Modelling the effect of multimodal pain rehabilitation

Abstract

Chronic pain is a cause of suffering in a large share of the population and a leading public health problem. Multimodal pain rehabilitation (MMR) is a multidisciplinary rehabilitation method commonly used to treat chronic pain. In Sweden, MMR providers contribute patient data to a national registry (the SQRP). A Bayesian multivariate linear regression model was fitted on SQRP patient data from 2009- 2016 (n=8168) targetting patient health change from before to after treatment as responses. The psychometric instruments MPI, SF-36, HADS and NRS were used to measure health, with summarized versions of MPI and SF-36 being utilized. Several significant effects were found, principally in that unemployment, low patient belief in recovery and having constant, as opposed to periodic, pain had significant negative effects on several responses. Additionally, the results of a previous study on dimension reduction of the MPI instrument were replicated.