Core concepts in pharmacoepidemiology: Key biases arising in pharmacoepidemiologic studies.

Pharmacoepidemiology has an increasingly important role in informing and improving clinical practice, drug regulation, and health policy. Therefore, unrecognized biases in pharmacoepidemiologic studies can have major implications when study findings are translated to the real world. We propose a simple taxonomy for researchers to use as a starting point when thinking through some of the most pervasive biases in pharmacoepidemiology. We organize this discussion according to biases best assessed with respect to the study population (including confounding by indication, channeling bias, healthy user bias, and protopathic bias), the study design (including prevalent user bias and immortal time bias), and the data source (including misclassification bias and missing data/loss to follow up). This tutorial defines, provides a curated list of recommended references, and illustrates through relevant case examples these key biases to consider when planning, conducting, or evaluating pharmacoepidemiologic studies.

Use of prescription drug samples in the US and implications for pharmacoepidemiologic research: a systematic search of the literature.

INTRODUCTION: Free drug samples are not captured in the pharmacy claims databases used in many pharmacoepidemiologic studies, which could lead to misclassification of drug exposure status and thus bias study results. AREAS COVERED: We systematically searched the literature in PubMed/MEDLINE, Embase, and Scopus from database inception to August 2020 for studies assessing the magnitude of exposure misclassification in pharmacy claims data associated with uncaptured drug sample utilization. Our review identified five US-based studies with substantially different characteristics, contexts, methods, and results. Taken together, these studies suggest that the risk of sample-related bias may be higher for (1) studies of newly approved, patented brand-only drugs in specific classes and contexts; (2) studies of populations where sample use is common and the unexposed cohort is small; and (3) studies where the outcomes of interest are expected to be early-onset or acute, with non-constant hazards. EXPERT OPINION: In light of declining overall trends in sample use, future research on sample-related exposure misclassification should focus on delineating bias across those modern contexts where sample use remains high and optimizing bias quantification methods to create a more standardized approach. Additionally, further assessment is warranted for other sources of misclassified exposure status in claims-based pharmacoepidemiology research.

Risk Minimisation Evaluation with Process Indicators and Behavioural or Health Outcomes in Europe: Systematic Review.

BACKGROUND: European Pharmacovigilance regulatory guidance recommends the evaluation of additional risk minimisation measures (aRMMs) with process indicators and outcomes. Evaluation of both measures within the same evaluation helps to establish the relationship between the implementation of aRMMs (across process indicators) and the impact on drug safety-related outcomes. The term risk minimisation evaluation (RMEv) was used to describe a study or group of studies that assesses the effectiveness of aRMMs for one specific product. OBJECTIVES: The objective of this systematic review was to describe the characteristics and results of RMEv that include both process indicators and outcomes as well as those of studies that conform the RMEv in Europe. METHODS: We conducted a systematic search in the European Union Register of Post-Authorization Studies, PubMed and grey literature (Google and abstracts of the International Conference on Pharmacoepidemiology and Therapeutic Risk Management) to identify studies that assessed the effectiveness of aRMMs including at least one European country, from 1 January, 2011 to 12 October, 2019. Identified studies linked to one product were considered part of the product RMEv. Only RMEv that included both process indicators and outcomes (behavioural and/or health/safety outcomes) were eligible. Data were abstracted from reports, manuscripts and abstracts. RESULTS: Eighteen of 102 (18%) RMEv had both process indicators and outcomes, and were included in this review. Of the 18 RMEv, ten consisted of one study only, five of two studies, and three of three or more studies. A total of 30 studies were included within the 18 RMEv. The designs of the studies were: 19 (63%) cross-sectional surveys (47% targeted patients and 89% healthcare professionals), 17 (57%) retrospective studies (47% using pre/post approach) and 3 (10%) prospective studies. Nineteen studies included process indicators that were receipt (n = 14), use (n = 12), knowledge (n = 17) and self-reported behaviour (n = 15). Regarding outcomes, 67% of the 18 RMEv evaluated behavioural outcomes and 50% health/safety outcomes. Three of the 18 RMEv evaluated both behavioural and health/safety outcomes. For five RMEv, correlations between process indicators and outcomes were performed, two at the patient level. Results were available for 14 of the 18 RMEv. In healthcare professional surveys, the median percentage was 57% for receipt, 92% for reading, 80% for use, 77% for knowledge and 74% for behaviour. In patient surveys, the median percentage was 56% for receipt, 87% for reading, 65% for use, 47% for knowledge and 69% for behaviour. Knowledge was better in healthcare professionals than patients (p < 0.05). Of the three RMEv with a correlation analysis, only one found a positive trend for a lower occurrence of outcomes as process indicators improved, though this was not statistically significant. CONCLUSIONS: A minority of RMEv assessed both process indicators and outcomes. More RMEv require approaches that correlate process indicators and outcomes at the patient level to evaluate more comprehensively the implementation of aRMMs.

Empirical assessment of case-based methods for drug safety alert identification in the French National Healthcare System database (SNDS): Methodology of the ALCAPONE project.

OBJECTIVES: To introduce the methodology of the ALCAPONE project. BACKGROUND: The French National Healthcare System Database (SNDS), covering 99% of the French population, provides a potentially valuable opportunity for drug safety alert generation. ALCAPONE aimed to assess empirically in the SNDS case-based designs for alert generation related to four health outcomes of interest. METHODS: ALCAPONE used a reference set adapted from observational medical outcomes partnership (OMOP) and Exploring and Understanding Adverse Drug Reactions (EU-ADR) project, with four outcomes-acute liver injury (ALI), myocardial infarction (MI), acute kidney injury (AKI), and upper gastrointestinal bleeding (UGIB)-and positive and negative drug controls. ALCAPONE consisted of four main phases: (1) data preparation to fit the OMOP Common Data Model and select the drug controls; (2) detection of the selected controls via three case-based designs: case-population, case-control, and self-controlled case series, including design variants (varying risk window, adjustment strategy, etc.); (3) comparison of design variant performance (area under the ROC curve, mean square error, etc.); and (4) selection of the optimal design variants and their calibration for each outcome. RESULTS: Over 2009-2014, 5225 cases of ALI, 354 109 MI, 12 633 AKI, and 156 057 UGIB were identified using specific definitions. The number of detectable drugs ranged from 61 for MI to 25 for ALI. Design variants generated more than 50 000 points estimates. Results by outcome will be published in forthcoming papers. CONCLUSIONS: ALCAPONE has shown the interest of the empirical assessment of pharmacoepidemiological approaches for drug safety alert generation and may encourage other researchers to do the same in other databases.

Evaluating the Utility of Coarsened Exact Matching for Pharmacoepidemiology Using Real and Simulated Claims Data.

Coarsened exact matching (CEM) is a matching method proposed as an alternative to other techniques commonly used to control confounding. We compared CEM with 3 techniques that have been used in pharmacoepidemiology: propensity score matching, Mahalanobis distance matching, and fine stratification by propensity score (FS). We evaluated confounding control and effect-estimate precision using insurance claims data from the Pharmaceutical Assistance Contract for the Elderly (1999-2002) and Medicaid Analytic eXtract (2000-2007) databases (United States) and from simulated claims-based cohorts. CEM generally achieved the best covariate balance. However, it often led to high bias and low precision of the risk ratio due to extreme losses in study size and numbers of outcomes (i.e., sparse data bias)-especially with larger covariate sets. FS usually was optimal with respect to bias and precision and always created good covariate balance. Propensity score matching usually performed almost as well as FS, especially with higher index exposure prevalence. The performance of Mahalanobis distance matching was relatively poor. These findings suggest that CEM, although it achieves good covariate balance, might not be optimal for large claims-database studies with rich covariate information; it might be ideal if only a few (<10) strong confounders must be controlled.

Pharmacy students' intentions to utilize pharmacoeconomics, pharmacoepidemiology, and health outcomes in future jobs.

INTRODUCTION: The objectives of this study were to explore factors associated with pharmacy students' intentions to utilize health outcomes by: (1) understanding opinions on health outcomes, (2) understanding the likelihood of using health outcomes in different settings, and (3) predicting pharmacy students' intentions to utilize health outcomes in future jobs. METHODS: This study surveyed second-year pharmacy students over two years. The survey contained four components: the theory of planned behavior, opinions on health outcomes, the likelihood of using health outcomes in different settings, and demographics. To predict pharmacy students' intentions to utilize health outcomes in future jobs, a multiple linear regression model was used with behavioral intention as the dependent variable. RESULTS: Of the 376 second-year pharmacy students surveyed, 229 responded (60.90%). Pharmacy students had a positive attitude (mean: 0.77, SD: 0.16), high level of subjective norm (mean: 0.75, SD: 0.18), high level of perceived behavioral control (mean: 0.74, SD: 0.15), and high level of behavioral intention (mean: 0.74, SD: 0.21). They thought health outcomes were important for their future jobs (mean: 0.76, SD: 0.22), and equally important as other courses in the doctor of pharmacy curriculum (mean: 0.49, SD: 0.23). Significant predictors of utilizing health outcomes in future jobs were attitude (0.21; 95% CI: 0.03, 0.40), subjective norm (0.38; 95% CI: 0.23, 0.54), and perceived behavioral control (0.45, 95% CI: 0.27, 0.63). CONCLUSIONS: The second-year pharmacy students in the program studied had positive opinions and expressed high likelihood of applying their health outcomes knowledge and skill after graduation.

An empirical assessment of immeasurable time bias in the setting of nested case-control studies: Statins and all-cause mortality among patients with heart failure.

PURPOSE: Immeasurable time bias exaggerates drug benefits in pharmacoepidemiologic studies due to exposure misclassification that occurs due to the lack of inpatient drug data in many healthcare databases. METHODS: To estimate the magnitude of immeasurable time bias and assess potential approaches to minimize it, we conducted a nested case-control study of statin use and mortality among heart failure patients using the South Korean nationwide healthcare database, which contains both inpatient and outpatient medication data. Using both inpatient and outpatient medication data to define the gold standard exposure definition, we assessed 10 different analytical methods in which exposure was defined using outpatient medication data only. We compared different methodological approaches to reduce immeasurable time bias: restricting to nonhospitalized patients, adjusting for hospitalization, weighting by either measurable time (nonhospitalized time during 90-d period) or outpatient time, and computing the odds ratios (ORs) using 90-day cumulative probability of exposure produced by the Kaplan-Meier product-limit estimator for cases and controls. RESULTS: The three approaches that most closely approximated the gold standard (hazard ratio [HR] 1.20; 95% confidence interval [CI], 1.05-1.37) were weighting by either measurable (HR 1.09; 95% CI, 0.92-1.28) or outpatient time (HR 1.14; 95% CI, 0.96-1.34) in the unexposed or by estimating the 90-day exposure probability (HR 1.31; 95% CI, 1.11-1.51). CONCLUSION: The use of one of these three methods may be suggested as an approach to minimize immeasurable time bias in nested case-control studies.

Engaging stakeholders in pharmacoepidemiology research: Current state and recommendations.

PURPOSE: Given current efforts to enhance patient-centered care and shared decision-making, the International Society of Pharmacoepidemiology Workgroup on Patient Engagement assessed patient and other stakeholder engagement in pharmacoepidemiology research and provides recommendations for the field. METHODS: A systematic review used MEDLINE and EMBASE to identify published literature from 2005 to 2016 addressing how stakeholders-patients, caregivers, and others-assisted researchers conducting pharmacoepidemiologic research. Three pairs of Workgroup members screened titles and abstracts to select articles for full-text review and analysis. Two Workgroup members abstracted the following data: research focus, characterization and role of stakeholders, and type(s) of engagement strategy employed. Data were summarized descriptively. RESULTS: We identified 5717 references for abstract screening. Of these, 69 met the criteria for full-text screening, and 11 were selected for data abstraction. Of these 11 studies, seven focused on the development of a research agenda and eight had stakeholders react or advise on an aspect of the study. Although patients were the most commonly identified stakeholders, advocacy groups and health care professionals were also frequently identified. Some studies reported the engagement of other stakeholders, including local government or policy experts. Engagement strategies varied, with five studies using more than one strategy. Studies often did not indicate the involvement of stakeholders in developing the study design or with implementation. CONCLUSIONS: Currently, few pharmacoepidemiology publications mention patient or other stakeholder engagement in the design, analysis, or reporting of research. This suggests that there are opportunities to expand stakeholder engagement and/or increase the transparency of reporting stakeholder engagement.

The value of a health insurance database to conduct pharmacoepidemiological studies in oncology.

Some concerns have emerged about the evidence of benefits on survival outcomes or quality of life of new anticancer drugs. In parallel, the decreased cancer mortality leads to an increased number of patients exposed to cancer treatment-related consequences. In this context, pharmacoepidemiology is crucial to assess anticancer drug use, effectiveness and safety in real life conditions. We aimed to describe strengths, limitations and considerations associated with the use of the French national health insurance database (système national des données de santé [SNDS]) to conduct pharmacoepidemiological studies in oncology. The SNDS represents a powerful tool in pharmacoepidemiology owing to its extensive coverage, accurate description and quantification of drug exposure and individual data on patients. The main limitations of this database ensue from the administrative nature resulting in technical difficulties in its management and gaps in availability of data. Another limitation is the lack of accurate identification of diseases, comorbidities or outcomes and potential confounding with notably the lack of data regarding cancer stage, prognosis or risk factors. Finally, the accurate identification of the nature of chemotherapy received by patients is sometimes complex. To minimize these limitations, several approaches and statistical methods could be used as highlighted by national or international initiatives. First, the SNDS may be linked with cancer registry or clinical data. Then, several data sources could be combined using meta-analytical methods. The development of methodological tools and the use of standardized methods are crucial to enhance the quality of studies that can impact clinical practice and guide public decision. Pharmacoepidemiological approaches and pharmacovigilance represent an important cornerstone in oncology for signal detection or long-term follow up of cancer patients. In this context, validated methods to identify cancer patients and to describe chemotherapy regimens within these data should be promoted and remain too scarce despite international guidelines. Moreover, limits and strength of each data sources should be systematically discussed according to the research question. Optimized and framed use of claims database represents a future challenge in onco-pharmacoepidemiology.

Use of propensity score and disease risk score for multiple treatments with time-to-event outcome: a simulation study.

Propensity score (PS) and disease risk score (DRS) are often used in pharmacoepidemiologic safety studies. Methods of applying these two balancing scores are extensively studied in binary treatment settings. However, the use of PS and DRS is not well understood in the case of non-ordinal multiple treatments. Some PS methods of multiple treatments have been implemented since the theoretical establishment. Nevertheless, most of the work applies to continuous or binary outcomes. Little work has been done for time-to-event outcomes. In this study, we extend the application of the PS and DRS methods to time-to-event outcomes in multiple treatment settings. The analytical approaches include weighing, matching, stratification, and regression. Simulation studies with rare event rates are conducted to evaluate the performances of different methods. Different treatment-covariates and outcome-covariates strength of associations are considered. Additionally, the impacts of imbalanced designs and large or limited PS overlaps are investigated on various analytical approaches. We found that the inverse probability treatment weighting with bootstrap variance estimator, the generalized PS matching, and the Cox regression estimated DRS in full cohort generally performed well in multiple treatment settings. This study aims to provide additional guidance for researchers on PS and DRS analyses in pharmacoepidemiologic observational studies.

Conditional validation sampling for consistent risk estimation with binary outcome data subject to misclassification.

PURPOSE: Misclassification of a binary outcome can introduce bias in estimation of the odds-ratio associated with an exposure of interest in pharmacoepidemiology research. It has been previously demonstrated that utilizing information from an internal randomly selected validation sample can help mitigate this bias. METHODS: Using a Monte Carlo simulation-based approach, we study the properties of misclassification bias-adjusted odds-ratio estimators in a contingency table setting. We consider two methods of internal validation sampling; namely, simple random sampling and sampling conditional on the original (possibly incorrect) outcome status. Additional simulation studies are conducted to investigate these sampling approaches in a multi-table setting. RESULTS: We demonstrate that conditional validation sampling, across a range of subsampling fractions, can produce better estimates than those based on an unconditional simple random sample. This approach allows for greater flexibility in the chosen categorical composition of the validation data, as well as the potential for obtaining a more efficient estimator of the odds-ratio. We further demonstrate that this relationship holds for the Mantel-Haenszel misclassification bias-adjusted odds-ratio in stratified samples. Recommendations for the choice of validation subsampling fraction are also provided. CONCLUSIONS: Careful consideration when choosing the sampling scheme used to draw internal validation samples can improve the properties of the outcome misclassification bias-adjusted odds-ratio estimator in a (multiple) contingency table.

Interest of pharmacoepidemiology for pharmacodynamics and analysis of the mechanism of action of drugs.

Pharmacology is often divided in separate branches, such as molecular and cellular pharmacology, pharmacokinetics, pharmacodynamics, experimental and/or preclinical pharmacology, clinical pharmacology (and therapeutics), pharmacogenetics, pharmacogenomics, pharmacovigilance, pharmacoepidemiology, pharmacoeconomics… This enumeration gives a global picture of different scientific areas, which are however dealing with the same question. Another mindset should be a global interactive and continuous approach, which could be designed as "human pharmacology". An original and attractive way to illustrate this continuous approach is to combine pharmacodynamics and pharmacovigilance and/or pharmacoepidemiologic data. Coupling disproportionality analyses in pharmacovigilance databases or computerized health databases, with pharmacological characteristics of drugs (receptor affinity, for example) allows investigating in humans, the mechanism of adverse drug reactions. Examples of such analyses investigating the risk of movement disorders, diabetes related to psychoactive drugs, or the risk of adverse cardiac outcomes with different drugs (classical drugs or protein kinase inhibitors) are given. The increasing number of research works investigating this topic underlines the importance of this relatively new approach, which gives significant inputs for the better knowledge of drug safety.

Quantifying bias reduction with fixed-duration versus all-available covariate assessment periods.

PURPOSE: Implementing a cohort study in longitudinal healthcare databases requires looking back over some covariate assessment period (CAP) preceding cohort entry to measure confounders. We used simulations to compare fixed-duration versus all-available CAPs for confounder adjustment in the presence of differences in available baseline time between exposure groups. METHODS: We simulated cohorts of 10 000 patients with binary variables for a single confounder, exposure, and outcome. Baseline time was simulated based on the observed distribution in a claims-based comparison of statin users versus nonusers. We compared bias after measuring confounders using fixed-duration and all-available CAPs, both when exposure groups had similar and discrepant amounts of available baseline time. RESULTS: When the comparison groups had similar amounts of baseline time, an all-available CAP was less biased than a fixed-duration CAP. When baseline time differed between comparison groups, the preferable CAP approach depended on the direction of confounding and which exposure group had higher covariate sensitivity. These findings were consistent in direction across sensitivity analyses. CONCLUSION: In certain settings of differential available baseline time between exposure groups, the all-available CAP was more biased than the fixed-duration CAP. The relative directions and strengths of confounding and misclassification biases are an important consideration when choosing between a fixed-duration or all-available CAP, but they are often unknown. Therefore, we recommend comparing the amount of available baseline time between exposure groups. When there is a large discrepancy, despite appropriate design choices, we recommend a fixed-duration approach to avoid potential increases in bias because of differential data availability.

Interest of pharmacoepidemiology in pharmacovigilance: Post-authorization safety studies in regulatory pharmacovigilance activity.

During the past few decades, it has been stated that a paradigm shift has occurred in the assessment and management of patient related drug safety. Some of these changes have resulted in a significant increase in the importance of pharmacoepidemiology and its use in pharmacovigilance. For European member states, the Pharmacovigilance Risk Assessment Committee (PRAC) is responsible for assessing the protocols and results of imposed and non-imposed post-authorization safety studies (PASS). Between 2013 and 2017, the total number of PASS during this 5-years period of the different products, including protocols and results, was 1062. The number of protocols of PASS is increasing over time, except in 2017 where a 25% decrease has been observed. Whereas, PASS results steadily increased over the 5years period. Between 2014 and 2017, about 29% (n=137) of PRAC reviewed protocols were imposed. The number of imposed PASS was almost constant over time with a mean of 34.3±7.6 imposed protocols per year and 3.5±1.74 imposed results per year. The need for the implementation of PASS for pharmacovigilance regulatory activities is increasing. Nevertheless, conducting such studies remains difficult.

Methodologic limitations of prescription opioid safety research and recommendations for improving the evidence base.

PURPOSE: The ongoing opioid epidemic has claimed more than a quarter million Americans' lives over the past 15 years. The epidemic began with an escalation of prescription opioid deaths and has now evolved to include secondary waves of illicit heroin and fentanyl deaths, while the deaths due to prescription opioid overdoses are still increasing. In response, the Centers for Disease Control and Prevention (CDC) moved to limit opioid prescribing with the release of opioid prescribing guidelines for chronic noncancer pain in March 2016. The guidelines represent a logical and timely federal response to this growing crisis. However, CDC acknowledged that the evidence base linking opioid prescribing to opioid use disorders and overdose was grades 3 and 4. METHODS: Motivated by the need to strengthen the evidence base, this review details limitations of the opioid safety studies cited in the CDC guidelines with a focus on methodological limitations related to internal and external validity. RESULTS: Internal validity concerns were related to poor confounding control, variable misclassification, selection bias, competing risks, and potential competing interventions. External validity concerns arose from the use of limited source populations, historical data (in a fast-changing epidemic), and issues with handling of cancer and acute pain patients' data. We provide a nonexhaustive list of 7 recommendations to address these limitations in future opioid safety studies. CONCLUSION: Strengthening the opioid safety evidence base will aid any future revisions of the CDC guidelines and enhance their prevention impact.

Medication history assessment in research: A randomized controlled trial comparing tablet-based (eMedHAT) versus structured interview (MedHAT).

PURPOSE: Accurate capture of medication use is important for high quality research. For epidemiologic studies, medication histories are the most common measure of exposure when trying to identify associations between medications and outcomes. Concomitant medications can alter the efficacy or safety of study drugs in clinical trials. However, there are few studies evaluating the accuracy and efficiency of methods to collect these histories. The objective of this study is to compare the accuracy of medication histories collected by structured interview to histories captured using a tablet-based application. METHODS: This was a randomized controlled trial. Subjects were instructed to record all prescription medications, non-prescription medications, vitamins, and dietary supplements in a diary for 30 days. At the end of the diary collection, subjects were randomized to providing a medication history during a structured interview by a trained research assistant (MedHAT) or using a tablet-based application (eMedHAT). The accuracy of these histories was compared using an adjusted analysis. We also measured the duration of the history collection and data entry. RESULTS: A total of 111 subjects were in the MedHAT group and 109 subjects were in the eMedHAT group. Recall of medications for the 30-day period was similar for MedHAT and eMedHAT (76.9% versus 75.2%, respectively). The total time required for researchers and subjects for history collection and data entry was 16 minutes shorter for the tablet-based method. CONCLUSIONS: Tablet-based medication histories were as accurate as histories obtained by research assistants and required less time for the researcher.

From a Viewpoint of Clinical Settings: Pharmacoepidemiology as Reverse Translational Research (rTR).

Clinical pharmacology and pharmacoepidemiology research may converge in practise. Pharmacoepidemiology is the study of pharmacotherapy and risk management in patient groups. For many drugs, adverse reaction(s) that were not seen and/or clarified during research and development stages have been reported in the real world. Pharmacoepidemiology can detect and verify adverse drug reactions as reverse translational research. Recently, development and effective use of medical information databases (MID) have been conducted in Japan and elsewhere for the purpose of post-marketing safety of drugs. The Ministry of Health, Labour and Welfare, Japan has been promoting the development of 10-million scale database in 10 hospitals and hospital groups as "the infrastructure project of medical information database (MID-NET)". This project enables estimation of the frequency of adverse reactions, the distinction between drug-induced reactions and basal health-condition changes, and usefulness verification of administrative measures of drug safety. However, because the database information is different from detailed medical records, construction of methodologies for the detection and evaluation of adverse reactions is required. We have been performing database research using medical information system in some hospitals to establish and demonstrate useful methods for post-marketing safety. In this symposium, we aim to discuss the possibility of reverse translational research from clinical settings and provide an introduction to our research.

The national healthcare system claims databases in France, SNIIRAM and EGB: Powerful tools for pharmacoepidemiology.

The French health care system is based on universal coverage by one of several health care insurance plans. The SNIIRAM database merges anonymous information of reimbursed claims from all these plans, linked to the national hospital-discharge summaries database system (PMSI) and the national death registry. It now covers 98.8% of the French population, over 66 million persons, from birth (or immigration) to death (or emigration), making it possibly the world's largest continuous homogeneous claims database. The database includes demographic data; health care encounters such as physician or paramedical visits, medicines, medical devices, and lab tests (without results); chronic medical conditions (ICD10 codes); hospitalisations with ICD10 codes for primary, linked and associated diagnoses, date and duration, procedures, diagnostic-related groups, and cost coding; date but currently not cause of death. The power of the database is correlatively great, and its representativeness is near perfect, since it essentially includes the whole country's population. The main difficulty in using the database, beyond its sheer size and complexity, is the administrative process necessary to access it. Recent legislative advances are making this easier. EGB (Echantillon Généraliste de Bénéficiaires) is the 1/97th random permanent representative sample of SNIIRAM, with planned 20-year longitudinal data (10 years at this time). Access time is 1 to 3 months, but its power is less (780 000 subjects). This is enough to study common issues with older drugs but may be limited for new products or rare events.