Epidemiology and determinants of randomized controlled trials discontinued for insufficient recruitment of participants

Briel, Matthias. Epidemiology and determinants of randomized controlled trials discontinued for insufficient recruitment of participants. 2020, Doctoral Thesis, University of Basel, Faculty of Science.


Official URL: https://edoc.unibas.ch/79114/

Downloads: Statistics Overview


Randomized controlled trials (RCTs) are the best tool we have to reliably evaluate benefits and harms of interventions in health care. However, the planning and conduct of high-quality RCTs is usually resource-intense and challenging, involving various stakeholders such as clinical researchers, patients, funding agencies, research ethics committees, and, potentially, industry and regulatory agencies. Many unforeseen things can happen during the course of an RCT, and, therefore, RCTs are often not conducted as originally planned, protocols are adapted, or RCTs are prematurely discontinued. RCT discontinuation can occur, for instance, due to unexpected harm or larger than expected benefit of the tested intervention, due to futility (i.e. very low probability of finding an effect of the intervention), strategic or administrative reasons, or insufficient recruitment of participants. If an RCT is prematurely discontinued because of recruitment problems, the research question may remain unanswered, time and effort of investigators and patients as well as substantial financial resources are wasted, and collected data and lessons learned from the trial experience may remain unpublished, raising ethical concerns. Thus, it is important to know, how often RCT discontinuations due to recruitment problems actually occur, to describe their characteristics including publication, and to investigate the root-causes and potential determinants, that can help to prevent recruitment failure or to identify RCTs at high risk of discontinuation early on in the process.
In a first quantitative project, we examined 1017 RCT protocols that were approved by one of six research ethics committees (Basel, Lucerne, Lausanne, Zurich, Freiburg (Germany), and Hamilton (Canada)) between 2000 and 2003. We retrospectively examined the completion status, reported reasons for RCT discontinuation, and the publication status of these RCTs through available files at research ethics committees, comprehensive literature searches, and investigator surveys. Overall, we included 894 initiated RCTs that enrolled patients and 123 RCTs enrolling healthy volunteers. Almost two thirds of included RCTs were industry-sponsored and about one third was investigator-sponsored. We found that 28% of RCTs enrolling patients were prematurely discontinued, while discontinuation was rare in RCTs with healthy volunteers (3%). In only 38% of discontinued RCTs we found evidence that the discontinuation was reported to an ethics committee. Poor recruitment was the main reason for RCT discontinuation. In a multivariable regression analysis, investigator-sponsorship (vs industry), an acute care setting such as emergency rooms or intensive care units, and smaller planned sample size were independently associated with higher rates of discontinuation. This suggests that investigator-sponsorship, acute care setting, and a smaller planned sample size are independent risk factors for recruitment failure. The publication rate among completed RCTs was 66% and among discontinued RCTs 45%; trial discontinuation was the strongest predictor for non-publication in multivariable regression.
In three subsequent qualitative studies we aimed to identify and describe root-causes for trial discontinuation due to poor recruitment. A systematic survey of 172 publications on discontinued RCTs revealed that 76% actually reported one or more reasons for recruitment failure; most frequently reported reasons were that investigators were overoptimistic about recruitment estimates, eligibility criteria were too narrow, and investigators or patients had prejudiced views on the intervention. The overall reporting of the recruitment process, however, was mostly poor with only few reports mentioning how eligible patients were identified, the number of patients assessed for eligibility, or who recruited patients into the trial. An interview study with 39 trial investigators and other stakeholders of clinical research in Switzerland identified the decentralized health care system with many small hospitals, a lack of standardized medical records across hospitals, and the universal health coverage of patients, the limited human and financial resources in the academic setting, underdeveloped research networks, a limited collaborative attitude among clinical researchers, and a lack of accessible information for patients about ongoing clinical studies as important reasons for insufficient recruitment. Finally, we conducted a matched comparison study in which we systematically compared RCTs discontinued for poor recruitment and completed RCTs with the same underlying research question. Based on 15 separate research questions and a total of 29 discontinued and 48 completed RCTs, we found that RCTs discontinued due to poor recruitment often had narrower eligibility criteria, were investigator- rather than industry-sponsored, were associated with a higher burden for patients and recruiters, sometimes used outdated control interventions, and were often launched later in time than successfully completed RCTs compromising uncertainty about tested interventions through emerging evidence. Whether a multi- or single centre setting was advantageous for patient recruitment seemed to depend on the research context.
Stakeholders in clinical research should consider the findings of this PhD thesis and need to agree on common goals and concerted efforts, so that a sustainable improvement in patient recruitment to RCTs can be achieved.
Advisors:Bucher, Heiner C. and Schumacher , Martin
Faculties and Departments:03 Faculty of Medicine > Departement Klinische Forschung > Clinical Epidemiology and Biostatistics CEB > Klinische Epidemiologie (Bucher H)
05 Faculty of Science
UniBasel Contributors:Bucher, Heiner C.
Item Type:Thesis
Thesis Subtype:Doctoral Thesis
Thesis no:13953
Thesis status:Complete
Number of Pages:V, 153
Identification Number:
  • urn: urn:nbn:ch:bel-bau-diss139539
edoc DOI:
Last Modified:05 Mar 2021 05:30
Deposited On:04 Mar 2021 10:42

Repository Staff Only: item control page