Longitudinal Data Analysis for Improving Patient Outcomes Following Hip Replacement Surgery

Abstract

Joint replacement surgery leaves many patients with postoperative pain and function limitations for extensive periods of time after surgery. This research will predict the likelihood of poor symptomatic recovery following surgery using preoperative patient data, including data on patient age, sex and comorbidities. The dataset to be analysed is total hip replacement data collected between 1989 and 2013 at the Royal Adelaide Hospital. Using the pain and function data collected repeatedly after surgery, longitudinal data analysis will be explored. The mortality information in the data will be used to explore the survival probability of patients based on different predictors using survival analysis. Repeated pain and function outcomes are modelled using mixed-effects modelling. The joint modelling of both survival and longitudinal models will be developed. Prediction methods surrounding these models will be compared to help assess the potential benefits of total hip replacement surgery for patients prior to surgery.