Optimal Screening for Prediction of Referral and Outcome (OSPRO) for Musculoskeletal Pain Conditions: Results From the Validation Cohort.

The sample was a convenience sample from 9 clinics, and 275 of 440 participants came from the Southeast, so the results may not apply to other regions or clinic settings. Only 279 of 440 participants completed 12-month follow-up (63.4%), which was lower than anticipated. Participants who completed follow-up were more likely to be younger, have higher income and more education, and have lower pain, lower region-specific disability and lower OSPRO-YF scores than those who did not complete follow-up. Those who did not complete follow-up were more likely to be uninsured, on disability, covered by Medicaid or worker compensation, and to have higher pain and pain-associated distress, so the models may need adjustment for these groups. All outcomes were self-reported, and 12-month comorbidity was not verified against medical records. The models did not include specific medical diagnoses, injury severity or individual treatment parameters. The OSPRO tools added small amounts of variance (0.01 to 0.07 at baseline; 0.04 to 0.07 for change), and the authors state that clinical relevance is limited. The study tested prediction, not whether using the tools changes treatment decisions or patient outcomes. Prediction of comorbidity outcomes was most affected by loss to follow-up. The OSPRO-YF total score was not weighted by its components, so it is unclear which psychological domains drive prediction.

The authors report no financial, conflicting or competing interests.

Do not read the prediction results as evidence that OSPRO screening will change care. The tools added only 0.01 to 0.07 of variance beyond baseline scores, and the study did not test whether using them changes treatment decisions or outcomes.