Has the reporting of patient-important outcomes improved in surgical trials? A meta-epidemiological study.

BACKGROUND

Clinical trials should ideally use patient-important outcomes to ensure their results are clinically relevant. We aimed to determine if the proportion of patient-important outcomes in surgical trials has changed over the last decade, and to determine whether patient-important outcomes are more likely to be specified as primary outcomes.

METHODS

The 350 most recent randomized controlled trials examining surgical interventions on humans and published in English were included. Outcomes were classified as patient-important, surrogate or laboratory using standardized definitions, along with either primary, secondary or unspecified. Mean proportions were calculated across all trials and a chi-squared test was used to detect the difference between time periods. Contingency tables were populated with each trial's outcomes, characterizing whether each outcome was patient-important or not, and whether it was specified as primary or secondary. Odds ratios were then combined in a random-effects meta-analysis to calculate a pooled odds ratio.

RESULTS

A total of 64% of all outcomes were patient-important. The mean (standard deviation) proportion of patient-important outcomes per trial was 66% (31.4), which significantly increased over the last decade, from 60% (31.6) in 2009. The mean proportion of primary outcomes which were patient-important increased from 64% (46.0) in 2009 to 77% (40.6) in 2019. Patient-important outcomes were not significantly associated with being a primary outcome and this did not change significantly over the decade.

CONCLUSION

Patient-important outcomes are still poorly represented as primary outcomes. The ongoing impact of updated reporting guidelines may improve the reporting of patient-important outcomes in surgical trials.