Value in Research: Achieving Validated Outcome Measurements While Mitigating Follow-up Cost

Background: Outcome-based research is dependent on effective follow-up, and often automated methods are augmented with costlier manual methods. The question remains as to whether the costly endeavor of achieving 80% follow-up is justified for patient-reported outcome measures (PROMs)-based research. This study evaluated associations between baseline patient characteristics and the required follow-up method, between the follow-up method and 1-year PROMs, and between baseline characteristics and 1-year PROMs for all patients compared with those for patients requiring only automated follow-up. Methods: The Orthopaedic Minimal Data Set Episode of Care (OME) database, which prospectively collects patient data and PROMs, was utilized to analyze 5,888 shoulder, hip, and knee surgical procedures at a large integrated health system. Patients were further grouped according to the method of follow-up (automated, manual, or non-responder). Associations between baseline characteristics and follow-up method were evaluated with multinomial logistic regression models. Associations of baseline characteristics with 1-year pain scores were evaluated with proportional odds logistic regression models. Results: Younger age was associated with a higher likelihood of requiring manual follow-up rather than automated follow-up for the knee surgery group (p < 0.001) and the shoulder surgery group (p < 0.001). The relative risk ratio of requiring the manual method for men undergoing a shoulder surgical procedure was 1.4 times that of women (p = 0.02). Better mental health and more education were associated with a higher likelihood of responding to automated follow-up for the hip surgery group (p < 0.001) and the knee surgery group (p = 0.001). There was no significant difference in distribution of 1-year pain scores between automated and manual follow-up methods for the knee surgery group (p = 0.51) and the shoulder surgery group (p = 0.17). There was a significant difference in 1-year pain scores for the hip surgery group (p = 0.03) that was not clinically meaningful. Conclusions: Baseline patient characteristics were significantly associated with follow-up requirements; however, there were no significant and clinically meaningful differences in 1-year PROMs. Limiting follow-up to automated methods may have the potential to transform the way that outcome-based research is designed and conducted to provide substantially better research value in large prospective cohorts.