Using Person-Centered Analytic Approaches as Primary Outcomes in Postoperative Pain and Recovery Research

Recovery from pain after surgery is usually studied in a cross-sectional manner. This lecture discusses how this method provides information only on the small number of individuals with continuing pain and fails to quantify the burden of pain over time since surgery in everyone. It describes how more frequent assessments of pain after surgery can be acquired and analyzed using growth curve models appropriate for person-centered forecasting and latent class analysis. The increased precision and statistical power of this approach compared to classic repeated measure ANOVA methods when examining interventions to speed recovery after surgery will also be discussed. Data from two recent interventional clinical trials which used daily assessments for 2 months after surgery to define person-centered modeled trajectory as the primary outcome measure will be contrasted with the traditional dichotomous, cross sectional outcome. Finally, challenges to gathering daily data in this research context and approaches to overcome those challenges will be presented.