Better reporting and greater homogeneity in outcome measures are seen in randomized trial protocols when guidelines exist

Abstract

Objective: Outcome Measures in Rheumatology promotes standardized outcome measures. No such organization exists for nephrology. We compared the reporting and homogeneity of outcome measures in registered protocols of randomized trials in rheumatology and nephrology.

Study design and setting: Data were extracted from protocols for rheumatoid arthritis or nephroprotection registered in ClinicalTrials.gov. We rated five outcome items (domain, specific measurement, specific metric, method of aggregating data, and time frame) to obtain a 5-point score. We split outcomes into clusters that could be pooled for meta-analysis, and assessed the proportion of trials and patients by cluster.

Results: We selected 75 protocols for rheumatology and 66 for nephrology. A high adjusted score for outcomes was associated with rheumatology protocols (odds ratio, 4.2; 95% confidence interval, 2.39, 7.39). We retained 13 clusters of outcomes for rheumatology, and one of one outcome (American College Rheumatology Criteria) could pool 87.1% of trials and 92.8% of patients. We retained eight clusters for nephrology, and one of four outcomes (assessing proteinuria) could pool 83.1% of trials and 44.7% of patients.

Conclusions: The reporting and homogeneity of outcomes is better in registered protocols of rheumatology than nephrology. The presence of international guidelines on outcome measurement may explain the differences.