Clinician-Level determinants of low-value acute back pain imaging in primary care

Immediate diagnostic imaging for patients experiencing acute low back pain (LBP) without red flags is generally considered low-value - inconsistent with professional guidelines, unlikely to improve clinical outcomes, and associated with multilevel harm - yet remains common. To examine factors associated with measured differences in low-value imaging utilization, we surveyed primary care clinicians from the highest- and lowest-performing (least vs. most low-value LBP imaging, respectively) primary care practices, linking clinician-reported attitudes, beliefs, and experiences with claims-based utilization. This cross-sectional survey study took place within a large health system in the southeastern United States. Primary care practices ( n = 48) were ranked based on history of low-value acute LBP imaging ordering rate determined using insurance claims data. Clinicians ( n = 93 physicians and advanced practice providers) from the 6 highest-performing and 6 lowest-performing practices were invited to participate. Survey questions focused on attitudes and beliefs related to low-value LBP imaging, perceptions of personal low-value LBP imaging performance, perceptions of drivers of and solutions for low-value LBP imaging, and clinician characteristics, including personal experience with LBP. LASSO regression identified responses associated with highest- vs. lowest-performers while controlling for clinician demographics. The survey had a 66% response rate. Across respondents, clinicians reported strong agreement with LBP imaging guidelines (9.1 ± 1.7 out of 10) and moderate trust in claims-based data (4.1 ± 1.9 out of 10). Personal history of LBP and greater years in practice were associated with lower odds of belonging to a highest-performing practice [OR = 0.70 (95%CI: 0.64–0.76) and OR = 0.87 (95%CI: 0.79–0.96), respectively]. Clinicians from lowest-performing practices were more likely to rate their performance as better than other clinicians in their practice compared with highest-performing clinicians (57% vs. 40%, p =.031). Across groups, patient demand was the most frequently identified driver of low-value LBP imaging. Clinicians from highest-performing practices more frequently recommended health system-focused strategies to reduce low-value imaging, whereas those from lowest-performing practices more frequently recommended patient education. Clinician characteristics, perceptions of performance, and preferred de-implementation strategies differed by practice performance. These findings may help inform the selection and targeting of strategies to de-implement low-value LBP imaging. This study was part of the Virginia Center for Health Innovation (VCHI) Smarter Care Virginia initiative and a participating health system’s intervention to reduce low-value back pain imaging. Both are registered on Clinicaltrials.gov: NCT04053335 (2023); NCT05413070 (2023).