Trial Registration

Trial registration is the practice of recording essential information about a clinical trial in a publicly accessible registry before participant enrollment begins. The registration record includes the study's objectives, design, population, interventions, outcome measures, and target sample size, and some registries also hold the analysis plan. Major registries include ClinicalTrials.gov (managed by the U.S. National Library of Medicine) and the International Standard Randomised Controlled Trial Number (ISRCTN) registry. The World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) is not itself a registry in which trials are registered: its search portal brings together records from the primary registries in the WHO Registry Network and from data providers such as ClinicalTrials.gov. Prospective registration, meaning registration before participant enrollment begins, is what the International Committee of Medical Journal Editors (ICMJE) recommends that journal editors require as a condition of considering a trial report for publication. The ICMJE's current wording is registration at or before the time of first patient consent for enrollment, and the policy has applied to trials starting enrollment after July 1, 2005. Laws and funders set their own rules, which can differ in timing: for example, under the US Food and Drug Administration Amendments Act (FDAAA 801), an applicable clinical trial must be registered on ClinicalTrials.gov no later than 21 days after the first participant is enrolled, so a trial can meet this legal deadline yet be registered too late for ICMJE journals.

Trial registration addresses two critical threats to the integrity of clinical evidence: publication bias and outcome reporting bias. Without registration, trials with negative or inconclusive results may never be published, creating a distorted literature that overestimates treatment effects. Outcome reporting bias occurs when researchers selectively report only those outcomes that showed statistically significant results, omitting prespecified outcomes that did not. By creating a public, timestamped record of the trial's intended design and outcomes, registration allows reviewers, regulators, and meta analysts to identify when trials have not been published and when published reports deviate from the original plan. This transparency is essential for evidence based medicine, systematic reviews, and regulatory decision making.

Before enrolling the first participant, investigators submit a registration record to a qualifying registry. The minimum registration dataset, defined by the WHO as its Trial Registration Data Set (version 1.3.1, 24 items), includes the trial's unique identifier, registration date, secondary identifiers, funding source, sponsor, contact information, title, research ethics approval, study conditions, interventions, inclusion and exclusion criteria, study type, study design, target sample size, recruitment status, primary and secondary outcomes, and key dates. Once registered, the trial receives a unique registration number (such as an NCT number from ClinicalTrials.gov) that is cited in all publications and communications. The record should be updated as the trial progresses, and in some jurisdictions the law requires summary results to be posted: under FDAAA 801 and its 2017 Final Rule (42 CFR Part 11), results of applicable clinical trials are generally due on ClinicalTrials.gov no later than one year after the primary completion date, and under the EU Clinical Trials Regulation (536/2014) sponsors must generally submit a summary of results to the Clinical Trials Information System (CTIS) within one year of the end of the trial. Some registries, such as ClinicalTrials.gov, also hold the protocol and statistical analysis plan, keep a public history of changes to the record, and post results including adverse events, creating a public record of the trial from inception through results.

Register all clinical trials prospectively, regardless of phase, design, or funding source. The ICMJE defines a clinical trial broadly as any research project that prospectively assigns participants to an intervention to evaluate its effects on health outcomes, encompassing both efficacy trials and comparative effectiveness studies. The ICMJE also encourages registration of studies with other designs, such as observational studies, and some journals expect systematic reviews to be registered (for example, on PROSPERO). Registration should be completed before the first participant is enrolled, and the record should show the trial's ethics approval status. If a trial was not prospectively registered, some registries allow retrospective registration, but this should be disclosed transparently in any resulting publication, and ICMJE member journals may decline to consider the manuscript.

A serious mistake is failing to register the trial before enrollment begins, which undermines the transparency benefits of registration and may prevent publication in major journals. Another common error is registering with vague or incomplete outcome specifications, such as listing "clinical outcomes" rather than specifying the exact outcome measure, time point, and analysis method. This vagueness allows post hoc outcome switching that registration is designed to prevent. Researchers sometimes fail to update the registry when protocol amendments change the primary outcome or analysis plan, creating discrepancies between the registration and the publication. Neglecting to post summary results as required by regulations (such as the FDA Amendments Act in the United States) is both a compliance violation and a contributor to results reporting bias.

A research team plans an NIH funded multicenter randomized trial of a digital mindfulness intervention for anxiety in college students. Before enrolling the first of their planned 400 participants, they register the trial on ClinicalTrials.gov, specifying the Generalized Anxiety Disorder 7 item scale (GAD 7) at 12 weeks as the primary outcome, the PHQ 9 and academic performance as secondary outcomes, and a mixed effects model with site as a random effect as the primary analysis. During the trial, they amend the protocol to add a six month follow up assessment and update the registry accordingly. After completing the trial, they post summary results on ClinicalTrials.gov within 12 months of the primary completion date, as the NIH policy on disseminating the results of NIH funded clinical trials requires. A reviewer comparing the published paper to the registration record confirms that the primary outcome, analysis, and sample size match the original registration, strengthening confidence in the results.