up until the earliest of occurrence of the outcome, loss to follow-up, end of data availability, end of the study period, or death.
Population
The study population will be the population that was included in the study EUPAS1000000440, of which this is a routinely repeated study. This study population will include all individuals aged 18 years and above with a primary diagnosis of one of the selected cancers (bone, brain, breast, colorectal, corpus uteri, kidney, leukaemia and lymphoma, liver, lung, melanoma, oesophageal, ovary, pancreas, prostate, stomach) during the inclusion period (from 01/01/2016 to 31/12/2022). Only individuals with an incident cancer diagnosis (excluding non-melanoma skin cancer), defined as a first cancer diagnosis after ≥365 days cancer-free history, will be included. Cancer cases and thromboembolic events will be identified based on appropriate computable phenotyping algorithms. Conditions in the OMOP CDM use the Systematised Nomenclature of Medicine (SNOMED) as the standard vocabulary for diagnosis codes. The International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) will also be considered for cancer diagnoses. Other eligibility criteria will include at least 365 days of database history prior to index date and at least 365 days between index date and end of data availability in the data source.
Variables
Exposure:
Not applicable.
Outcome:
The outcomes will include thromboembolic events, specifically: deep vein thrombosis (DVT), pulmonary embolism (PE), venous thromboembolism (VTE, composite of DVT and PE), pelvic venous thrombosis (PVT), splanchnic vein thrombosis (SVT, including hepatic and extra-hepatic vein thrombosis), retinal vein thrombosis (RVT, including retinal central vein thrombosis), and disseminated intravascular coagulation (DIC).
Relevant covariates:
The following covariates will be assessed at index date: age group in years (18–34, 35–44, 45–54, 55–64, 65–74, 75–84, and ≥85), sex, and study subperiod (2016–2019 and 2020–2022). These variables will be used to stratify the results.
Data sources
1. Belgium: IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium)
2. Denmark: Danish Data Health Registries (DK-DHR)
3. Estonia: Estonian Biobank (EBB)
4. Finland: Finnish Care Register for Health Care (FinOMOP-THL)
5. Germany: IQVIA Disease Analyzer Germany (IQVIA DA Germany)
6. Netherlands: Integrated Primary Care Information (IPCI)
7. Spain: The Information System for Research on Primary Care (SIDIAP)
8. United Kingdom: Clinical Practice Research Datalink GOLD (CPRD GOLD)
9. United Kingdom: UK BioBank (UKBB)
Study size
No sample size will be calculated, as this is an exploratory study which will not test a specific hypothesis. Based on the results of the study EUPAS1000000440, the expected number of person counts will be the lowest for DIC (during 1-year follow-up: 5 in FinOMOP-THL – 171 in SIDIAP, with 0 counts in CPRD GOLD, EBB, IPCI, IQVIA DA Germany, and IQVIA LPD Belgium) and highest for VTE (during 1-year follow-up: 27 in IQVIA LPD Belgium – 4,597 in FinOMOP-THL).
Statistical analysis
Analyses will be conducted separately for each data source and carried out in a federated manner, allowing analyses to be run locally without sharing individual-level data.
Objective 1
The probabilities of not having thromboembolic events at 6-month intervals within 5 years in adults with each type of selected cancer will be assessed using the R package CohortSurvival, accounting for a competing risk of death.
Objective 2
The median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer will be assessed using the R package CohortSurvival.
The R package CohortSurvival is designed to work with data in the OMOP CDM format to extract and summarise survival data applying the Kaplan-Meier method. The analyses will be conducted for the overall cohorts as well as by strata of age group, sex, and study subperiod.
Absence of diagnosis codes will be interpreted as a lack of the conditions themselves. A minimum cell count of 5 will be used when reporting results, with any smaller count reported as “<5” and zero counts as “0”.

4. AMENDMENTS AND UPDATES
None
5. MILESTONES
Study milestones and deliverables
Planned dates
Final Study Protocol
To be confirmed by EMA
Creation of Analytical code
31 October 2025
Execution of Analytical Code on the data
14 November 2025
Draft Study Report
Depending on IRB approvals
Final Study Report
Depending on IRB approvals
*Planned dates are dependent on obtaining approvals from the internal review boards of the data sources.
6. RATIONALE AND BACKGROUND
This study is a routine repeated study of a previous DARWIN EU® study (EUPAS1000000440) focused on estimating the incidence rates of venous thromboembolic events in adults newly diagnosed with any of the selected cancers (bone, brain, breast, colorectal, corpus uteri, kidney, leukaemia and lymphoma, liver, lung, melanoma, oesophageal, ovary, pancreas, prostate, stomach) during the period 2016–2022 and describing the individuals’ characteristics at the time of cancer diagnosis. This study is now being repeated with the same study population and outcome, but a different objective, i.e., to estimate the time to onset of thromboembolic events in adults with selected types of cancer.
7. RESEARCH QUESTION AND OBJECTIVES
Research question
The aim of this study is to estimate time to onset of venous thromboembolic events in adults with each type of selected cancer.
Research objectives
The specific objectives of the study are:
1. To estimate the probability of not having thromboembolic events at 6-month intervals within 5 years in adults with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
2. To estimate median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
8. RESEARCH METHODS
8.1. Study design
A cohort study will be conducted. The study will comprise:
• a characterisation study to address objective 1, assessing the probability of not having thromboembolic events at 6-month intervals within 5 years in adults with each type of selected cancer, overall and stratified by age group, sex, and study subperiod
• a characterisation study to address objective 2, assessing median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer, overall and stratified by age group, sex, and study subperiod.
The study design to address objective 1, including assessment windows, is visualised in Figure 1. For objective 2, we will subset the cohort of adults with cancer with thromboembolic events.

Figure 1. Graphical depiction of the study design.
a. The censor date will be the earliest of occurrence of the outcome, loss to follow-up, end of data availability, or death.
8.2. Follow-up
For both objectives, follow-up in the survival analysis will start on the date of cancer diagnosis (index date) and end on the earliest of occurrence of the outcome (thromboembolic event), loss to follow-up, end of data availability, or death.
8.3. Study population with inclusion and exclusion criteria
Objective 1
Inclusion criteria
• First diagnosis of a selected cancer (index date) between 01/01/2016 and 31/12/2022
• Age ≥18 years at cancer diagnosis
• Minimum 365 days of available history before the cancer diagnosis date
• Cancer diagnosis date ≥365 days prior to end of data availability of the data source
Exclusion criteria
• History of any cancer diagnosis ever before the selected cancer diagnosis date
• Outcome during the year before the cancer diagnosis date
Objective 2
Inclusion criteria
• Included in the study population of objective 1
• Occurrence of the outcome (thromboembolic event) during follow-up.


8.4. Study setting and data sources
This study will be conducted using routinely collected data from different health care settings from 9 data sources in the DARWIN EU® network of data partners from 8 countries across Europe, of which 7 EU member states (Table 1). All data were a priori mapped to the OMOP CDM.
Table 1. Data sources.
Country
Name of Data source
Health Care setting
Type of Data
Number of active individuals
Calendar period covered by each data source
Contributing to
BE
IQVIA LPD Belgium
Primary care
EHRs
189k
2015–2025
All objectives
DK
DK-DHR
All settings
EHRs, registries, claims
5.98M
1995–2025
All objectives
EE
EBB
Primary care, hospital care (IP and OP)
EHRs, claims, registries, biobank
212k
2004–2025
All objectives
FI
FinOMOP-THL
Hospital care (IP and OP)
EHRs, registries
5.7M
2011–2025
All objectives
DE
IQVIA DA Germany
Primary care
EHRs
4.48M
1992–2025
All objectives
NL
IPCI
Primary care
EHRs
1.33M
2006–2025
All objectives
ES
SIDIAP
Primary care
EHRs
5.95M
2006–2025
All objectives
GB
CPRD GOLD
Primary care, hospital care (OP)
EHRs
2.83M
1987–2025
All objectives
GB
UKBB
Primary care (up to 2017), hospital care (IP and OP, up to November 2022)
EHRs, registries, biobank
500k
1940–2025
All objectives
Countries: BE=Belgium, DE=Germany, DK=Denmark, EE=Estonia, ES=Spain, FI=Finland, GB=United Kingdom of Great Britain and Northern Ireland, NL=The Netherlands
Data sources: IQVIA LPD=IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium); DK-DHR=Danish Data Health Registries; EBB=Estonian Biobank; FinOMOP-THL=Finnish Care Register for Health Care; IQVIA DA=IQVIA Disease Analyzer Germany (IQVIA DA Germany); IPCI=Integrated Primary Care Information; SIDIAP=The Information System for Research in Primary Care; CPRD GOLD=Clinical Practice Research Datalink GOLD; UKBB=UK BioBank
Types of data: EHR=electronic health record, IP=inpatient, OP=outpatient
Number of active subjects: k=thousands, M=millions

Data sources selection
These data sources fulfil the criteria required in terms of data quality, completeness, timeliness, and representativeness for the cohort study while covering different regions of Europe (Annex II).



8.5. Study period
The study period is from 01/01/2016 to the most recent data available for each contributing data source.
8.6. Variables
8.6.1. Exposure
None.
8.6.2. Outcome
All objectives
The thromboembolic event outcomes in this study are identical to those in EUPAS1000000440, of which this is a routinely repeated study:
• Deep vein thrombosis (DVT)
• Pulmonary embolism (PE)
• Venous thromboembolism (VTE, composite of DVT and PE)
• Pelvic venous thrombosis (PVT)
• Splanchnic vein thrombosis (SVT), including hepatic and extra-hepatic vein thrombosis
• Retinal vein thrombosis (RVT), including retinal central vein thrombosis
• Disseminated intravascular coagulation (DIC)
Each of the specific thromboembolic events will be a primary, binary outcome. The outcome will be assessed at any diagnosis position in the electronic health record and from any of the care settings in each data source.
The list of concepts, from EUPAS1000000440 and based on SNOMED codes and aligned with previous studies that used OMOP CDM and VTE as an outcome (Burn et al., 2022), is provided in Annex IV.
8.6.3. Intercurrent events (only for causal studies)
Not applicable.
8.6.4. Covariates, including confounders, effect modifiers, and other variables
All Objectives
• Sex
◦ Female/male
• Age groups (years) at cancer diagnosis date:
◦ 18–34
◦ 35–44
◦ 45–54
◦ 55–64
◦ 65–74
◦ 75–84
◦ ≥85
Age at cancer diagnosis date will be calculated using January 1st of the year of birth as proxy for the actual birthday. Date/month is either not present or cannot be made available for governance reasons. If available, date is often set to first of the month for personal privacy.
• Study subperiod:
◦ 2016–2019
◦ 2020–2022
The status of each covariate will be assessed at the cancer diagnosis date. Each covariate will be used to stratify the results.
8.7. Study size
No sample size will be calculated, as this is a characterisation study which will not test a specific hypothesis. In addition, we will use data from the study population that was included in EUPAS1000000440, of which this is a routinely repeated study. Thus, the sample size is driven by the availability of data for adults with both cancer and a thromboembolic event. Based on the results of the study EUPAS1000000440, the expected number of person counts is lowest for DIC (during 1-year follow-up: 5 in FinOMOP-THL – 171 in SIDIAP, with 0 counts in CPRD GOLD, EBB, IPCI, IQVIA DA Germany, and IQVIA LPD Belgium) and highest for VTE (during 1-year follow-up: 27 in IQVIA LPD Belgium – 4,597 in FinOMOP-THL).
8.8. Analysis
8.8.1. Federated network analyses
All analyses will be conducted separately for each data source, and will be carried out in a federated manner, allowing analyses to be run locally without sharing individuals’ data.
Before sharing the study package, test runs of the analytics will be performed on a subset of the data sources and quality control checks will be performed. After all the tests are passed (see Annex III. Operational and reporting considerations), the final package will be released in a version-controlled study repository for execution against all the participating data sources.
8.8.2. Data privacy protection
The data partners will locally execute the analytics against the OMOP CDM in R Studio and review and approve the default aggregated results. They will then be made available to the Principal Investigators and study team in secure online repository of DTZ (Data Transfer Zone). All results will be locked and timestamped for reproducibility and transparency. The study results of all data sources will be checked, after which they are made available to the team, and the Study Dissemination Phase can start. All analyses will be conducted separately for each database, and will be carried out in a federated manner, allowing analyses to be run locally without sharing individual-level data. Cell counts <5 will be suppressed when reporting results to comply with the data source’s privacy protection regulations.
8.8.3. Statistical model specification and assumptions of the analytical approach considered
Objective 1
The probability of not having thromboembolic events at 6-month intervals within 5 years (6, 12, 18, 24, 30, 36, 42, 48, 54, and 60 months) in adults with each type of selected cancer will be calculated based on OMOP CDM mapped data using the R package CohortSurvival, developed by DARWIN EU®. The key measure of not having events at 6-month intervals within 5 years in adults with each type of selected cancer will be the probability, i.e., the proportion of individuals in the cohort who have not yet experienced a thromboembolic event after each interval, accounting for the competing risk of death. Individuals will be censored at the date of loss to follow-up, the end of data availability, or the end of the study period. The estimateCompetingRiskSurvival() function will be used in the analysis.
Objective 2
The median time to onset of venous thromboembolic events in a cohort of adults with thromboembolic events with each type of selected cancer will be calculated based on OMOP CDM mapped data using the R package CohortSurvival, developed by DARWIN EU®. The key measure of median time to onset of venous thromboembolic events will be median survival time obtained using the estimateSingleEventSurvival() function.
All objectives
The R package CohortSurvival is designed to work with data in the OMOP CDM format to extract and summarise survival data applying the Kaplan-Meier method. For objective 1, death will be accounted for as a competing risk. The analyses will be conducted for the overall cohorts as well as by strata of age group, sex, and study subperiod.
The absence of diagnosis codes will be interpreted as the absence of the conditions themselves. A minimum cell count of 5 will be used when reporting results, with any smaller count reported as “<5” and zero counts as “0”.
Sensitivity analysis
Not applicable.
8.8.4. Output
Output will include a PDF report including an executive summary, and tables and figures. Mock versions of the intended tables and figures are listed below.
• Table 1. Attrition of study participants (objectives 1 and 2).
• Figure 1. Probability of not having thromboembolic event after the first cancer diagnosis (objective 1).
This figure will be plotted for each combination of cancer type and outcome, faceted by data source. The main report will include 15 such figures: one for each type of cancer. In the meta-analysis results of EUPAS1000000440, of which this is a routinely repeated study, the most common outcome was VTE across all cancer types, except for liver cancer. In liver cancer, the most common outcome was SVT. Therefore, Figure 1 in the main report will be plotted for SVT in liver cancer and for VTE in all other cancer types. The corresponding figures for all other outcomes across all cancer types will be available in Shiny.
• Table 2. Median time to thromboembolic event after first cancer diagnosis in those with occurrence of thromboembolic event (objective 2).
The main report will include 15 such tables: one for each cancer type.
An interactive dashboard (Shiny) will be generated by incorporating all the results (tables and figures) included in the PDF report mentioned above. Specifically, the Shiny will contain:
• Overall results:
◦ Objective 1: 15 cancer types * 7 outcomes = 105 figures as presented in Figure 1.
◦ Objective 2: 15 tables as presented in Table 2.
• Stratified results:
◦ By age group:
▪ Objective 1: 7 age groups * 105 figures = 735 figures
▪ Objective 2: 7 age groups * 15 tables = 105 tables
◦ By sex:
▪ Objective 1: 2 sexes * 105 figures = 210 figures
▪ Objective 2: 2 sexes * 15 tables = 30 tables
◦ By study subperiod:
▪ Objective 1: 2 subperiods * 105 figures = 210 figures
▪ Objective 2: 2 subperiods * 15 tables = 30 tables.

Table 1. Attrition of study participants.


IQVIA LPD
DK-DHR
EBB
FinOMOP-THL
IQVIA DA
IPCI
SIDIAP
CPRD GOLD
UKBB
Bone cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Brain cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Breast cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Colorectal cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Corpus uteri cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Oesophageal cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Kidney cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Liver cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Lung cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Lymphoma and Leukaemia
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Melanoma
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Ovarian cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Pancreatic cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Prostate cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









Stomach cancer
Qualifying initial records










No prior history of cancer










Outcome during follow-up









N=the number of individuals meeting each criterion.
IQVIA LPD=IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium); DK-DHR=Danish Data Health Registries; EBB=Estonian Biobank; FinOMOP-THL=Finnish Care Register for Health Care; IQVIA DA=IQVIA Disease Analyzer Germany (IQVIA DA Germany); IPCI=Integrated Primary Care Information; SIDIAP=The Information System for Research in Primary Care; CPRD GOLD=Clinical Practice Research Datalink GOLD; UKBB=UK BioBank.


Figure 1. Cumulative probability of not having thromboembolic event after the first cancer diagnosis accounting for a competing risk of death.

Table 2. Median time in days (95% CI) to thromboembolic event after first cancer diagnosis.

IQVIA LPD
DK-DHR
EBB
IQVIA DA
IPCI
SIDIAP
CPRD GOLD
UKBB
FinOMOP-THL
Deep vein thrombosis (DVT)









PE and DVT combined (VTE)









Pelvic vein thrombosis (PVT)









Pulmonary embolism (PE)









Retinal vein thrombosis (RVT)









Splanchnic extrahepatic vein thrombosis (SVT)









IQVIA LPD=IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium); DK-DHR=Danish Data Health Registries; EBB=Estonian Biobank; FinOMOP-THL=Finnish Care Register for Health Care; IQVIA DA=IQVIA Disease Analyzer Germany (IQVIA DA Germany); IPCI=Integrated Primary Care Information; SIDIAP=The Information System for Research in Primary Care; CPRD GOLD=Clinical Practice Research Datalink GOLD; UKBB=UK BioBank; NC=Not calculated (fewer than 5 events).

9. STRENGTHS AND LIMITATIONS
The study will be informed by routinely collected healthcare data, and therefore, data quality issues must be considered. In particular, the identification of individuals with cancer and thromboembolic events may vary across data sources. While relatively few false positives are expected, false negatives may be more likely, especially for primary care data sources that lack patient-level linkage to secondary care data. We expect misclassification to be minimal in registry data sources and in the primary care data sources where cancer diagnoses have been previously validated. Underestimation of thromboembolic events is also possible, particularly for rare events with complex diagnoses, such as RVT and SVT.
Given the large number and diverse nature of participating data sources, it is important to note that differences in patient representations might have resulted from disparate coding practices and specifics of data capture. The granularity or detail of concepts representing clinical facts can vary across source terminologies (e.g., ICD-10, Read codes), influencing how information is later transformed into standardised vocabularies [1]. The preliminary code lists created to identify individuals with cancer include codes from standard vocabularies used in cancer registries, such as ICD-O-3 codes. However, most data sources capture information on cancer diagnoses using SNOMED codes, which may not be granular enough to cover all the topology and morphology of cancer [2]. ICD-O-3 codes are only available at DK-DHR, EBB, and UKBB.

10. REFERENCES
1. Ostropolets A, Reich C, Ryan P, Weng C, Molinaro A, DeFalco F, et al. Characterizing database granularity using SNOMED-CT hierarchy. AMIA Annu Symp Proc. 2021 Jan 25;2020:983–92.
2. Campbell WS, Campbell JR, West WW, McClay JC, Hinrichs SH. Semantic analysis of SNOMED CT for a post-coordinated database of histopathology findings. J Am Med Inform Assoc. 2014 Sept 1;21(5):885–92.

11. ANNEXES
ANNEX I. Description of data sources
DATA SOURCES DESCRIPTION
IQVIA Longitudinal Patient Database Belgium (IQVIA LPD Belgium)
#
Section
Description
1
Database Identification and country
IQVIA LPD Belgium (IQVIA Longitudinal Patient Database Belgium) Belgium
2
Data partner information section
IQVIA IQVIA Europe
3
Coverage and timespan
Data collection since: 2005 Extent: Nation-wide. Panel of 300 GPs in Belgium. The panel is maintained as a representative sample of the primary care physician population in Belgium, according to three criteria known to influence prescribing: age, sex, and geographical distribution.
4
Healthcare setting / type of data
Primary care – gps. Ambulatory visits, with diagnosis, prescriptions, procedures, and laboratory tests.
5
Data collection process
Outpatient electronic health records. Records are entered by GPs at the healthcare encounter.
6
General representativeness
The panel of contributing physicians (a stable 300 GPs) is maintained as a representative sample of the primary care physician population in Belgium, according to three criteria known to influence prescribing: age, sex, and geographical distribution. The panel consists of a stable 300 GPs that are geographically well spread. The total number of active GPs in Belgium is 15,602. The regional geographical spread of physicians in the LPD data is also representative of the distribution across the country: 57% GPs in the North (compared to 54% nationally), 31% in the South (33% nationally), and 12% in Brussels (13%).The provider of the data has more than 2,250 GPs under contract so in case of a drop out a replacement is easily found.
7
Data content /source coding
No information on source coding.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. The patient ID is per practice. So a patient can have different IDs in the DB, one per practice. In Belgium, patients are typically registered at only one GP practice, so duplication should be minimal.
9
Quality control (database specific)
No QC. Integrity constraints only.
10
Linkage
No linkage.
11
Vital status
Death information is derived from healthcare events.
12
Limitations
No database-specific limitations documented. General limitations for the data type applicable.
13
Main references
No main reference provided.
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111116 Website: https://iqvia.com



Danish Data Health Registries (DK-DHR)
#
Section
Description
1
Database Identification and country
DK-DHR (Danish Data Health Registries) Denmark
2
Data partner information section
Danish Medicines Agency (DKMA), Data Analytics Centre (DAC)
3
Coverage and timespan
Data collection since: 1995, Extent: Nationwide. The data is representative of the entire Danish population.
4
Healthcare setting / type of data
Community pharmacists, and secondary care – specialists (ambulatory or hospital outpatient care), and hospital inpatient care. The following data elements are collected: diagnosis (including rare diseases and pregnancy data), hospital admissions, discharge and ICU data, Cause of death, Drug prescription retrievals, vaccination and contraception, Procedures, and Sociodemographic information (sex and age but no information on income, education, occupation).
5
Data collection process
Outpatient electronic health records, and Inpatient hospital electronic health records, and Registries, and Other. All causes of deaths, all retrieved drug prescriptions, all records of vaccinations, all hospital inpatient and outpatients contacts including disease diagnoses and hospital surgical and non-surgical procedures, histologically confirmed incident cancers, laboratory test results for the entire Danish population from 1/1/1995 onwards.
6
General representativeness
The data is representative of the entire Danish population. Healthcare is free in Denmark, so we do not expect any bias in data collection based on socio-economic status.
7
Data content /source coding
Diagnoses and causes of death are collected using the ICD-10 vocabulary. ATC and RxNorm are used for Drugs. SNOMED codes are used for Procedures.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC, ICDO3, cancer Modifier). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. .
9
Quality control (database specific)
The data we have received relating to nationwide Danish Health Data registries offer an opportunity for large-scale, population-based studies with several advantages 1) Their large size improves the precision of estimates and enables the study of rare exposures and outcomes with long-term latency, 2) Inclusion of nearly all individuals in the target population ensures that the data reflect routine clinical care and all clinical segments of the source population, 3) Data are collected independently of each research study, thus minimising certain types of bias, e.g., non-response, and the influence from attention to the research question on the diagnostic process. Before the source data is sent to us, the Danish Health Data Authority does running and comprehensive checks of the registry table data validity of the variables, breaks in data, changes in variable coding, missingness, etc. We perform checks of missingness/completeness in relation to requested variables. In essence, we are receiving a dump of a mirror of the data that is controlled by the SDS. The documentation performed by SDS is available online, in Danish primarily https://www.esundhed.dk/Dokumentation (all variables), but also in English https://sundhedsdatastyrelsen.dk/da/english/health_data_and_registers/national_health_registers
10
Linkage
There is no linkage in this data source.
11
Vital status
The Cause of Death registry (DAR) is used, the cause of death is collected using ICD-10 codes.
12
Limitations
There are no clinical measurements in the data. DK-DHR has the following limitations, which may be relevant confounders for certain complex Darwin EU studies:
• We lack information on key socio-economic status (SES) factors, such as occupation, education, and income. These variables may be important for analysis in some studies.
• We only have complete data on lifestyle factors (such as smoking status and weight) for pregnant women.
• We have no information on patient contacts in primary care (visits to the GP). Consequently, the incidence of chronic diseases like Type 2 Diab
13
Main references
Schmidt M, Schmidt SAJ,Adelborg K,Sundbøll J,Laugesen K,Ehrenstein V,Sørensen HT "The Danish health care system and epidemiological research: from health care contacts to database records." Clinical epidemiology (2019): 31372058
14
Link to HMA-EMA catalogue and database webpage
Website: https://sundhedsdatastyrelsen.dk/da/english/health_data_and_registers/healthdatadenmark HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111217

Estonian Biobank (EBB)
#
Section
Description
1
Database Identification and country
EBB (Estonian Biobank) Estonia
2
Data partner information section
University of Tartu Institute of Computer Science
3
Coverage and timespan
Data collection since: 2004 Extent: Nation-wide. EBB is a nation-wide database containing records from 2004 onwards. Estonian population-based cohort size of 211,800 participants (01/01/2024) aged 18 years and older recruited at GP offices, private practices, and hospitals or in the recruitment offices of the Estonian Genome Center.
4
Healthcare setting / type of data
Primary care – GPs, and community pharmacists, and primary care specialists (e.g. paediatricians), and secondary care – specialists (ambulatory or hospital outpatient care), and hospital inpatient care. Registry which collects electronic records from the biobank and cohort study.
5
Data collection process
Data is retrieved by Estonian Biobank once a year from national registries. The insurance claims are requested from Estonian Health Insurance Fund. The inpatient and outpatient electronic health records are requested from National Health Information System. The cancer registry and cause of death registry information is requested from The National Institute for Health Development. The data is sent to the national registry by the healthcare providers.

6
General representativeness
The age, sex, and geographical distribution closely reflect those of the Estonian adult population and encompass close to 5approximately 20% of adult population. Female participants are over-represented in EBB. Overall, 3.4% of Estonian men and 5.5% of Estonian women are represented in EBB. Older people tend to participate less frequently, however, all age groups are well represented.
7
Data content /source coding
All participants have undergone a standardized health assessment, including provision of blood samples for purification of DNA, white blood cells, and plasma, and completed a questionnaire covering various health-related topics, such as lifestyle, diet, and clinical diagnoses. Diseases and health problems are recorded as ICD-10 codes and prescribed medicine according to the ATC classification and local package codes. Procedures and services are coded with NOMESCO classifier and local service codes.  
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. There is one national identifier that allows linking together all encounters across databases.
9
Quality control (database specific)
The quality control procedures in the Estonian Biobank aim to remove the most obvious mistakes in the data, misspellings, impossible dates, duplicates. Before performing the ETL, several problems are fixed on the source data. Since the ETL procedures are used for a number of different datasets (from the same national sources), we have a growing number of pre-processing steps that correspond to the issues we have discovered previously in the data, such as checking for the presence of critical values, harmonizing date and unit of measurement formats, checking the validity of certain entries against classifiers, etc.
10
Linkage
Follow‐up data are available via linkage with national health‐related registries and via re‐examination of participants. Furthermore, electronic health records are updated for phenotypic outcome information every year. The EBB database is regularly linked with national registries, hospital databases, and the databases of the Estonian Health Insurance Fund (EHIF) and the National Health Information System (NHIS)
11
Vital status
Vital status (death date and causes of death) are obtained from the Causes of Death Registry.
12
Limitations
Participation in EBB cohort is voluntary, therefore the biobank does not represent a random sample and could be subject to recruitment bias. Although recruitment was open to everyone, there is a disproportion between ethnic Estonians and ethnic Russians in the biobank, with Estonians being overrepresented.

13
Main references
Milani, L., Alver, M., Laur, S. et al. The Estonian Biobank’s journey from biobanking to personalized medicine. Nat Commun 16, 3270 (2025). https://doi.org/10.1038/s41467-025-58465-3
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111114 Website: https://genomics.ut.ee/en/content/estonian-biobank

Finnish Care Register for Health Care (FinOMOP-THL)
#
Section
Description
1
Database Identification and country
FinOMOP-THL (Finnish Care Register for Health Care) Finland
2
Data partner information section
Finnish Institute for Health and Welfare (THL) Department of Knowledge Brokers
3
Coverage and timespan
Data collection since: 1998 Extent: Nation-wide. The current CDM population comprises all persons having been alive and residing in Finland since the beginning of 2011.
4
Healthcare setting / type of data
Primary care – gps, and primary care specialists (e.g. paediatricians), and secondary care – specialists (ambulatory or hospital outpatient care), and hospital inpatient care. The THL database covers both public and private, primary, and specialised inpatient and outpatient health care encounters in Finland, starting from 2011. The entire public sector and private inpatient encounters have been included since 2011, while private outpatient encounters, including occupational care, are included since 2020. Since 1998, the register has covered both public outpatient and inpatient specialized care and private inpatient care (TerveysHilmo). Since 2009, the Finnish National Vaccination Register is covered (complete since 2020). The vaccination register covers all vaccinations from the public sector and from a large part of private vaccination providers, with the data coverage from both sections being very good from 2020 onwards. Since 2011, the register has covered public primary care (AvoHilmo). Since 2020, the register has covered private outpatient care and occupational care. In addition, the CDM also contains positive COVID-19 test results from the Finnish National Infectious Diseases Register, which is maintained by THL.
5
Data collection process
Outpatient electronic health records, and Inpatient hospital electronic health records, and Registries. Data is entered by clinicians upon healthcare contact and processed by THL.
6
General representativeness
The THL data has national coverage and is therefore well representative of the Finnish population. Using the complete population as a basis for the person table also serves to facilitate calculations on a population level, e.g. incidence rates.
7
Data content /source coding
The following coding systems have been OMOP-mapped, typically to a good level of completeness: ICD10fi Finnish Extension, ATC, Toimenpideluokitus (procedure classification adapted from the Nordic Classification of Surgical Procedures (NCSP)), Terveydenhuollon erikoisalat (Hilmo specific provider speciality), Rokotustapa (AR/YDIN National classification for vaccine administration), Tupakointistatus (AR/YDIN National classification for smoking status). Vaccinations are identified on product level based on batch number, trade name, vaccine title, and ATC-code. This is mapped on brand and type in the OMOP CDM.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. Each patient in THL has a unique identifier.
9
Quality control (database specific)
The source data collection undergoes a structural and semantic validation before entry into the source database. Additionally, some coded variables undergo quality assessment against the respective code systems post entry into the database. The source registers are also assessed for completeness and coverage, with the aim of improving future collection in the areas where data is lacking.
10
Linkage
THL is already a linkage of multiple Finnish registries (see above).
11
Vital status
The National Population registry data forms the basis for forming the patient population. This ensures an up-to-date location (municipality of residence) of patients, as well as complete death occurrences (although not the cause of death).
12
Limitations
No database-specific limitations documented. General limitations for the data type applicable.
13
Main references
Häkkinen, Pirjo; Mölläri, Kaisa; Saukkonen, Sanna-Mari; Väyrynen, Riikka; Mielikäinen, Lasse; Järvelin, Jutta "Hilmo - Sosiaali- ja terveydenhuollon hoitoilmoitus 2020 : Määrittelyt ja ohjeistus : Voimassa 1.1.2020 alkaen" Terveyden ja hyvinvoinnin laitos (2019):
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111187 Website: https://thl.fi/fi/tilastot-ja-data/ohjeet-tietojen-toimittamiseen/hoitoilmoitusjarjestelma-hilmo

IQVIA Disease Analyzer Germany (IQVIA DA Germany)
#
Section
Description
1
Database Identification and country
IQVIA DA Germany (IQVIA Disease Analyzer Germany) Germany
2
Data partner information section
IQVIA
3
Coverage and timespan
Data collection since: 1989 Extent: Nation-wide. GP and specialists in Germany using specific patient management software.
4
Healthcare setting / type of data
Primary care – gps, and primary care specialists (e.g. paediatricians). Diagnoses, medication, and procedures from an ambulatory setting. Medications are recorded as prescriptions of marketed products.
5
Data collection process
Outpatient electronic health records. By clinicians at healthcare contact.
6
General representativeness
No specific details on general representativeness given.
7
Data content /source coding
Prescription is on product code level (German PZN), ICD10, NFC, Local lab coding.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. There can be patients registered under different ID numbers, because there is no linkage between different GPs.
9
Quality control (database specific)
Data is quality checked on plausibility.
10
Linkage
No.
11
Vital status
Death information is derived from medical events.
12
Limitations
No database-specific limitations documented. General limitations for the data type applicable.
13
Main references
No main reference provided.
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/104282 Website: https://www.iqvia.com/

Integrated Primary Care Information (IPCI)
#
Section
Description
1
Database Identification and country
IPCI (Integrated Primary Care Information) Netherlands
2
Data partner information section
Erasmus University Medical Center Department of Medical Informatics
3
Coverage and timespan
Data collection since: 2006 Extent: Nation-wide. IPCI is a Dutch database that contains patient records from 2006 onwards. However, it mainly covers the central part of the country, including the most densely populated area (the ‘Randstad’) and non-urban areas. IPCI contains information on all patients registered with GPs responsible for non-emergency care and referrals. A patient is registered at birth or at first encounter with the GP.
4
Healthcare setting / type of data
Primary care – gps. Data is collected from primary care EHR. This includes demographic information, complaints and symptoms, diagnoses, laboratory test results, lifestyle factors (in limited amount), and correspondence with secondary care, such as referral and discharge letters.
5
Data collection process
Outpatient electronic health records. Data is entered into the EHR system by the GPs, during or after the visit. Data is aggregated by Erasmus MC data managers and combined in one harmonized database. Several checks are done on this database to ensure correct data processing. Persons are mostly uniquely identified, with the exception of when persons change GP practice (when the same individual can receive several different identifiers).
6
General representativeness
More than 99% of the Dutch population has health insurance, and almost all citizens are registered with a general practitioner. Over 12 months, around 78% of the population has at least one contact with their GP. IPCI included around 350 GP practices out of around 5000 in the country (~ 7%). The demographic composition of the IPCI population mirrors that of the general Dutch population in terms of age and sex.
7
Data content /source coding
Dutch GPs use mainly Dutch standard codes, like ICPC-1 and Diagnostische Bepalingen maintained by NHG. And for therapy the G-Standard is used, maintained by ZIndex.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. Patients can be registered under different IDs. However, in the Netherlands, patients typically have one GP and changing practice is uncommon.
9
Quality control (database specific)
Prior to each data release, extensive quality control steps are performed, e.g., comparison of patient characteristics between practices, and checks to identify abnormal temporal data patterns in practices. For each practice, around 200 quality indicators are obtained. Of these indicators, a quarter refer to population characteristics, e.g. number of birth and mortalities relative to practice size, temporal consistency. The other indicators are based on medical data, e.g. distribution of measurement values, frequencies of diagnoses and procedures relative to age, completeness of data. The indicators are combined in a couple of quality scores for each practice. For these scores, cut-off values for acceptable quality have been defined. Practices with a score below a cut-off are excluded for research. This approach has shown to be very important, for example to check if data from practices that just joined the database are at an acceptable level of quality. The details of the approach, like the cut-off values for acceptance, are based on years of experience. In addition, trends are compared with the previous database release. Extensive quality control steps are performed before each data release. These include comparing patient characteristics between practices and checks to identify abnormal temporal data patterns in practices. Additional checks include over 200 indicators related to population characteristics (e.g., reliability of birth and mortality rates) and medical data (e.g., availability of durations of prescriptions and completeness of laboratory results). Records of low quality are excluded from the database.
10
Linkage
Linkage requires additional approval steps and needs to be assessed on a case-by-case basis. IPCI is not routinely linked with other databases.
11
Vital status
Vital status (death date and cause) is collected based on GP records.
12
Limitations
The main limitation comes with the fact that IPCI is limited to GP records, and although it contains information on referrals and discharge letters, it may not fully capture specific hospital information. IPCI does not include coded/detailed data about medications/procedures/test results from the hospital or other care-providers.
13
Main references
de Ridder MAJ, de Wilde M,de Ben C,Leyba AR,Mosseveld BMT,Verhamme KMC,van der Lei J,Rijnbeek PR "Data Resource Profile: The Integrated Primary Care Information (IPCI) database, The Netherlands." International journal of epidemiology (2022): 35182143
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/42618 Website: http://www.ipci.nl

The Information System for the Development of Research on Primary Care (SIDIAP)
#
Section
Description
1
Database Identification and country
SIDIAP (The Information System for the Development of Research in Primary Care) Catalunya, Spain
2
Data partner information section
IDIAPJGol
3
Coverage and timespan
Data collection since: 2006 Extent: Regional. The SIDIAP database contains records of around 6 million people residing in Catalonia, estimated to be representing around 76% of the Catalan population.
4
Healthcare setting / type of data
Primary care – gps, and hospital inpatient care. SIDIAP captured data includes routine visits, sociodemographic information, diagnoses, laboratory tests, drugs (prescribed and dispensed), referrals, and lifestyle information.
5
Data collection process
Outpatient electronic health records, and Inpatient hospital electronic health records, and Other. Data is entered by primary care physicians upon healthcare contact, supplemented with hospital discharge records. The Institut Catala de la Salut is the owner of the data and acts as the data controller.
6
General representativeness
It was previously shown that the captured SIDIAP population is highly representative of the entire Catalan region in terms of geographic, age, and sex distributions.
7
Data content /source coding
SIDIAP data covers all services that occur at the Primary Care Centres, as well as support services, such as sexual and reproductive health or home end-of-life care. Drugs are coded in ATC-WHO terminology in the source data. Health outcomes are captured in ICD-10CM codes. The SIDIAP contains all laboratory tests and results performed in primary health centres. Demographics, geographical, as well as socio-economic factors are recorded for each patient.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. No.
9
Quality control (database specific)
Internal and external validation processes are carried out to determine the data quality of the SIDIAP information at each data update. These include stratifying the data by geographical regions and year in order to identify differences in data collection that need to be harmonized (e.g. recording of specific information under different codes). The measurement units of variables measuring one characteristic are also homogenized (e.g. transformation of the data from every laboratory that measures haemoglobin to grams per decilitre). Visual inspection of all data included in the database by week is also conducted, allowing one to see temporal patterns in the registry of a certain variable. With this information, the SIDIAP team can issue recommendations to researchers about the most common variable(s) where certain information is recorded (e.g., there are several variables with information concerning the women’s menopausal status and with these visual inspection tools the SIDIAP team can inform the researchers about which related variables have the largest number of records and could be more helpful to capture menopause). Data availability (longitudinally and reliability), plausibility (range checks and unusual values), and consistency are inspected through visualisation tools. In addition, before accessing the data for a requested project, research teams have access to a quality-control report. This document contains counts, years, percentiles, maximums and minimums, incidences, and prevalence of the data requested for the project, allowing detection of inconsistencies in the data extraction prior to data delivery. External validation processes of the SIDIAP database mainly include assessing the data recorded in SIDIAP through linkage to external gold standard data sources, by analysing free text, or by sending questionnaires to health professionals.
10
Linkage
SIDIAP is linked to a hospital discharge database, pharmacy dispensation, and primary care laboratories. It can also be linked to other registries in Catalonia on a project-by-project basis.
11
Vital status
Mortality is fully captured in SIDIAP. The cause of death is not available but can be linked to the Spanish death registry on a project-by-project basis.
12
Limitations
The SIDIAP data is not representative of individuals not using public primary care, and conditions that are usually followed by specialist care might not be properly captured. In addition, there is limited information on lifestyle variables (not always requested in primary care visit and, therefore, the information is missing in many cases). Patients are followed until Death or when transferring to another primary health care centre that does not contribute to SIDIAP.
13
Main references
Recalde M, Rodríguez C,Burn E,Far M,García D,Carrere-Molina J,Benítez M,Moleras A,Pistillo A,Bolíbar B,Aragón M,Duarte-Salles T "Data Resource Profile: The Information System for Research in Primary Care (SIDIAP)." International journal of epidemiology (2022): 35415748
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/50190 Website: https://www.sidiap.org/index.php/en

Clinical Practice Research Datalink GOLD (Oxford) (CPRD GOLD)
#
Section
Description
1
Database Identification and country
CPRD GOLD (Clinical Practice Research Datalink GOLD ) United Kingdom
2
Data partner information section
University of Oxford NDORMS
3
Coverage and timespan
Data collection since: 1987 Extent: Nation-wide. CPRD GOLD consists of patients in contributing practices using Vision software. Historically this covered the whole of the UK, but the number of contributing practices in the England is dropping. In January 2025 only 3 practices from England were a part of CPRD GOLD, while historical patient data were from the whole of the UK, and will continue to be so. In the future, no practices from England will be present, only practices from Scotland, Wales, and Northern Ireland.
4
Healthcare setting / type of data
Primary care – gps, and primary care specialists (e.g. paediatricians), and secondary care – specialists (ambulatory or hospital outpatient care), and hospital inpatient care. CPRD GOLD data include patient demographics, biological measurements, clinical symptoms and diagnoses, referrals to specialist/hospital and their outcome, laboratory tests/results, and prescribed medications.
5
Data collection process
Outpatient electronic health records. Data is entered by clinicians into the EHR. Data is processed by CPRD and provides data releases for research.
6
General representativeness
CPRD GOLD has been assessed and found to be broadly representative of the UK general population in terms of age, gender, and ethnicity. In CPRD GOLD in January 2025 there were 2,730,707 current acceptable patients (i.e. registered at currently contributing practices that use Vision software, excluding transferred out, deceased patients, and those flagged by CPRD as not acceptable for clinical research for data quality issues). This equals to 4.07%, based on the UK population estimates of 67,026,300 from the Office of National Statistics (mid-2023). Current patients are only from Scotland, Wales, and Northern Ireland. Historically, GOLD does contain data from England as well.
7
Data content /source coding
Gemscript, Read, dm+d
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. In GOLD, a patient can be registered under different ID numbers upon changing practice or re-registration. Researchers are not able to identify these patients, as the data are anonymised. However, GOLD covers less than 5% of the current UK GP practices and it is unlikely that an individual who does change GP practice ends up in another GP practice which uses the Vision software and accepts the CPRD data collection agreement. The very small number of duplicated IDs will have different observation periods and should not have an impact on the data analyses.
9
Quality control (database specific)
CPRD GOLD only includes practices whose data quality is assessed to be up-to-standard (uts). Each practice is associated to an uts date set when the data quality standards become satisfactory, and CPRD recommend using only longitudinal data starting from this uts date. Every time CPRD collect the EHR from a practice, checks are run for the data quality standards and if they are not adequate, the EHR is not accepted. When the data quality becomes acceptable again, CPRD updates the practice uts date. CPRD also check data quality standards at the patient level and associate each patient to a flag, reporting if its data is acceptable for clinical research. Only patients with acceptable data quality are included in the population to be mapped to CDM.
10
Linkage
CPRD GOLD can be linked to several sources, however our Oxford OMOP CDM is only linked to the CPRD GOLD Ethnicity Record and to the CPRD Townsend Deprivation Index at Practice Level
11
Vital status
Vital status is retrieved from the GP records. Population registry (ONS) data can be requested on a study-by-study basis and linked. This data only covers England and is planned to be mapped to OMOP in the future. The cause of death is not captured.
12
Limitations
The main limitation is due to the fact that CPRD GOLD is limited to GP records, and although it contains information on referrals and discharge letters, it may not fully capture specific hospital information. Events from hospital and specialist care are not covered.
13
Main references
Sanchez-Santos MT, Axson EL,Dedman D,Delmestri A "Data Resource Profile Update: CPRD GOLD." International journal of epidemiology (2025): 40499193
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111113 Website: https://cprd.com

UK BioBank (UKBB)
#
Section
Description
1
Database Identification and country
UKBB (UK BioBank) United Kingdom
2
Data partner information section
Oxford University NDORMS
3
Coverage and timespan
Data collection since: 2006 Extent: Nation-wide. People recruited from whole of the UK.
4
Healthcare setting / type of data
Primary care – gps, and primary care specialists (e.g. paediatricians), and secondary care – specialists (ambulatory or hospital outpatient care), and hospital inpatient care, and other (specify). UK Biobank is made by a rich variety of data sources, which include genetic data, primary care data, hospital inpatient data, death data, and cancer registry.
5
Data collection process
Inpatient hospital electronic health records, and Registries, and Biobank. The baseline assessment is consisting of both patient reported data (questionnaire) and physical measurements. GP and hospital data, as well as death and cancer registry records, are linked afterwards and are subject to data validation: https://biobank.ctsu.ox.ac.uk/~bbdatan/Data_cleaning_overall_doc_showcase_v1.pdf
6
General representativeness
The database population consists of volunteers aged 40-69years. We can expect that volunteers might have been more willing to participate if living closer to one of the 22 recruitment centres or if more interested in health issues compared to the general population. These aspects might have introduced an unavoidable bias in the cohort.
7
Data content /source coding
READ2, READ3, DM+D, ICD9, ICD10, OPCS3, OPCS4, ICD-O-3 are used.
8
Data Harmonization
The data has been mapped to the OMOP CDM v5.4 and the OMOP standard vocabularies (SNOMED, RxNorm, LOINC). The format, structural and semantic conformance has been verified upon onboarding into the DARWIN EU® data network. No.
9
Quality control (database specific)
All UK BioBank data are provided already curated and each of the many datasets have specific curation algorithms and procedures. As always, primary care and hospital data, which come from real-world setting, need special attention regarding data quality. Please refer to the link below for specific details https://biobank.ndph.ox.ac.uk/showcase/showcase/docs/primary_care_data.pdf https://biobank.ndph.ox.ac.uk/showcase/showcase/docs/HospitalEpisodeStatistics.pdf
10
Linkage
The database contains liked data from death and cancer registry, GP and hospital for the participants.
11
Vital status
Linked to the national death registry.
12
Limitations
The UKBB source data will not be updated anymore and have no new records after December 2022. There is no day's supply information captured in the source. GP prescription data are available for 45% of the cohort. There is no information on dispensed medicines. GP laboratory tests and results are not available.
13
Main references
Hewitt J, Walters M,Padmanabhan S,Dawson J "Cohort profile of the UK Biobank: diagnosis and characteristics of cerebrovascular disease." BMJ open (2016): 27006341
14
Link to HMA-EMA catalogue and database webpage
HMA-EMA Catalogue entry: https://catalogues.ema.europa.eu/data-source/1111233 Website: https://www.ukbiobank.ac.uk/

 
ANNEX II. Fitness for use assessment
Data source justification for inclusion and key characteristics
The selected data sources met the criteria required to capture outcomes of interest and relevant data, enabling a patient-level characterisation of newly diagnosed individuals with cancer across different European settings and regions. The main criterion was a meaningful number of person counts for the population of interest (individuals with cancer) and outcomes (thromboembolic events) assessed at the feasibility stage for all data sources included in the study. Data sources were also selected based on European representativeness. Not all data sources had records of all outcomes of interest.
Additional criteria reflect other data quality domains assessed at the DARWIN EU® data partners onboarding stage. With every new release of the data partners OMOP CDM, the DARWIN EU® coordination centre also receives new results of the CdmOnboarding, DashboardExport, and DataQualityDashboard packages and assesses the quality of the data. No open quality issues related to the study population and outcomes were present for any of the data sources selected for the study.
Relevance was also assessed based on the previous research related to the study population of the outcome. Previously, IPCI, EBB, SIDIAP, CPRD, IQVIA DA Germany, UKBB, and EBB were used in studies with thromboembolic events as an outcome. (Ali et al., 2020, Mercadé-Besora et al., 2024, Li et al., 2022, Voss et al, 2023). CPRD, FinOMOP-THL, UKBB, SIDIAP, IPCI, and UKBB were used in studies repeated on individuals with cancer (Hagberg et al., 2023; Corby et al., 2024, Leinonen et al., 2017, Smith et al., 2024; Recalde et al, 2019; van Soest 2008; Chen et al., 2024). 
In addition to that, DK-DHR is a nationwide, fully representative data source that includes information from the National Patient Registry and the National Cancer Register.

Design elements
Operational definition
Data elements for valid capture
Criticality of the quality of the element, including justification where relevant
Study population
Objective 1
Inclusion criteria
• First diagnosis of a selected cancer (index date) between 01/01/2016 and 31/12/2022
• Age ≥18 years at index date
• Minimum 365 days of available history before index date
• Index date ≥365 days prior to end of data availability of the data source
Exclusion criteria
• Diagnoses of multiple primary tumours at index date
• History of cancer diagnosis ever before index date
• Outcome during the year prior to index date
Objective 2
Inclusion criteria
• Included in the study population of objective 1
• Occurrence of the outcome during follow-up
• First diagnosis of a selected cancer
Diagnosis of:
• Deep vein thrombosis (DVT)
• Pulmonary embolisms (PE)
• Venous thromboembolism (VTE, composite of DVT and PE)
• Pelvic venous thrombosis (PVT)
• Splanchnic vein thrombosis (SVT), including hepatic and extra-hepatic vein thrombosis
• Retinal vein thrombosis (RVT), including retinal central vein thrombosis
• Disseminated intravascular coagulation (DIC)
Low/Medium/High
Treatment/ exposure
Not applicable.

Low/Medium/High
Comparator group (if relevant)
Not applicable.

Low/Medium/High
Outcomes (if relevant)
All objectives
The thromboembolic event outcomes in this study are identical to those in EUPAS1000000440, of which this is a routinely repeated study:
• Deep vein thrombosis (DVT)
• Pulmonary embolisms (PE)
• Venous thromboembolism (VTE, composite of DVT and PE)
• Pelvic venous thrombosis (PVT)
• Splanchnic vein thrombosis (SVT), including hepatic and extra-hepatic vein thrombosis
• Retinal vein thrombosis (RVT), including retinal central vein thrombosis
• Disseminated intravascular coagulation (DIC)
Each of the specific thromboembolic events was a primary, binary outcome. The outcome was assessed at any diagnosis position in the electronic health record and from any of the care settings in each data source.
The list of concepts, from EUPAS1000000440 and based on SNOMED codes and aligned with previous studies that used OMOP CDM and VTE as an outcome (Burn et al., 2022), is provided in Annex IV.
Diagnosis of:
• Deep vein thrombosis (DVT)
• Pulmonary embolisms (PE)
• Venous thromboembolism (VTE, composite of DVT and PE)
• Pelvic venous thrombosis (PVT)
• Splanchnic vein thrombosis (SVT), including hepatic and extra-hepatic vein thrombosis
• Retinal vein thrombosis (RVT), including retinal central vein thrombosis
• Disseminated intravascular coagulation (DIC)
Low/Medium/High
Covariates (including confounders if relevant)
All Objectives
• Sex
◦ Female/male
• Age groups (years) at index date:
◦ 18–34
◦ 35–44
◦ 45–54
◦ 55–64
◦ 65–74
◦ 75–84
◦ ≥85
Age at index date will be calculated using January 1st of the year of birth as proxy for the actual birthday. Date/month is either not present or cannot be made available for governance reasons. If available, date is often set to first of the month for personal privacy.
• Study subperiod:
◦ 2016–2019
◦ 2020-2022
The status of each covariate will be assessed at index date. Each covariate will be used to stratify the results.
• Date of first diagnosis of selected cancer (index date)
• Sex
• Age at index date
Low/Medium/High
Follow-up time (if relevant)
For both objectives, follow-up will start on the date of cancer diagnosis (index date) and end on the earliest of occurrence of the outcome, loss to follow-up, end of data availability, death, or end of the study period (31/12/2024).
• Date of first diagnosis of selected cancer (index date)
• Date of thromboembolic event
• Data of end of data availability
• Death date
Low/Medium/High
EMA Data Quality Framework for EU medicines regulation: application to Real-World Data for more information (https://www.ema.europa.eu/system/files/documents/other/data-quality-framework-eu-medicines-regulation-application-real-world-data_en.pdf)
ANNEX III. Operational and reporting considerations
DATA MANAGEMENT
Data management
All data sources have previously mapped their data to the OMOP common data model. This enables the use of standardised analytics and using DARWIN EU® tools across the network, since the structure of the data and the terminology system is harmonised. The OMOP CDM was developed and maintained by the Observational Health Data Sciences and Informatics (OHDSI) initiative and is described in detail on the wiki page of the CDM: https://ohdsi.github.io/CommonDataModel and in The Book of OHDSI: http://book.ohdsi.org.
The analytic code for this study will be written in R and will use standardized analytics wherever possible. Each data partner will execute the study code against their data source containing patient-level data and then return the results (csv files), which will only contain aggregated data. The results from each of the contributing data sites will then be combined in tables and figures for the study report.
Data storage and protection
For this study, participants from various EU member states will process personal data from individuals that is collected in national/regional electronic health record data sources. Due to the sensitive nature of this personal medical data, it is important to be fully aware of ethical and regulatory aspects and to strive to take all reasonable measures to ensure compliance with ethical and regulatory issues on privacy.
All data sources used in this study are already used for pharmaco-epidemiological research and have a well-developed mechanism to ensure that European and local regulations dealing with ethical use of the data and adequate privacy control are adhered to. In agreement with these regulations, rather than combining person level data and performing only a central analysis, local analyses will be run, which generate non-identifiable aggregate summary results. 
The output files are stored in the DARWIN EU® Remote Research Environment (RRE). These output files do not contain any data that allow identification of subjects included in the study. The RRE implements further security measures to ensure a high level of stored data protection to comply with the local implementation of the General Data Protection Regulation (GDPR) (EU) 679/20161 in the various member states. 
QUALITY CONTROL
Data source quality control
When defining drug cohorts, non-systemic products will be excluded from the list of included codes summarised on the ingredient level.
When defining cohorts for indications, a systematic search of possible codes for inclusion will be identified using the CodelistGenerator R package (https://github.com/darwin-eu/CodelistGenerator). This package allows the user to define a search strategy and will use this to query the vocabulary tables of the OMOP common data model so as to find potentially relevant codes. In addition, the CohortDiagnostics (https://github.com/OHDSI/CohortDiagnostics) and DrugExposureDiagnostics (https://cran.r-project.org/web/packages/DrugExposureDiagnostics/index.html) R packages will be run, if needed, to assess the use of different codes across the data sources contributing to the study and identify any codes potentially omitted in error. The DrugExposureDiagnostics package evaluates ingredient-specific attributes and patterns in drug exposure records.
The study code will be based on DARWIN EU® R packages: IncidencePrevalence to estimate Incidence and Prevalence, DrugUtilisation to characterise the drug use, and CohortCharacteristics to characterise the cohort by indication. These packages will include numerous automated unit tests to ensure the validity of the codes, alongside software peer review and user testing. The R package will be made publicly available via GitHub. 
PLANS FOR DISSEMINATING AND COMMUNICATING STUDY RESULTS
A PDF report including an executive summary, and the specified tables and/or figures will be submitted to EMA by the DARWIN EU® CC upon completion of the study.
An interactive dashboard incorporating all the results (tables and figures) will be provided alongside the PDF report. The full set of underlying aggregated data used in the dashboard will also be made available, if requested.

ANNEX IV. List of stand-alone documents
Concepts to define individuals with cancer (populations of interest) are available in a stand-alone document (DARWIN_EU_P3_C3_005 _Cancer_phenotypes.xlsx). Tables S1 to S7 include concepts used to define outcomes (thromboembolic events).

Table S1. List of concepts used to define deep vein thrombosis (DVT).
Concept ID
Concept Name
Domain
Vocabulary
762047
Acute bilateral thrombosis of subclavian veins
Condition
SNOMED
762148
Acute deep vein thrombosis of bilateral iliac veins
Condition
SNOMED
37169261
Acute deep vein thrombosis of bilateral lower limbs following procedure
Condition
SNOMED
37169249
Acute deep vein thrombosis of bilateral upper limbs following procedure
Condition
SNOMED
35616028
Acute deep vein thrombosis of left iliac vein
Condition
SNOMED
35615035
Acute deep vein thrombosis of left lower limb following procedure
Condition
SNOMED
35615031
Acute deep vein thrombosis of left upper limb following procedure
Condition
SNOMED
43531681
Acute deep vein thrombosis of lower limb
Condition
SNOMED
35616027
Acute deep vein thrombosis of right iliac vein
Condition
SNOMED
35615034
Acute deep vein thrombosis of right lower limb following procedure
Condition
SNOMED
35615030
Acute deep vein thrombosis of right upper limb following procedure
Condition
SNOMED
44782746
Acute deep venous thrombosis
Condition
SNOMED
44782751
Acute deep venous thrombosis of axillary vein
Condition
SNOMED
762008
Acute deep venous thrombosis of bilateral axillary veins
Condition
SNOMED
760875
Acute deep venous thrombosis of bilateral calves
Condition
SNOMED
765155
Acute deep venous thrombosis of bilateral ileofemoral veins
Condition
SNOMED
762017
Acute deep venous thrombosis of bilateral internal jugular veins
Condition
SNOMED
762417
Acute deep venous thrombosis of bilateral legs
Condition
SNOMED
761461
Acute deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
762020
Acute deep venous thrombosis of bilateral popliteal veins
Condition
SNOMED
765546
Acute deep venous thrombosis of bilateral tibial veins
Condition
SNOMED
762004
Acute deep venous thrombosis of both upper extremities
Condition
SNOMED
44782742
Acute deep venous thrombosis of calf
Condition
SNOMED
44782747
Acute deep venous thrombosis of femoral vein
Condition
SNOMED
762015
Acute deep venous thrombosis of ileofemoral vein of left leg
Condition
SNOMED
765541
Acute deep venous thrombosis of ileofemoral vein of right lower extremity
Condition
SNOMED
44782748
Acute deep venous thrombosis of iliofemoral vein
Condition
SNOMED
44782752
Acute deep venous thrombosis of internal jugular vein
Condition
SNOMED
762009
Acute deep venous thrombosis of left axillary vein
Condition
SNOMED
760876
Acute deep venous thrombosis of left calf
Condition
SNOMED
765540
Acute deep venous thrombosis of left femoral vein
Condition
SNOMED
765922
Acute deep venous thrombosis of left internal jugular vein
Condition
SNOMED
762418
Acute deep venous thrombosis of left lower extremity
Condition
SNOMED
761462
Acute deep venous thrombosis of left pelvic vein
Condition
SNOMED
618482
Acute deep venous thrombosis of left peroneal vein
Condition
SNOMED
765537
Acute deep venous thrombosis of left upper extremity
Condition
SNOMED
44782767
Acute deep venous thrombosis of lower extremity as complication of procedure
Condition
SNOMED
44782761
Acute deep venous thrombosis of pelvic vein
Condition
SNOMED
762022
Acute deep venous thrombosis of politeal vein of right leg
Condition
SNOMED
44782743
Acute deep venous thrombosis of popliteal vein
Condition
SNOMED
762021
Acute deep venous thrombosis of popliteal vein of left leg
Condition
SNOMED
762010
Acute deep venous thrombosis of right axillary vein
Condition
SNOMED
760877
Acute deep venous thrombosis of right calf
Condition
SNOMED
762013
Acute deep venous thrombosis of right femoral vein
Condition
SNOMED
762018
Acute deep venous thrombosis of right internal jugular vein
Condition
SNOMED
762419
Acute deep venous thrombosis of right lower extremity
Condition
SNOMED
765229
Acute deep venous thrombosis of right pelvic vein
Condition
SNOMED
618681
Acute deep venous thrombosis of right peroneal vein
Condition
SNOMED
762005
Acute deep venous thrombosis of right upper extremity
Condition
SNOMED
44782745
Acute deep venous thrombosis of thigh
Condition
SNOMED
44782744
Acute deep venous thrombosis of tibial vein
Condition
SNOMED
762026
Acute deep venous thrombosis of tibial vein of left leg
Condition
SNOMED
765156
Acute deep venous thrombosis of tibial vein of right leg
Condition
SNOMED
44782421
Acute deep venous thrombosis of upper extremity
Condition
SNOMED
44782766
Acute deep venous thrombosis of upper extremity as complication of procedure
Condition
SNOMED
37171353
Acute ischemia of colon due to thrombosis of mesenteric vein
Condition
SNOMED
37170675
Acute ischemia of small intestine due to thrombosis of mesenteric vein
Condition
SNOMED
762048
Acute thrombosis of left subclavian vein
Condition
SNOMED
45757410
Acute thrombosis of mesenteric vein
Condition
SNOMED
762049
Acute thrombosis of right subclavian vein
Condition
SNOMED
36712892
Acute thrombosis of splenic vein
Condition
SNOMED
44782762
Acute thrombosis of subclavian vein
Condition
SNOMED
4179911
Axillary vein thrombosis
Condition
SNOMED
37109253
Bilateral acute deep vein thrombosis of femoral veins
Condition
SNOMED
618678
Bilateral acute deep venous thrombosis of peroneal veins
Condition
SNOMED
609003
Bilateral deep femoral vein thrombophlebitis
Condition
SNOMED
3179900
Bilateral deep vein thromboses
Condition
Nebraska Lexicon
40478951
Bilateral deep vein thrombosis of lower extremities
Condition
SNOMED
609002
Bilateral femoral vein thrombophlebitis
Condition
SNOMED
608965
Bilateral iliac vein thrombophlebitis
Condition
SNOMED
1245776
Bilateral popliteal vein thrombophlebitis
Condition
SNOMED
609006
Bilateral tibial vein thrombophlebitis
Condition
SNOMED
4042396
Deep thrombophlebitis
Condition
SNOMED
4046884
Deep vein thrombosis of leg related to air travel
Condition
SNOMED
3655221
Deep vein thrombosis of lower extremity due to intravenous drug use
Condition
SNOMED
4133004
Deep venous thrombosis
Condition
SNOMED
761013
Deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
37163011
Deep venous thrombosis of calf
Condition
SNOMED
45773536
Deep venous thrombosis of femoropopliteal vein
Condition
SNOMED
763942
Deep venous thrombosis of left lower extremity
Condition
SNOMED
1075379
Deep venous thrombosis of left posterior tibial vein
Condition
SNOMED
761980
Deep venous thrombosis of left upper extremity
Condition
SNOMED
443537
Deep venous thrombosis of lower extremity
Condition
SNOMED
4133975
Deep venous thrombosis of pelvic vein
Condition
SNOMED
40480555
Deep venous thrombosis of peroneal vein
Condition
SNOMED
1075377
Deep venous thrombosis of posterior tibial vein
Condition
SNOMED
4322565
Deep venous thrombosis of profunda femoris vein
Condition
SNOMED
763941
Deep venous thrombosis of right lower extremity
Condition
SNOMED
1075378
Deep venous thrombosis of right posterior tibial vein
Condition
SNOMED
761928
Deep venous thrombosis of right upper extremity
Condition
SNOMED
4207899
Deep venous thrombosis of tibial vein
Condition
SNOMED
4028057
Deep venous thrombosis of upper extremity
Condition
SNOMED
193512
Embolism and thrombosis of the renal vein
Condition
SNOMED
435565
Embolism and thrombosis of the vena cava
Condition
SNOMED
4258295
Embolism from thrombosis of vein of distal lower extremity
Condition
SNOMED
40481089
Embolism from thrombosis of vein of lower extremity
Condition
SNOMED
40479840
Embolism from thrombosis of vein of thigh
Condition
SNOMED
4119760
Iliofemoral deep vein thrombosis
Condition
SNOMED
4124856
Inferior mesenteric vein thrombosis
Condition
SNOMED
608964
Left iliac vein thrombophlebitis
Condition
SNOMED
602592
Left peroneal vein thrombophlebitis
Condition
SNOMED
600938
Left subclavian vein thrombophlebitis
Condition
SNOMED
37164448
Lemierre syndrome
Condition
SNOMED
4281689
Phlegmasia alba dolens
Condition
SNOMED
4284538
Phlegmasia cerulea dolens
Condition
SNOMED
3185768
Popliteal vein thrombosis
Condition
Nebraska Lexicon
4309333
Postoperative deep vein thrombosis
Condition
SNOMED
1245858
Postpartum acute deep vein thrombosis
Condition
SNOMED
46285905
Provoked deep vein thrombosis
Condition
SNOMED
608963
Right iliac vein thrombophlebitis
Condition
SNOMED
602583
Right peroneal vein thrombophlebitis
Condition
SNOMED
600939
Right subclavian vein thrombophlebitis
Condition
SNOMED
4033521
Splenic vein thrombosis
Condition
SNOMED
4055089
Superior mesenteric vein thrombosis
Condition
SNOMED
4230403
Thrombophlebitis of axillary vein
Condition
SNOMED
4069561
Thrombophlebitis of deep femoral vein
Condition
SNOMED
761831
Thrombophlebitis of deep vein of bilateral lower limbs
Condition
SNOMED
761830
Thrombophlebitis of deep vein of left lower limb
Condition
SNOMED
761808
Thrombophlebitis of deep vein of left upper limb
Condition
SNOMED
761832
Thrombophlebitis of deep vein of right lower limb
Condition
SNOMED
761809
Thrombophlebitis of deep vein of right upper limb
Condition
SNOMED
4221821
Thrombophlebitis of deep veins of lower extremity
Condition
SNOMED
440750
Thrombophlebitis of deep veins of upper extremities
Condition
SNOMED
4203618
Thrombophlebitis of femoropopliteal vein
Condition
SNOMED
4176614
Thrombophlebitis of iliac vein
Condition
SNOMED
764715
Thrombophlebitis of internal jugular vein
Condition
SNOMED
608904
Thrombophlebitis of left axillary vein
Condition
SNOMED
761821
Thrombophlebitis of left deep femoral vein
Condition
SNOMED
761819
Thrombophlebitis of left femoral vein
Condition
SNOMED
609000
Thrombophlebitis of left popliteal vein
Condition
SNOMED
609005
Thrombophlebitis of left tibial vein
Condition
SNOMED
4318407
Thrombophlebitis of mesenteric vein
Condition
SNOMED
608903
Thrombophlebitis of right axillary vein
Condition
SNOMED
761820
Thrombophlebitis of right deep femoral vein
Condition
SNOMED
761818
Thrombophlebitis of right femoral vein
Condition
SNOMED
609001
Thrombophlebitis of right popliteal vein
Condition
SNOMED
609004
Thrombophlebitis of right tibial vein
Condition
SNOMED
4205652
Thrombophlebitis of subclavian vein
Condition
SNOMED
4110339
Thrombophlebitis of the anterior tibial vein
Condition
SNOMED
4111868
Thrombophlebitis of the common iliac vein
Condition
SNOMED
4110343
Thrombophlebitis of the external iliac vein
Condition
SNOMED
439314
Thrombophlebitis of the femoral vein
Condition
SNOMED
4109877
Thrombophlebitis of the internal iliac vein
Condition
SNOMED
4112171
Thrombophlebitis of the popliteal vein
Condition
SNOMED
4112172
Thrombophlebitis of the posterior tibial vein
Condition
SNOMED
4250765
Thrombophlebitis of tibial vein
Condition
SNOMED
42538533
Thrombosis of iliac vein
Condition
SNOMED
44811347
Thrombosis of internal jugular vein
Condition
SNOMED
765049
Thrombosis of left peroneal vein
Condition
SNOMED
4317289
Thrombosis of mesenteric vein
Condition
SNOMED
4203836
Thrombosis of subclavian vein
Condition
SNOMED
4175649
Thrombosis of the popliteal vein
Condition
SNOMED
4153353
Traumatic thrombosis of axillary vein
Condition
SNOMED
46285904
Unprovoked deep vein thrombosis
Condition
SNOMED
37163265
Venous thromboembolism due to thrombosis of vein of lower limb
Condition
SNOMED

Table S2. List of concepts used to define pulmonary embolism (PE).
Concept ID
Concept Name
Domain
Vocabulary
608954
Acute cor pulmonale due to septic pulmonary embolism
Condition
SNOMED
4120091
Acute massive pulmonary embolism
Condition
SNOMED
45768439
Acute pulmonary embolism
Condition
SNOMED
45768888
Acute pulmonary thromboembolism
Condition
SNOMED
762808
Infarction of lung due to embolus
Condition
SNOMED
40480461
Infarction of lung due to iatrogenic pulmonary embolism
Condition
SNOMED
4108681
Postoperative pulmonary embolus
Condition
SNOMED
37160752
Postoperative pulmonary thromboembolism
Condition
SNOMED
1244882
Pulmonary artery embolism due to foreign body
Condition
SNOMED
440417
Pulmonary embolism
Condition
SNOMED
37109911
Pulmonary embolism due to and following acute myocardial infarction
Condition
SNOMED
37016922
Pulmonary embolism on long-term anticoagulation therapy
Condition
SNOMED
43530605
Pulmonary embolism with pulmonary infarction
Condition
SNOMED
4253796
Pulmonary microemboli
Condition
SNOMED
4121618
Pulmonary thromboembolism
Condition
SNOMED
36713113
Saddle embolus of pulmonary artery
Condition
SNOMED
35615055
Saddle embolus of pulmonary artery with acute cor pulmonale
Condition
SNOMED
40479606
Septic pulmonary embolism
Condition
SNOMED
4119607
Subacute massive pulmonary embolism
Condition
SNOMED

Table S3. List of concepts used to define venous thromboembolism (VTE).
Concept ID
Concept Name
Domain
Vocabulary
762047
Acute bilateral thrombosis of subclavian veins
Condition
SNOMED
608954
Acute cor pulmonale due to septic pulmonary embolism
Condition
SNOMED
762148
Acute deep vein thrombosis of bilateral iliac veins
Condition
SNOMED
37169261
Acute deep vein thrombosis of bilateral lower limbs following procedure
Condition
SNOMED
37169249
Acute deep vein thrombosis of bilateral upper limbs following procedure
Condition
SNOMED
35616028
Acute deep vein thrombosis of left iliac vein
Condition
SNOMED
35615035
Acute deep vein thrombosis of left lower limb following procedure
Condition
SNOMED
35615031
Acute deep vein thrombosis of left upper limb following procedure
Condition
SNOMED
43531681
Acute deep vein thrombosis of lower limb
Condition
SNOMED
35616027
Acute deep vein thrombosis of right iliac vein
Condition
SNOMED
35615034
Acute deep vein thrombosis of right lower limb following procedure
Condition
SNOMED
35615030
Acute deep vein thrombosis of right upper limb following procedure
Condition
SNOMED
44782746
Acute deep venous thrombosis
Condition
SNOMED
44782751
Acute deep venous thrombosis of axillary vein
Condition
SNOMED
762008
Acute deep venous thrombosis of bilateral axillary veins
Condition
SNOMED
760875
Acute deep venous thrombosis of bilateral calves
Condition
SNOMED
765155
Acute deep venous thrombosis of bilateral ileofemoral veins
Condition
SNOMED
762017
Acute deep venous thrombosis of bilateral internal jugular veins
Condition
SNOMED
762417
Acute deep venous thrombosis of bilateral legs
Condition
SNOMED
761461
Acute deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
762020
Acute deep venous thrombosis of bilateral popliteal veins
Condition
SNOMED
765546
Acute deep venous thrombosis of bilateral tibial veins
Condition
SNOMED
762004
Acute deep venous thrombosis of both upper extremities
Condition
SNOMED
44782742
Acute deep venous thrombosis of calf
Condition
SNOMED
44782747
Acute deep venous thrombosis of femoral vein
Condition
SNOMED
762015
Acute deep venous thrombosis of ileofemoral vein of left leg
Condition
SNOMED
765541
Acute deep venous thrombosis of ileofemoral vein of right lower extremity
Condition
SNOMED
44782748
Acute deep venous thrombosis of iliofemoral vein
Condition
SNOMED
44782752
Acute deep venous thrombosis of internal jugular vein
Condition
SNOMED
762009
Acute deep venous thrombosis of left axillary vein
Condition
SNOMED
760876
Acute deep venous thrombosis of left calf
Condition
SNOMED
765540
Acute deep venous thrombosis of left femoral vein
Condition
SNOMED
765922
Acute deep venous thrombosis of left internal jugular vein
Condition
SNOMED
762418
Acute deep venous thrombosis of left lower extremity
Condition
SNOMED
761462
Acute deep venous thrombosis of left pelvic vein
Condition
SNOMED
618482
Acute deep venous thrombosis of left peroneal vein
Condition
SNOMED
765537
Acute deep venous thrombosis of left upper extremity
Condition
SNOMED
44782767
Acute deep venous thrombosis of lower extremity as complication of procedure
Condition
SNOMED
44782761
Acute deep venous thrombosis of pelvic vein
Condition
SNOMED
762022
Acute deep venous thrombosis of politeal vein of right leg
Condition
SNOMED
44782743
Acute deep venous thrombosis of popliteal vein
Condition
SNOMED
762021
Acute deep venous thrombosis of popliteal vein of left leg
Condition
SNOMED
762010
Acute deep venous thrombosis of right axillary vein
Condition
SNOMED
760877
Acute deep venous thrombosis of right calf
Condition
SNOMED
762013
Acute deep venous thrombosis of right femoral vein
Condition
SNOMED
762018
Acute deep venous thrombosis of right internal jugular vein
Condition
SNOMED
762419
Acute deep venous thrombosis of right lower extremity
Condition
SNOMED
765229
Acute deep venous thrombosis of right pelvic vein
Condition
SNOMED
618681
Acute deep venous thrombosis of right peroneal vein
Condition
SNOMED
762005
Acute deep venous thrombosis of right upper extremity
Condition
SNOMED
44782745
Acute deep venous thrombosis of thigh
Condition
SNOMED
44782744
Acute deep venous thrombosis of tibial vein
Condition
SNOMED
762026
Acute deep venous thrombosis of tibial vein of left leg
Condition
SNOMED
765156
Acute deep venous thrombosis of tibial vein of right leg
Condition
SNOMED
44782421
Acute deep venous thrombosis of upper extremity
Condition
SNOMED
44782766
Acute deep venous thrombosis of upper extremity as complication of procedure
Condition
SNOMED
37171353
Acute ischemia of colon due to thrombosis of mesenteric vein
Condition
SNOMED
37170675
Acute ischemia of small intestine due to thrombosis of mesenteric vein
Condition
SNOMED
4120091
Acute massive pulmonary embolism
Condition
SNOMED
45768439
Acute pulmonary embolism
Condition
SNOMED
45768888
Acute pulmonary thromboembolism
Condition
SNOMED
762048
Acute thrombosis of left subclavian vein
Condition
SNOMED
45757410
Acute thrombosis of mesenteric vein
Condition
SNOMED
762049
Acute thrombosis of right subclavian vein
Condition
SNOMED
36712892
Acute thrombosis of splenic vein
Condition
SNOMED
44782762
Acute thrombosis of subclavian vein
Condition
SNOMED
4179911
Axillary vein thrombosis
Condition
SNOMED
37109253
Bilateral acute deep vein thrombosis of femoral veins
Condition
SNOMED
618678
Bilateral acute deep venous thrombosis of peroneal veins
Condition
SNOMED
609003
Bilateral deep femoral vein thrombophlebitis
Condition
SNOMED
3179900
Bilateral deep vein thromboses
Condition
Nebraska Lexicon
40478951
Bilateral deep vein thrombosis of lower extremities
Condition
SNOMED
609002
Bilateral femoral vein thrombophlebitis
Condition
SNOMED
608965
Bilateral iliac vein thrombophlebitis
Condition
SNOMED
1245776
Bilateral popliteal vein thrombophlebitis
Condition
SNOMED
609006
Bilateral tibial vein thrombophlebitis
Condition
SNOMED
44782732
Chronic pulmonary embolism
Condition
SNOMED
45768887
Chronic pulmonary thromboembolism
Condition
SNOMED
45771016
Chronic pulmonary thromboembolism without pulmonary hypertension

Condition
SNOMED
4042396
Deep thrombophlebitis
Condition
SNOMED
4046884
Deep vein thrombosis of leg related to air travel
Condition
SNOMED
3655221
Deep vein thrombosis of lower extremity due to intravenous drug use
Condition
SNOMED
4133004
Deep venous thrombosis
Condition
SNOMED
761013
Deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
37163011
Deep venous thrombosis of calf
Condition
SNOMED
45773536
Deep venous thrombosis of femoropopliteal vein
Condition
SNOMED
763942
Deep venous thrombosis of left lower extremity
Condition
SNOMED
1075379
Deep venous thrombosis of left posterior tibial vein
Condition
SNOMED
761980
Deep venous thrombosis of left upper extremity
Condition
SNOMED
443537
Deep venous thrombosis of lower extremity
Condition
SNOMED
4133975
Deep venous thrombosis of pelvic vein
Condition
SNOMED
40480555
Deep venous thrombosis of peroneal vein
Condition
SNOMED
1075377
Deep venous thrombosis of posterior tibial vein
Condition
SNOMED
4322565
Deep venous thrombosis of profunda femoris vein
Condition
SNOMED
763941
Deep venous thrombosis of right lower extremity
Condition
SNOMED
1075378
Deep venous thrombosis of right posterior tibial vein
Condition
SNOMED
761928
Deep venous thrombosis of right upper extremity
Condition
SNOMED
4207899
Deep venous thrombosis of tibial vein
Condition
SNOMED
4028057
Deep venous thrombosis of upper extremity
Condition
SNOMED
193512
Embolism and thrombosis of the renal vein
Condition
SNOMED
435565
Embolism and thrombosis of the vena cava
Condition
SNOMED
4258295
Embolism from thrombosis of vein of distal lower extremity
Condition
SNOMED
40481089
Embolism from thrombosis of vein of lower extremity
Condition
SNOMED
40479840
Embolism from thrombosis of vein of thigh
Condition
SNOMED
4119760
Iliofemoral deep vein thrombosis
Condition
SNOMED
43530934
Induced termination of pregnancy complicated by pulmonary embolism
Condition
SNOMED
762808
Infarction of lung due to embolus
Condition
SNOMED
40480461
Infarction of lung due to iatrogenic pulmonary embolism
Condition
SNOMED
4124856
Inferior mesenteric vein thrombosis
Condition
SNOMED
608964
Left iliac vein thrombophlebitis
Condition
SNOMED
602592
Left peroneal vein thrombophlebitis
Condition
SNOMED
600938
Left subclavian vein thrombophlebitis
Condition
SNOMED
37164448
Lemierre syndrome
Condition
SNOMED
4281689
Phlegmasia alba dolens
Condition
SNOMED
4284538
Phlegmasia cerulea dolens
Condition
SNOMED
3185768
Popliteal vein thrombosis
Condition
Nebraska Lexicon
4309333
Postoperative deep vein thrombosis
Condition
SNOMED
4108681
Postoperative pulmonary embolus
Condition
SNOMED
37160752
Postoperative pulmonary thromboembolism
Condition
SNOMED
1245858
Postpartum acute deep vein thrombosis
Condition
SNOMED
46285905
Provoked deep vein thrombosis
Condition
SNOMED
1244882
Pulmonary artery embolism due to foreign body
Condition
SNOMED
440417
Pulmonary embolism
Condition
SNOMED
37109911
Pulmonary embolism due to and following acute myocardial infarction
Condition
SNOMED
3655209
Pulmonary embolism due to and following ectopic pregnancy
Condition
SNOMED
3655210
Pulmonary embolism due to and following molar pregnancy
Condition
SNOMED
37016922
Pulmonary embolism on long-term anticoagulation therapy
Condition
SNOMED
43530605
Pulmonary embolism with pulmonary infarction
Condition
SNOMED
4253796
Pulmonary microemboli
Condition
SNOMED
4121618
Pulmonary thromboembolism
Condition
SNOMED
4236271
Recurrent pulmonary embolism
Condition
SNOMED
608963
Right iliac vein thrombophlebitis
Condition
SNOMED
602583
Right peroneal vein thrombophlebitis
Condition
SNOMED
600939
Right subclavian vein thrombophlebitis
Condition
SNOMED
36713113
Saddle embolus of pulmonary artery
Condition
SNOMED
35615055
Saddle embolus of pulmonary artery with acute cor pulmonale
Condition
SNOMED
40479606
Septic pulmonary embolism
Condition
SNOMED
4033521
Splenic vein thrombosis
Condition
SNOMED
4119607
Subacute massive pulmonary embolism
Condition
SNOMED
4055089
Superior mesenteric vein thrombosis
Condition
SNOMED
4230403
Thrombophlebitis of axillary vein
Condition
SNOMED
4069561
Thrombophlebitis of deep femoral vein
Condition
SNOMED
761831
Thrombophlebitis of deep vein of bilateral lower limbs
Condition
SNOMED
761830
Thrombophlebitis of deep vein of left lower limb
Condition
SNOMED
761808
Thrombophlebitis of deep vein of left upper limb
Condition
SNOMED
761832
Thrombophlebitis of deep vein of right lower limb
Condition
SNOMED
761809
Thrombophlebitis of deep vein of right upper limb
Condition
SNOMED
4221821
Thrombophlebitis of deep veins of lower extremity
Condition
SNOMED
440750
Thrombophlebitis of deep veins of upper extremities
Condition
SNOMED
4203618
Thrombophlebitis of femoropopliteal vein
Condition
SNOMED
4176614
Thrombophlebitis of iliac vein
Condition
SNOMED
764715
Thrombophlebitis of internal jugular vein
Condition
SNOMED
608904
Thrombophlebitis of left axillary vein
Condition
SNOMED
761821
Thrombophlebitis of left deep femoral vein
Condition
SNOMED
761819
Thrombophlebitis of left femoral vein
Condition
SNOMED
609000
Thrombophlebitis of left popliteal vein
Condition
SNOMED
609005
Thrombophlebitis of left tibial vein
Condition
SNOMED
4318407
Thrombophlebitis of mesenteric vein
Condition
SNOMED
608903
Thrombophlebitis of right axillary vein
Condition
SNOMED
761820
Thrombophlebitis of right deep femoral vein
Condition
SNOMED
761818
Thrombophlebitis of right femoral vein
Condition
SNOMED
609001
Thrombophlebitis of right popliteal vein
Condition
SNOMED
609004
Thrombophlebitis of right tibial vein
Condition
SNOMED
4205652
Thrombophlebitis of subclavian vein
Condition
SNOMED
4110339
Thrombophlebitis of the anterior tibial vein
Condition
SNOMED
4111868
Thrombophlebitis of the common iliac vein
Condition
SNOMED
4110343
Thrombophlebitis of the external iliac vein
Condition
SNOMED
439314
Thrombophlebitis of the femoral vein
Condition
SNOMED
4109877
Thrombophlebitis of the internal iliac vein
Condition
SNOMED
4112171
Thrombophlebitis of the popliteal vein
Condition
SNOMED
4112172
Thrombophlebitis of the posterior tibial vein
Condition
SNOMED
4250765
Thrombophlebitis of tibial vein
Condition
SNOMED
42538533
Thrombosis of iliac vein
Condition
SNOMED
44811347
Thrombosis of internal jugular vein
Condition
SNOMED
765049
Thrombosis of left peroneal vein
Condition
SNOMED
4317289
Thrombosis of mesenteric vein
Condition
SNOMED
4203836
Thrombosis of subclavian vein
Condition
SNOMED
4175649
Thrombosis of the popliteal vein
Condition
SNOMED
4153353
Traumatic thrombosis of axillary vein
Condition
SNOMED
46285904
Unprovoked deep vein thrombosis
Condition
SNOMED
37163265
Venous thromboembolism due to thrombosis of vein of lower limb
Condition
SNOMED

Table S4. List of concepts used to define pelvic vein thrombosis (PVT) (concept sets included all descendants of listed concepts).
Concept ID
Concept Name
Domain
Vocabulary
762148
Acute deep vein thrombosis of bilateral iliac veins
Condition
SNOMED
35616028
Acute deep vein thrombosis of left iliac vein
Condition
SNOMED
35616027
Acute deep vein thrombosis of right iliac vein
Condition
SNOMED
765155
Acute deep venous thrombosis of bilateral ileofemoral veins
Condition
SNOMED
761461
Acute deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
762015
Acute deep venous thrombosis of ileofemoral vein of left leg
Condition
SNOMED
765541
Acute deep venous thrombosis of ileofemoral vein of right lower extremity
Condition
SNOMED
761462
Acute deep venous thrombosis of left pelvic vein
Condition
SNOMED
44782761
Acute deep venous thrombosis of pelvic vein
Condition
SNOMED
765229
Acute deep venous thrombosis of right pelvic vein
Condition
SNOMED
608965
Bilateral iliac vein thrombophlebitis
Condition
SNOMED
765152
Chronic deep vein thrombosis of bilateral iliac veins
Condition
SNOMED
35616026
Chronic deep vein thrombosis of left iliac vein
Condition
SNOMED
761439
Chronic deep vein thrombosis of left pelvic vein
Condition
SNOMED
46271548
Chronic deep vein thrombosis of pelvic vein
Condition
SNOMED
35616025
Chronic deep vein thrombosis of right iliac vein
Condition
SNOMED
761441
Chronic deep vein thrombosis of right pelvic vein
Condition
SNOMED
765542
Chronic deep venous thrombosis of bilateral ileofemoral veins
Condition
SNOMED
761440
Chronic deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
765543
Chronic deep venous thrombosis of left ileofemoral vein
Condition
SNOMED
762016
Chronic deep venous thrombosis of right ileofemoral vein
Condition
SNOMED
761013
Deep venous thrombosis of bilateral pelvic veins
Condition
SNOMED
4133975
Deep venous thrombosis of pelvic vein
Condition
SNOMED
608964
Left iliac vein thrombophlebitis
Condition
SNOMED
4285751
Pelvic thrombophlebitis in puerperium
Condition
SNOMED
608963
Right iliac vein thrombophlebitis
Condition
SNOMED
4176614
Thrombophlebitis of iliac vein
Condition
SNOMED
4317290
Thrombophlebitis of pelvic vein
Condition
SNOMED
4111868
Thrombophlebitis of the common iliac vein
Condition
SNOMED
4110343
Thrombophlebitis of the external iliac vein
Condition
SNOMED
4109877
Thrombophlebitis of the internal iliac vein
Condition
SNOMED
42538533
Thrombosis of iliac vein
Condition
SNOMED
4319327
Thrombosis of pelvic vein
Condition
SNOMED

Table S5. List of concepts used to splanchnic vein thrombosis (SVT).
Concept ID
Concept Name
Domain
Vocabulary
37171353
Acute ischemia of colon due to thrombosis of mesenteric vein
Condition
SNOMED
37170675
Acute ischemia of small intestine due to thrombosis of mesenteric vein
Condition
SNOMED
45757410
Acute thrombosis of mesenteric vein
Condition
SNOMED
36712892
Acute thrombosis of splenic vein
Condition
SNOMED
196715
Budd-Chiari syndrome
Condition
SNOMED
4301208
Hepatic vein thrombosis
Condition
SNOMED
4124856
Inferior mesenteric vein thrombosis
Condition
SNOMED
4092406
Portal thrombophlebitis
Condition
SNOMED
199837
Portal vein thrombosis
Condition
SNOMED
4033521
Splenic vein thrombosis
Condition
SNOMED
4055089
Superior mesenteric vein thrombosis
Condition
SNOMED
4318407
Thrombophlebitis of mesenteric vein
Condition
SNOMED
4317289
Thrombosis of mesenteric vein
Condition
SNOMED

Table S6. List of concepts used to define retinal vein thrombosis (RVT).
Concept ID
Concept Name
Domain
Vocabulary
437544
Arterial retinal branch occlusion
Condition
SNOMED
3657106
Bilateral occlusion of branch retinal arteries
Condition
SNOMED
37310623
Bilateral occlusion of central retinal arteries
Condition
SNOMED
37169454
Bilateral vascular occlusion of retina of eyes
Condition
SNOMED
4336004
Branch macular artery occlusion
Condition
SNOMED
4339013
Branch retinal vein occlusion with macular edema
Condition
SNOMED
4334248
Branch retinal vein occlusion with neovascularization
Condition
SNOMED
4199035
Branch retinal vein occlusion with no neovascularization
Condition
SNOMED
437540
Central retinal artery occlusion
Condition
SNOMED
313761
Central retinal vein occlusion
Condition
SNOMED
4208221
Central retinal vein occlusion - ischemic
Condition
SNOMED
4208222
Central retinal vein occlusion - non-ischemic
Condition
SNOMED
4339010
Central retinal vein occlusion with macular edema
Condition
SNOMED
4334246
Central retinal vein occlusion with neovascularization
Condition
SNOMED
4338905
Cilioretinal artery occlusion
Condition
SNOMED
42535735
Combined occlusion by thrombus of retinal artery and retinal vein
Condition
SNOMED
4102317
Incipient occlusion of retinal vein
Condition
SNOMED
4083482
Macular branch retinal vein occlusion
Condition
SNOMED
37206377
Occlusion of branch of retinal vein of left eye
Condition
SNOMED
37206378
Occlusion of branch of retinal vein of right eye
Condition
SNOMED
37206381
Occlusion of central retinal vein of left eye
Condition
SNOMED
37206380
Occlusion of central retinal vein of right eye
Condition
SNOMED
36713329
Occlusion of left branch retinal artery
Condition
SNOMED
37207955
Occlusion of left central retinal artery
Condition
SNOMED
3657873
Occlusion of left cilioretinal artery
Condition
SNOMED
36713330
Occlusion of right branch retinal artery
Condition
SNOMED
37207895
Occlusion of right central retinal artery
Condition
SNOMED
3657872
Occlusion of right cilioretinal artery
Condition
SNOMED
4334245
Retinal artery occlusion
Condition
SNOMED
4324290
Retinal phlebitis
Condition
SNOMED
440392
Retinal vascular occlusion
Condition
SNOMED
3183076
Right branch retinal artery occlusion
Condition
Nebraska Lexicon
4216561
Thrombophlebitis of retinal vein
Condition
SNOMED
4187790
Thrombosis of retinal vein
Condition
SNOMED
3657847
Vascular occlusion of retina of left eye
Condition
SNOMED
3657848
Vascular occlusion of retina of right eye
Condition
SNOMED
312622
Venous retinal branch occlusion
Condition
SNOMED

Table S7. List of concepts used to define disseminated intravascular coagulation (DIC).
Concept ID
Concept Name
Domain
Vocabulary
37117819
Acquired purpura fulminans
Condition
SNOMED
436093
Disseminated intravascular coagulation
Condition
SNOMED
4028488
Purpura fulminans
Condition
SNOMED

ANNEX V. ENCePP checklist for study protocols
ENCePP Checklist for Study Protocols (Revision 4)
Doc.Ref. EMA/540136/2009
Adopted by the ENCePP Steering Group on 15/10/2018
Study title: DARWIN EU® - Time to onset of thromboembolic events in adults with selected types of cancer


EU PAS Register® number: Study not registered yet
Study reference number (if applicable): P4-C2-017

Section 1: Milestones
Yes
No
N/A
Section Number
1.1 Does the protocol specify timelines for




1.1.1 Start of data collection1
X


8.5
1.1.2 End of data collection2
X


8.5
1.1.3 Progress report(s)
X


5
1.1.4 Interim report(s)
X


5
1.1.5 Registration in the EU PAS Register®
X


5
1.1.6 Final report of study results.
X


5
Comments:


Section 2: Research question
Yes
No
N/A
Section Number
2.1 Does the formulation of the research question and objectives clearly explain:




2.1.1 Why the study is conducted? (e.g. to address an important public health concern, a risk identified in the risk management plan, an emerging safety issue)
X


6
2.1.2 The objective(s) of the study?
X


7
2.1.3 The target population? (i.e. population or subgroup to whom the study results are intended to be generalised)
X


8.3
2.1.4 Which hypothesis(-es) is (are) to be tested?


X

2.1.5 If applicable, that there is no a priori hypothesis?


X

Comments:


Section 3: Study design
Yes
No
N/A
Section Number
3.1 Is the study design described? (e.g. cohort, case-control, cross-sectional, other design)
X


8.1
3.2 Does the protocol specify whether the study is based on primary, secondary or combined data collection?
X


8.4
3.3 Does the protocol specify measures of occurrence? (e.g., rate, risk, prevalence)
X


8.8.3
3.4 Does the protocol specify measure(s) of association? (e.g. risk, odds ratio, excess risk, rate ratio, hazard ratio, risk/rate difference, number needed to harm (NNH))


X

3.5 Does the protocol describe the approach for the collection and reporting of adverse events/adverse reactions? (e.g. adverse events that will not be collected in case of primary data collection)


X

Comments:


Section 4: Source and study populations
Yes
No
N/A
Section Number
4.1 Is the source population described?
X


Annex I
4.2 Is the planned study population defined in terms of:




4.2.1 Study time period
X


8.5
4.2.2 Age and sex
X


8.3
4.2.3 Country of origin
X


8.4
4.2.4 Disease/indication
X


8.3
4.2.5 Duration of follow-up
X


8.2
4.3 Does the protocol define how the study population will be sampled from the source population? (e.g. event or inclusion/exclusion criteria)
X


8.3
Comments:


Section 5: Exposure definition and measurement
Yes
No
N/A
Section Number
5.1 Does the protocol describe how the study exposure is defined and measured? (e.g. operational details for defining and categorising exposure, measurement of dose and duration of drug exposure)


X

5.2 Does the protocol address the validity of the exposure measurement? (e.g. precision, accuracy, use of validation sub-study)


X

5.3 Is exposure categorised according to time windows?


X

5.4 Is intensity of exposure addressed?
(e.g. dose, duration)


X

5.5 Is exposure categorised based on biological mechanism of action and taking into account the pharmacokinetics and pharmacodynamics of the drug?


X

5.6 Is (are) (an) appropriate comparator(s) identified?


X

Comments:


Section 6: Outcome definition and measurement
Yes
No
N/A
Section Number
6.1 Does the protocol specify the primary and secondary (if applicable) outcome(s) to be investigated?
X


8.6.2
6.2 Does the protocol describe how the outcomes are defined and measured?
X


8.6.2
6.3 Does the protocol address the validity of outcome measurement? (e.g. precision, accuracy, sensitivity, specificity, positive predictive value, use of validation sub-study)
X


9
6.4 Does the protocol describe specific outcomes relevant for Health Technology Assessment? (e.g. HRQoL, QALYs, DALYS, health care services utilisation, burden of disease or treatment, compliance, disease management)


X

Comments:


Section 7: Bias
Yes
No
N/A
Section Number
7.1 Does the protocol address ways to measure confounding? (e.g. confounding by indication)


X

7.2 Does the protocol address selection bias? (e.g. healthy user/adherer bias)
X


9
7.3 Does the protocol address information bias? (e.g. misclassification of exposure and outcomes, time-related bias)
X


9
Comments:


Section 8: Effect measure modification
Yes
No
N/A
Section Number
8.1 Does the protocol address effect modifiers? (e.g. collection of data on known effect modifiers, sub-group analyses, anticipated direction of effect)


X

Comments:


Section 9: Data sources
Yes
No
N/A
Section Number
9.1 Does the protocol describe the data source(s) used in the study for the ascertainment of:




9.1.1 Exposure? (e.g. pharmacy dispensing, general practice prescribing, claims data, self-report, face-to-face interview)


X

9.1.2 Outcomes? (e.g. clinical records, laboratory markers or values, claims data, self-report, patient interview including scales and questionnaires, vital statistics)
X


Annex I
9.1.3 Covariates and other characteristics?
X


Annex I
9.2 Does the protocol describe the information available from the data source(s) on:




9.2.1 Exposure? (e.g. date of dispensing, drug quantity, dose, number of days of supply prescription, daily dosage, prescriber)


X

9.2.2 Outcomes? (e.g. date of occurrence, multiple event, severity measures related to event)
X


Annex I
9.2.3 Covariates and other characteristics? (e.g. age, sex, clinical and drug use history, co-morbidity, co-medications, lifestyle)
X


Annex I
9.3 Is a coding system described for:




9.3.1 Exposure? (e.g. WHO Drug Dictionary, Anatomical Therapeutic Chemical (ATC) Classification System)


X

9.3.2 Outcomes? (e.g. International Classification of Diseases (ICD), Medical Dictionary for Regulatory Activities (MedDRA))
X


Annex I
9.3.3 Covariates and other characteristics?
X


Annex I
9.4 Is a linkage method between data sources described? (e.g. based on a unique identifier or other)


X

Comments:


Section 10: Analysis plan
Yes
No
N/A
Section Number
10.1 Are the statistical methods and the reason for their choice described?
X


8.8.3
10.2 Is study size and/or statistical precision estimated?


X

10.3 Are descriptive analyses included?
X


8.8.3
10.4 Are stratified analyses included?
X


8.8.3
10.5 Does the plan describe methods for analytic control of confounding?


X

10.6 Does the plan describe methods for analytic control of outcome misclassification?


X

10.7 Does the plan describe methods for handling missing data?
X


8.8.3
10.8 Are relevant sensitivity analyses described?


X

Comments:


Section 11: Data management and quality control
Yes
No
N/A
Section Number
11.1 Does the protocol provide information on data storage? (e.g. software and IT environment, database maintenance and anti-fraud protection, archiving)
X


Annex III
11.2 Are methods of quality assurance described?
X


Annex III
11.3 Is there a system in place for independent review of study results?
X


Annex III
Comments:


Section 12: Limitations
Yes
No
N/A
Section Number
12.1 Does the protocol discuss the impact on the study results of:




12.1.1 Selection bias?
X


9
12.1.2 Information bias?
X


9
12.1.3 Residual/unmeasured confounding?
(e.g. anticipated direction and magnitude of such biases, validation sub-study, use of validation and external data, analytical methods).
X


9
12.2 Does the protocol discuss study feasibility? (e.g. study size, anticipated exposure uptake, duration of follow-up in a cohort study, patient recruitment, precision of the estimates)
X


8.7
Comments:


Section 13: Ethical/data protection issues
Yes
No
N/A
Section Number
13.1 Have requirements of Ethics Committee/ Institutional Review Board been described?
X


Annex III
13.2 Has any outcome of an ethical review procedure been addressed?


X

13.3 Have data protection requirements been described?
X


Annex III
Comments:


Section 14: Amendments and deviations
Yes
No
N/A
Section Number
14.1 Does the protocol include a section to document amendments and deviations?
X


4
Comments:


Section 15: Plans for communication of study results
Yes
No
N/A
Section Number
15.1 Are plans described for communicating study results (e.g. to regulatory authorities)?
X


Annex III
15.2 Are plans described for disseminating study results externally, including publication?
X


Annex III
Comments:



ANNEX VI. Glossary
Additional definitions are available in the EMA Glossary of terms https://www.ema.europa.eu/en/about-us/glossaries.
Aggregated Data
Data collected and combined from multiple sources to generate summary information, typically anonymised.
Benefit-Risk Assessment
Evaluation of the positive therapeutic effects of a medicine compared to its risks (e.g., side effects).
Common Data Model (CDM)
A standardized data structure that enables data from multiple sources to be harmonized, making analysis consistent and reproducible. DARWIN EU® utilises the OMOP CDM maintained by the OHDSI community.
Complex Studies (C3)
Studies requiring the development or customisation of specific study designs, protocols, and Statistical Analysis Plans (SAPs), with extensive collection or extraction of data. Examples include etiological studies measuring the strength and determinants of an association between an exposure and the occurrence of a health outcome in a defined population considering sources of bias, potential confounding factors, and effect modifiers.
Coordination Centre (CC)
The central hub responsible for managing and overseeing the activities within DARWIN EU®. It is based at Erasmus University Medical Centre in Rotterdam, the Netherlands.
Data Access
The process of obtaining permission to use specific datasets for regulatory or scientific studies.
Data Quality Framework
A set of standards and procedures to ensure accuracy, completeness, timeliness, and consistency of data used in DARWIN EU®.
Data Source
A database or repository of structured health-related data, such as electronic health records (EHRs), insurance claims, or registries.
DARWIN EU®
The European Medicines Agency's (EMA) federated network of real-world data sources designed to generate evidence to support regulatory decision-making.
EMA (European Medicines Agency)
The regulatory body responsible for the evaluation and supervision of medicinal products in the EU, overseeing DARWIN EU®.
Evidence Generation
The process of analysing real-world data to produce scientific information that can inform healthcare or regulatory decisions.
Federated Network
A data infrastructure where data remain at their original location but can be analysed in a harmonised way across multiple partners using a common model and tools.
GDPR (General Data Protection Regulation)
The EU regulation governing the protection of personal data and privacy, crucial to how DARWIN EU® handles health data.
Health Technology Assessment (HTA)
A systematic evaluation of properties and impacts of health technology, often using DARWIN EU® data to support assessments.
Metadata
Descriptive information about a data source (e.g., its content, quality, and structure), essential for identifying relevant databases in DARWIN EU® studies.
Off-the-Shelf Studies (OTS)
Studies for which a standard protocol per study/analysis type and standardised analytics may be developed and applied or adapted, typically relating to a descriptive research question. This includes studies on disease epidemiology, for example, the estimation of the prevalence or incidence of health outcomes in defined time periods and population groups, or drug utilisation studies at the population or individual level.
OHDSI (Observational Health Data Sciences and Informatics)
An open-science collaborative community that develops tools and standards (including the OMOP CDM) to enable large-scale analytics of observational health data. OHDSI provides the technical and scientific foundation for DARWIN EU®’s analytical ecosystem.
Patient-Level Data
Data related to individuals, de-identified, used for longitudinal or detailed analyses.
OMOP (Observational Medical Outcomes Partnership)
A common data model (CDM) that standardises the structure and content of observational healthcare data, enabling systematic analysis across disparate datasets. DARWIN EU® uses the OMOP CDM to ensure interoperability and consistency in real-world evidence generation.
Real-World Data (RWD)
Data relating to individual health status or healthcare delivery that is collected from routine clinical practice rather than from randomised controlled trials.
Real-World Evidence (RWE)
Clinical evidence derived from the analysis of RWD, used to inform decisions by regulators, payers, or clinicians.
Regulatory Decision-Making
The process by which authorities like EMA assess data to authorise, monitor, or modify the use of medicines in the EU.
Routine Repeated Studies (RR)
Studies that are either Off-the-Shelf or Complex studies repeated on a regular basis, following the same protocol and study code, but with updated data and/or different data partners.
Study Protocol
A detailed plan describing how a specific real-world study will be conducted, including objectives, design, data sources, and analyses.
Very Complex Studies (C4)
Studies which cannot rely only on electronic health care databases, or which would require complex methodological work, for example, due to the occurrence of events that cannot be defined by existing diagnosis codes, including events that do not yet have a diagnosis code, where it may be necessary to combine a diagnosis code with other data such as results of laboratory investigations. These studies might require the collection of data prospectively, or the inclusion of new (not previously onboarded) data sources.
Name of the main author of the protocol:
Melissa Leung
Date: 10/10/2025

Signature:
M. Leung