Prognostic and prescriptive clinical prediction rules for low back pain: a systematic review

Abstract

Background: Populations of patients with low back pain (LBP) are generally considered to be comprised of smaller subgroups who differ in regard to their prognosis and relative response to given interventions. Identifying these subgroups and targeting interventions provides individual and population level benefits. Clinical prediction rules (CPRs) are tools that function to facilitate such stratification of care and have been developed in increasing numbers for LBP in recent years. Previous reviews on this topic have limited their scope to CPRs developed for specific interventions, a particular health profession or to a particular stage of CPR development. Consequently, it is probable that many relevant prognostic and prescriptive CPRs for LBP have not yet been identified in previous reviews and their readiness for clinical application at this time remains undetermined. Purpose(s): The purpose of this review was to identify prognostic and prescriptive CPRs relevant to the non-surgical management of LBP, and to assess their readiness for clinical application. Method(s): A systematic review of studies reporting on the derivation, validation or impact analysis of prognostic or prescriptive CPRs relevant to the non-surgical management of LBP was conducted. A sensitive electronic search strategy across 7 databases was employed in addition to handsearching and citation tracking. Following the removal of duplicates, two independent reviewers screened the titles and abstracts of identified records for eligibility. Potentially eligible studies were reviewed by full-text by both reviewers with concordance determining whether a study was included or excluded. Disagreements were resolved by consensus or if needed, by a third reviewer. The quality appraisal of included studies was conducted by two reviewers using tools relevant to each study's specific research design (QUIPS or PEDro scale), and against methodological standards for CPR development studies. Result(s): 10,005 unique records were screened for eligibility with 35 subsequently included. These studies report on the development of 30 prognostic/prescriptive CPRs for LBP with most (n = 20) derived using patients receiving a specific treatment program. Three CPRs were identified to have undergone validation in one or more studies - the 'Cassandra rule' for predicting which patients with LBP are more likely to develop significant functional restrictions, and the 5-item and 2-item Flynn manipulation CPRs that function to predict which patients being treated with lumbopelvic manipulation are likely to experience a favorable prognosis. No impact analysis studies were identified in this review. Conclusion(s): Most of the CPRs identified in this review are in their initial stage of development. The evidence considered in this review provides emergent confidence in the limited predictive performance of the 'Cassandra rule' and the 5-item Flynn manipulation CPR in similar patient populations and comparable clinical settings. Neither tool however has been evaluated for its ability to beneficially impact clinical practice. Implications: The majority of prognostic/prescriptive CPRs for LBP require further development before they can confidently be applied in clinical practice. The 'Cassandra rule' and the 5-item Flynn manipulation CPR may be applied in specific clinical circumstances with some confidence regarding their known predictive performance.