Depression affects millions worldwide with both pharmacological and psychological therapies widely applied, both with limited treatment success. Many clinical trials have been undertaken to test and compare treatments for depression. In collaboration with Medicines Discovery Catapult (MDC), we developed a comprehensive database of depression-related clinical trials, including information on the availability of additional biological samples such as blood or DNA. We systematically extracted structured and unstructured data from two major clinical trial registries: the US (https://clinicaltrials.gov/, using the Aggregated Analysis of ClinicalTrials.gov (AACT) database) and the EU Clinical trials register (https://www.clinicaltrialsregister.eu/). Summary data from clinical trial records and resulting publications were extracted on interventions, conditions, drugs, sample sizes, trial phase, status, and start and end dates, participant demographics, and sponsors. To identify trials likely to include blood samples or genetic information, we applied a semantic similarity approach using vector-based natural language processing to score trial records based on textual indicators. This methodology prioritised trials most likely to contain genetic information by measuring similarity between predefined query terms and clinical trial text fields. We identified 8,853 unique clinical trials registered between 1987 and 2024, with the majority (86%) from the AACT database. In total, 3,659 (41%) trials involved a drug intervention, 1,160 of which were testing antidepressants. Selective serotonin reuptake inhibitors (SSRIs) were the most commonly tested drug class (n=899), with escitalopram the most frequently studied drug (n=322). Other trial interventions included behavioural interventions and digital devices. SSRI trials peaked around 2009, while most trials since 2015 have been of NMDA receptor antagonists such as ketamine. Trial sponsors included universities, healthcare organisations, and pharmaceutical companies, with the latter sponsoring a higher proportion of drug intervention trials. Vector scores prioritised trials likely to have biological samples or genetic data, and initial validation confirmed the accuracy of these indicators. Participant demographic data and final sample size were incomplete for many trials. This comprehensive resource of clinical trials in depression provides a valuable overview of the landscape of trials performed since 1987. The dataset will be shared as an open-access repository on publication. It will enable researchers to identify clinical trials with genetic samples to expand pharmacogenetic studies, ultimately working toward the goal of personalised antidepressant prescribing.
Stewart et al. (2025) studied this question.