Ability of Primary Care Health Databases to Assess Medicinal Products Discussed by the European Union Pharmacovigilance Risk Assessment Committee

Abstract

This study measured the exposure to different categories of medicinal products discussed by the European Union (EU) Pharmacovigilance Risk Assessment Committee from September to November 2018 in four electronic primary care health databases: IQVIA Medical Research Data‐UK, IQVIA Medical Research Data‐France, IQVIA Medical Research Data‐Germany, and Clinical Practice Research Datalink Aurum, in the entire lifespan of each database until August 31, 2018. The assessment of 83 centrally authorized products and 45 nationally authorized products showed that coverage was better for products marketed for longer duration and worse for orphan drugs. The ability to detect associations against hypothetical comparators was better for more common events and for larger effect sizes. Coverage of advanced therapies was worse for those typically administered in a specialized rather than primary care setting. This study shows that to enable better informed regulatory decisions there is a need to access complementary data sources, particularly capturing secondary care prescribing.

Apalutamide for prostate cancer: Multicentre and multidisciplinary real‐world study of 227 patients

“…Indeed, this is not surprising as the evaluation and reporting of AE are more strict within a clinical trial and patients in real‐world practice may only refer to the most severe symptoms or these may be underreported if already managed by family doctors or other specialities. 19 , 20 Our results are similar to the lower frequency of AE found in a similar real‐world study by Hussain and colleagues. 21 …”

“…Indeed, this is not surprising as the evaluation and reporting of AE are more strict within a clinical trial and patients in real-world practice may only refer to the most severe symptoms or these may be underreported if already managed by family doctors or other specialities. 19,20 Our results are similar to the lower frequency of AE found in a similar real-world study by Hussain and colleagues. 21 When analyzing the first therapy administered after apalutamide discontinuation, a predominance of Note: The numerators and denominators per variable adjusted according to the amount of valid data in this study are described in Table S2.…”

Randomized Controlled Trials Versus Real World Evidence: Neither Magic Nor Myth

“…As RWE is becoming part of the picture, there are more questions to be answered in addition to the analytics approach, including limitations of real‐world data availability (especially from specialist settings), 27 quality, and completeness, comparability of outcomes collected in the real‐world vs. more rigorously assessed or adjudicated data collected in RCTs, and matching of outcomes in RCTs, such as quality of life data, that may not be assessed in routine clinical practice.…”

Incomplete capture of apremilast in Clinical Practice Research Datalink Aurum: An example of exposure misclassification of specialty treatments in United Kingdom general practice databases

“…10 A recent study of 128 newly approved medicinal products, including apremilast, found capture to be low for many products in primary care data sources from Germany, France, and the United Kingdom. 11 Capture was particularly low for centrally authorized products which include biologics, advanced-therapies, orphan medicines, treatments for certain specified diseases and for drugs prescribed in secondary and tertiary care. This study assessed the presence of any prescription for each drug in the database.…”

“…Researchers should expect to miss exposure information for patients in all UK general practice data sources for all treatments categorized as Red or Amber including biologics, as well as oral, small molecule treatments such as apremilast. 6,11 Researchers should also assess the potential for significant underestimation of exposed person-time even in patients with a recorded prescription for these treatments and consider potential biases introduced when classification varies (e.g., Red vs. Amber) by location or over time. GP questionnaires provided additional exposure information, at significant time and cost.…”