Predicting pain recovery in patients with acute low back pain: a validation study of a clinical prediction model

Abstract

The prognosis of acute low back pain is typically favorable; however, there is substantial variability in the clinical course of these patients. The possibility of identifying the probability of individual patients recovering at specific time points may be useful to clinicians in the decision-making process. A clinical prediction model was recently developed to predict the probability of recovery from pain in patients with acute low back pain who continue to have pain approximately one week after initially seeking care. A validation study of this prediction model was then conducted with a sample of patients from Denmark. The model presented an acceptable performance. Although the results of the previous studies are promising and the model may be useful to use in the Brazilian population, a new validation study is necessary to test the performance of the model in this population, considering the possible differences between the individuals from Denmark. The aim of this study will be to determine the usefulness of this new clinical prediction model for the Brazilian population, based on the model performance (discrimination and calibration) to predict pain recovery at one week, one month and three months in patients who still have low back pain one week after the initial consultation. We will recruit 350 patients with acute low back pain with moderate pain intensity. The outcome of interest will be number of days for recovery from pain. Discrimination will be assessed according to the C-statistic, and calibration will be assessed by graphs of agreement between the observed proportions and the predicted probabilities. (AU)