A survey of statistical methods utilized for analysis of randomized controlled trials of behavioral interventions

ObjectivesGiven the many statistical analysis options used for randomized controlled trials (RCTs) of behavioral interventions and the lack of clear guidance for analysis selection, the present study aimed to characterize the predominate statistical analyses utilized in RCTs in palliative care and behavioral research and to highlight the relative strengths and weaknesses of each of these methods as guidance for future researchers and reform. MethodsAll RCTs published between 2015 and 2021 were systematically extracted from 4 behavioral medicine journals and analyzed based on prespecified inclusion criteria. Two independent raters classified each of the manuscripts into 1 of 5 RCT analysis strategies. ResultsThere was wide variation in the methods used. The 2 most prevalent analyses for RCTs were longitudinal modeling and analysis of covariance. Application of method varied significantly by sample size. Significance of resultsEach statistical analysis presents its own unique strengths and weaknesses. The information resulting from this research may prove helpful for researchers in palliative care and behavioral medicine in navigating the variety of statistical methods available. Future discussion around best practices in RCT analyses is warranted to compare the relative impact of interventions in a more standardized way.

Automated summary

This study aimed to characterize the predominant statistical analysis methods used in RCTs in palliative care and behavioral research and highlight their relative strengths and weaknesses. The most common analyses were longitudinal modeling and analysis of covariance. The application of these methods varied significantly based on sample size. The findings of this study may guide future researchers in selecting appropriate statistical methods and promote discussion on best practices in RCT analyses.