Low back pain prognosis: Structured review of the literature

The purpose of this paper was to review and appraise pertinent articles to gain a better understanding of critical methodological issues necessary to properly design a high-quality back pain prognosis study. The review concentrated on back pain prognosis studies with epidemiologically sound designs focusing on work-disability outcomes and utilizing survival analytic methods. Nine papers were reviewed. There were few well-designed studies that achieved good scientific quality with minimal flaws. The outcomes were well defined in each paper. The age and sex characteristics of the cohorts were described in six papers and an adequate description of the study site occurred in five papers. All papers employed suitable mathematical/statistical techniques, but only one paper discussed accuracy and predictive value. No paper addressed the issue of reproducibility of the predictor variables or the final model. Most papers derived models that were clinically sensible, and the ease of use for clinicians was high. A recommended course of action for use by future patients/therapists in prognostication was rarely documented. To date, prognosis has been an inadequately studied aspect of the continuum from back injury to recovery. Researchers and clinicians interested in prognosis research need to overcome the limitations of past designs and address the methodological guidelines outlined to improve the quality of future prognosis studies