Objective: To gain understanding about why a controlled intervention to reduce variability in lumbar spine imaging test use rates for low back pain patients was ineffective among internal medicine and family practice physicians in a large US health maintenance organization.
Design: We retrospectively analyzed data from focus groups that had been conducted prior to the implementation of the intervention. The physicians were asked about the factors that influence their decisions to order such tests.
Study participants: Internal medicine and family practice physicians in the intervention group. MAIN STUDY FINDINGS: Most of the variability in physicians' imaging test ordering appeared to occur in the care of patients with back pain of non-traumatic origin who had no radicular symptoms. Within that clinical context, nonclinical factors such as patient age and work status, time constraints, access problems and ambiguity about internal referral processes were important factors in physicians' decisions. Especially relevant were tensions and conflicts the physicians face as they attempted to meet conflicting role obligations in the health maintenance organization. These tensions raised issues of patient trust in their physicians and in medical care organizations, and it appeared that imaging test orders sometimes served social and symbolic functions in resolving them.
Conclusion: Our findings suggest that gaining information from focus groups prior to designing physician behavior change interventions may aid the design of more effective interventions.