Article Text
Abstract
Objectives There are ongoing calls for prompt, unrestricted access to Clinical Study Reports (CSRs) to support more robust evidence synthesis to enhance clinical decision making regarding medications. Between 2016–2018, the European Medicines Agency (EMA) provided access on their Clinical Data website to documents including CSRs, which had been submitted for regulatory drug approval. This study aims to describe the characteristics of available CSR documents, specifically medication and trial characteristics, as well as document length and type. For a subset of included trials which were labelled as ‘pivotal’, this study will quantify the timeliness of access to trial results from these CSRs compared with trial registry (ClinicalTrials.gov) and journal publication.
Method This is a cross-sectional analysis of drug regulatory documents published by the EMA between 2016–2018. Summary information and individual CSR files for all available medications were accessed via the EMA’s Clinical Data website. Using document filenames along with recommended structure for CSR nomenclature, individual trials in each submission were identified. Duplicate trials were excluded. Information was extracted on 142 medications for which documents were submitted, involving a total of 6,317 documents across 2,115 trials. Data were extracted including authorisation status, approval pathway and Anatomical Therapeutic Chemical (ATC) classification. Most CSRs (64.1%) had been submitted for initial marketing authorisation. Median and interquartile range (IQR) were reported for the number and length (number of pages) of documents and number of trials. For 264 pivotal trials, the trial phase, time from date of release of documents by the EMA and time to publication of results in journals and/or registries were reported.
Results There was a median of 15 (5–46) documents, 5 (2–14) trials and 9,629 (2,711–26,673) pages per submission for drug approval. The median pages per submission was lower for authorised, biosimilar, and exceptional circumstances medications, and higher for initial marketing authorisation, withdrawn procedure, conditional approval and orphan status medications. There was a median of 1 (1–4) document and 336 (21–1,192) pages per trial. Most pivotal trials, 153 (58.1%), were phase 3; 50 (19%) were phase 1 and 38 (14.4%) were phase 2. No trial registry results were identified for 64 trials (26.1%), no journal publications were identified for 41 (16.7%), and 33 trials (13.5%) had neither. For most pivotal trials with published results (94.5% n=189), journal publication/registry results were available prior to EMA publication. However, for 11 (5.5%) pivotal trials, EMA publication was the earliest source, available a median of 523 days (range 18–1,320) before a corresponding journal publication.
Conclusions The EMA Clinical Data website was found to contain lengthy documentation for a large number of trials, including detailed unpublished information. The volume of information available varied depending on the medication class and approval pathways. CSRs were the only source of data for many trials. CSRs were a timely data source, available over a year before journal or trial registry publication for some trials. Open access to currently unpublished trial data should be promoted, to support timely decision-making and better patient outcomes.