Evaluation of a risk-stratification strategy to improve primary care for low back pain: the MATCH cluster randomized trial protocol

Despite numerous options for treating back pain and the increasing healthcare resources devoted to this problem, the prevalence and impact of back pain-related disability has not improved. It is now recognized that psychosocial factors, as well as physical factors, are important predictors of poor outcomes for back pain. A promising new approach that matches treatments to the physical and psychosocial obstacles to recovery, the STarT Back risk stratification approach, improved patients’ physical function while reducing costs of care in the United Kingdom (UK). This trial evaluates implementation of this strategy in a United States (US) healthcare setting.

Six large primary care clinics in an integrated healthcare system in Washington State were block-randomized, three to receive an intensive quality improvement intervention for back pain and three to serve as controls for secular trends. The intervention included 6 one-hour training sessions for physicians, 5 days of training for physical therapists, individualized and group coaching of clinicians, and integration of the STarT Back tool into the electronic health record. This prognostic tool uses 9 questions to categorize patients at low, medium or high risk of persistent disabling pain with recommendations about evidence-based treatment options appropriate for each subgroup. Patients at least 18 years of age, receiving primary care for non-specific low back pain, were invited to provide data 1–3 weeks after their primary care visit and follow-up data 2 months and 6 months (primary endpoint) later. The primary outcomes are back-related physical function and pain severity. Using an intention to treat approach, intervention effects on patient outcomes will be estimated by comparing mean changes at the 2 and 6 month follow-up between the pre- and post-implementation periods. The inclusion of control clinics permits adjustment for secular trends. Differences in change scores by intervention group and time period will be estimated using linear mixed models with random effects. Secondary outcomes include healthcare utilization and adherence to clinical guidelines.

This trial will provide the first randomized trial evidence of the clinical effectiveness of implementing risk stratification with matched treatment options for low back pain in a United States health care delivery system.

NCT02286141. Registered November 5, 2014.