The objectives of this review was:

1. To assess the effects of alternative care models versus usual care on quality of care.

2. To assess the effects of alternative care models versus usual care on self-reported health outcomes.

Method In this Cochrane review, we included randomised trials in any language and healthcare setting. Eligible studies included adults (≥18 years) with non-specific low back pain, regardless of symptom duration. Any type of alternative care model was considered. Eligible comparators included usual care, if the same care was provided/accessed in intervention and comparator arms. Major outcomes include: 1) referral or use of any imaging modality for the lumbar spine, 2) prescription or use of opioid medication, 3) referral or use of surgery for the lumbar spine, 4) admission to hospital for non-specific LBP, 5) pain, 6) back-related function, and 7) adverse events. The primary comparison was any alternative care model compared to usual care for outcomes closest to 12 months.

Results Forty-five trials (25,427 participants) were included. There was substantial clinical diversity across alternative care models; healthcare delivery was commonly altered through coordination of care and management of care processes (n=22) or information and communication technology (n=13). Moderate certainty evidence suggests alternative care models probably result in little or no change in referral/use of imaging (RR [95% CI]: 0.93 [0.86 to 1.00]) or prescription/use of opioids (RR [95% CI]: 0.96 [0.89 to 1.04]); downgraded due to serious concerns of indirectness (diversity in outcome measures). Low certainty evidence suggests alternative care models may result in little or no clinically important improvement in pain (mean difference [95% CI]: -0.34 [-0.47 to -0.20] on 11-point scale) or back-related function (standardised mean difference [95% CI]: -0.16 [-0.21 to -0.07]); downgraded due to very serious concerns of risk of bias. We are uncertain of the effect of alternative care models on the remaining outcomes.

Conclusions While alternative care models for individuals with non-specific low back pain have been proposed as an opportunity to reduce the evidence-practice gap, our review found a paucity of evidence for their value in either improving the quality of care or improving health-related outcomes. The strength of evidence can be improved by future studies that are at low risk of performance/detection bias and the use of standardised direct quality of care measures. Due to the considerable clinical diversity in interventions and healthcare settings in this review, decision-makers need to carefully consider their local arrangements, the evidence-practice gaps that need addressing, and the barriers/enablers to evidence-based care when applying the findings of this review.