The Reverse Continuous Fragility Index of Randomized Controlled Trials Comparing Capsular Closure Versus Noncapsular Closure During Hip Arthroscopy

Abstract

Background:

The fragility index has been utilized in the literature to better characterize the limited categorical interpretation of the traditional P value in randomized controlled trials (RCTs). The reverse continuous fragility index (rCFI) aims to broaden this metric to include the analysis of nonsignificant, continuous outcome results, specifically on the topic of capsular closure versus nonclosure during hip arthroscopy.

Purpose/Hypothesis:

The purpose was to characterize the rCFI of statistically nonsignificant results in RCTs comparing capsular repair versus noncapsular repair in hip arthroscopy for the treatment of FAIS. It was hypothesized that (1) nonsignificant differences in clinical outcomes between the capsular closure and noncapsular closure groups would be statistically robust, as demonstrated by rCFI values exceeding the rate of loss to follow-up, and (2) rCFI would be greater than traditional rFI values and CFI values reported in the sports medicine literature.

Study Design:

Meta-analysis; Level of evidence, 2

Methods:

The PubMed, Cochrane, and Embase databases were queried from inception to October 2024 with combinations of the following search terms: “hip arthroscopy,” “capsule, “capsular,” and “randomized controlled trials.” Studies were included if they were randomized trials that compared capsular closure with nonclosure and had at least 1 nonsignificant outcome reported. The rCFI calculation was performed on primary nonsignificant outcomes across all studies to obtain a mean rCFI. Multivariate linear regression was performed to determine study characteristics and variables associated with higher rCFI values.

Results:

Six studies with 416 patients were included in this analysis. A total of 136 outcomes with nonsignificant results were identified across 6 studies, with 6 of these outcomes representing each study’s primary outcome. The mean rCFI across all studies was 16.333 (SD, 6.121). A mean of 11.03% (SD, 5.79%) of patients were lost to follow-up. The rCFI exceeded the number of patients lost to follow-up for all analyzed studies. Multivariate regression showed that sample size was significantly predictive of high rCFI (P = .018).

Conclusion:

Provisional assessment of rCFI suggests that the nonsignificant differences reported by RCTs comparing hip capsular closure versus nonclosure after interportal capsulotomies are relatively robust. Increased sample sizes in RCTs are associated with greater robustness of nonsignificant results.