A Smaller Sample Size Is Associated With Lower Rates of Reporting of Harms in Randomized Controlled Trials Cited in the AAOS Clinical Practice Guidelines for Anterior Cruciate Ligament Injuries

Abstract

Background:

Randomized controlled trials (RCTs) significantly influence clinical decision-making, necessitating comprehensive reporting of trial outcomes. However, previous studies have demonstrated that reporting of harms among RCTs is often inadequate.

Purpose/Hypothesis:

The purpose of this study was to evaluate reporting of harms among RCTs cited within the American Academy of Orthopaedic Surgeons (AAOS) clinical practice guidelines (CPGs) for anterior cruciate ligament (ACL) injuries, with an emphasis on study characteristics associated with adherence. We hypothesized that >50% of these trials would have incomplete adherence to the checklist items outlined in the Consolidated Standards of Reporting Trials (CONSORT) extension for harms and that a larger sample size would be associated with greater compliance.

Study Design:

Cross-sectional study.

Methods:

We identified RCTs cited in the AAOS CPGs for the management of ACL injuries and assessed compliance with the 18-item CONSORT extension for harms checklist. Descriptive statistics were used to summarize overall adherence to the checklist items, and linear regression analysis evaluated adherence over time.

Results:

Our analysis included 116 RCTs, the majority of which were single-center studies (81.0%). Most trials had sample sizes of 50-100 patients (45.7%) or 101-500 patients (37.9%). On average, trials adhered to 9.25 of 18 CONSORT items (51.4%), with the most compliant study meeting 15 of 18 items (83.3%) and the least compliant meeting 3 of 18 items (16.7%). Overall, 18 RCTs (15.5%) reported ≤33% of items, and 22 RCTs (19.0%) reported ≥67% of items. Studies with >500 patients had significantly higher adherence compared with studies with <50 patients (P = .046). There was no significant difference in adherence based on the disclosure of funding sources (P = .85) or the implementation of blinding (P = .37). Interrupted time series regression analysis demonstrated no significant change in reporting both before (R² = 0.017; P = .62) and after (R² = 0.21; P = .16) the release of the checklist in 2004.

Conclusion:

On average, RCTs cited within the AAOS CPGs for the management of ACL injuries adhered to 51.4% of CONSORT checklist requirements, with considerable variability across studies. Improvements in reporting practices are warranted to support evidence-based treatment decisions.