Does the implementation of a care pathway for patients with hip or knee osteoarthritis lead to fewer diagnostic imaging and referrals by general practitioners? A pre-post-implementation study of claims data

Abstract

Background: The Dutch care for hip and knee osteoarthritis (OA) is of good quality, but there is room for improvement regarding the efficient use of diagnostic imaging and conservative treatment. Therefore a stepped-care approach, in the shape of the care pathway 'Better exercise in osteoarthritis', was implemented to reduce the number of diagnostic imaging requested by GPS and referrals of GPS to orthopaedic care. Methods: In 2015, the pathway is implemented with the use of educational meetings, distributing guidelines and incorporating reminders in the GPS' referral application. To evaluate the effect of the pathway on the diagnostic and referral behaviour of GPS, hip and knee related health insurance claims are used together with claims of other joints and of a control region for comparison. The average number of claims and the percentage change in the post-implementation period are described. Binary logistic regression analysis is used to examine the interaction between region (intervention and control) and period (pre- A nd post-implementation). Using random sampling of patient records, information about the practical application of the pathway and the number of hip or knee arthroplasties is added. Results: In both regions, the number of diagnostic imaging decreased and the number of initial orthopaedic consultations increased during the post-implementation period. Significant interaction effects were found in knee-related diagnostics (p ≤ 0.001) and diagnostics of other joints (p = 0.039). No significant interaction effects were found in hip-related diagnostics (p = 0.060) and in initial orthopaedic consultation claims of hip (p = 0.979), knee (p = 0.281), and other joints (p = 0.464). Being referred according to the pathway had no significant effect on the probability of undergoing arthroplasty. Conclusion: The implementation of the pathway had a positive effect on GPS diagnostic behaviour related to the knee, but not to the hip. The referral behaviour of GPS to orthopaedic care needs attention for future interventions and research, since an increase (instead of a desired decrease) in the number of initial orthopaedic consultations was found. Focusing on the entire width of care for hip and knee OA could be helpful.