Assessing selective outcome reporting in sleep disorder trials: A registry-publication comparison study

Prespecified outcomes form the foundation of clinical trial validity. Despite federal registration requirements under the FDAAA Final Rule, discrepancies between registered and published outcomes remain common across medical fields. The extent of outcome modifications in interventional sleep disorder trials has not been systematically examined using registry version histories.

We conducted a cross-sectional registry-publication comparison of 54 interventional sleep disorder trials registered on ClinicalTrials.gov between January 2017 and December 2024. Using the platform's "History of Changes" tool, we compared earliest and most recent registry versions to identify outcome modifications. Changes were classified as high, moderate, or low impact. Corresponding publications were reviewed to assess disclosure. Predictors of modification were evaluated using Firth penalized logistic regression.

Substantive outcome modifications were present in 94.4% of trials, with a median of five modifications per trial. High-impact modifications affecting primary outcomes occurred in 48.1% of trials. Changes most frequently involved clarification or increased specification (61.1%), outcome additions (57.4%), and removals (40.7%). The majority of modifications occurred after primary completion (60.8%) or publication (37.3%). Disclosure was rare: 3.9% of trials acknowledged modifications in the registry and 5.9% in the corresponding publication. Industry funding was associated with lower odds of high-impact modification (OR 0.2, 95% CI 0.0-0.7; p = 0.01).

Outcome modifications are nearly universal in sleep disorder trials and are rarely disclosed. Strengthening preregistration practices, editorial oversight, and registry transparency may improve the reliability of the sleep medicine evidence base.