The Continuous Fragility Index of Statistically Significant Findings in Studies Based on High Levels of Evidence Comparing Interventions for Femoroacetabular Impingement Syndrome

Abstract

Background: Critical analysis of studies with high level of evidence has relied on the significance set by the reported P values. However, this strategy steers readers toward categorical interpretation of the data; therefore, a more comprehensive approach of data analysis is warranted. The continuous fragility index (CFI) allows for frailty interpretation of any given study’s continuous outcome results. Purpose: To calculate the CFI of high-level quality studies reporting significant continuous outcomes on comparison of treatment modalities for management of femoroacetabular impingement syndrome. Study Design: Meta-analysis; Level of evidence, 2. Methods: Three databases (PubMed, Embase, Cochrane) were queried from inception to February 2024 utilizing Boolean operators to combine variations of the following search terms: “femoroacetabular impingement, randomized controlled trials or prospective cohort.” Studies were included if level of evidence 1 or 2, and a statistically significant outcome was reported for any continuous outcome. CFI calculation was performed for all significant outcomes to obtain a study-specific mean CFI and also for the primary outcome of each study. Mean CFI was also calculated for outcomes reported to be significant in >3 studies. Additionally, multivariable linear regression was utilized for assessment of variables associated with achievement of a higher CFI. Results: Thirteen studies totaling 1316 patients were included for analysis: 11 level of evidence 1 and 2 level of evidence 2 studies. A total of 48 outcomes reaching significance were extracted, with 8 representing primary outcomes. Study-specific mean (SD) CFI was 8 (9.3), whereas primary outcome mean CFI was 12.5 (12). In 4 of the 13 studies, the number of hips lost to follow-up was greater than the study CFI. Outcome-specific mean CFI was obtained for 5 outcomes reported in >3 studies. Multivariable linear regression revealed that larger sample size and greater journal impact factor had a significant positive association with a higher overall CFI value ( P < .05). Conclusion: The mean (SD) number of patient outcome events needed to reverse the significance of a continuous outcome (ie, CFI) was 8.0 (9.3). Nearly one-third of studies had a CFI less than the reported loss to follow-up, reflecting the need for better patient compliance to attain less fragile statistical results. Larger sample size and greater journal impact factor were both predictive of a higher CFI.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 7 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.