Physicians with stronger clinical knowledge are measurably less likely to order low-value imaging and other unnecessary services. That is the central finding of a study published Monday, July 13, in JAMA Internal Medicine, which matched the performance of roughly 7,100 US internists on a continuous knowledge assessment against utilization records for 25 services widely regarded as wasteful. The pattern held across every category examined: the better a doctor scored, the fewer unwarranted tests their patients received.
The stakes are enormous. By estimates cited in the paper, the United States spends about $100 billion every year on low-value care — think imaging for an uncomplicated headache. Trimming that waste without harming patients has frustrated policymakers for decades, and the new analysis points to a lever that rarely gets top billing in the debate: what the ordering physician actually knows.
How the study was conducted
The research team turned to the Longitudinal Knowledge Assessment (LKA), a program the American Board of Internal Medicine (ABIM) launched in 2022 as an alternative to the traditional high-stakes recertification exam. Participants answer 30 questions per quarter across a five-year cycle and receive immediate feedback on each response. About 92,000 physicians are currently enrolled. For this analysis, the researchers used scores from the program’s first year.
Those scores were cross-referenced with claims data for roughly 900,000 Medicare fee-for-service beneficiaries treated in outpatient settings during 2022 and 2023. The yardstick was a set of 25 well-documented low-value services, including colorectal cancer screening in patients older than 85, CT for uncomplicated rhinosinusitis, MRI for rheumatoid arthritis, and imaging for syncope, plantar fasciitis and nonspecific low back pain.
The gaps consistently favored the most knowledgeable clinicians. Patients of physicians in the top quartile were about 8% less likely to receive any of the 25 services (28.6% versus 31%). They also underwent 16% fewer unnecessary diagnostic and preventive tests (8.7% versus 10.4%), roughly 4% less imaging (13.2% versus 13.8%) and 11% fewer unwarranted cancer screenings (12.3% versus 13.8%).
Senior author Bradley Gray, PhD, a researcher at ABIM, described the analysis as “extremely rigorous” and argued that medical expertise may be a powerful lever for cutting unnecessary testing. Lead author Jonathan L. Vandergrift, MS, a senior researcher at the board, framed the open question more cautiously: whether interventions that raise knowledge — rather than merely measure it — would actually reduce wasteful ordering still needs to be tested.
What counts as low-value care
As researchers and medical societies define it, a low-value service is one whose expected benefit to the patient is small or nonexistent relative to its costs and risks. The tests themselves are not “bad.” A head CT saves lives in trauma, yet adds almost nothing for a primary headache without red flags. Clinical context is what separates a sound indication from waste.
The damage rarely stops at the first scan. Incidental findings trigger follow-up studies, biopsies and consultations in a cascade that drains resources, exposes patients to additional radiation and fuels avoidable anxiety. International campaigns such as Choosing Wisely — created by the foundation affiliated with ABIM itself — have spent more than a decade trying to break that cycle with specialty-led lists of “don’t do” recommendations.
Implications for clinical practice
For imaging departments, the findings shift part of the conversation upstream. Radiologists ultimately read whatever arrives on the worklist, but the volume of unwarranted studies is decided at the point of ordering. Appropriateness criteria, clinical decision support embedded in the electronic health record and structured feedback to referring physicians all gain weight when the evidence shows that knowledge itself shapes ordering behavior.
The study also lands at a moment when imaging is betting heavily on technology to manage demand, from triage algorithms to models that map breast cancer risk over time. Smart tools can flag a questionable order, yet the final call remains human — and it draws on precisely the kind of expertise the LKA tries to measure. Continuing education starts to look less like an administrative box to tick and more like an efficiency strategy.
There is a financial subtext as well. Health systems already struggling with aging digital infrastructure in radiology can ill afford scanner time, reading capacity and follow-up cascades spent on studies that never should have been requested. Reducing low-value orders frees capacity for the exams that genuinely change management.
Limitations and what comes next
The paper carries a limitation disclosed by the authors themselves: three of the four work at ABIM, and the fourth has received consulting fees from the board — a relevant potential conflict of interest, given that the results cast the organization’s own assessment product in a favorable light. The design is also observational, so it demonstrates association rather than causation. Physicians who order fewer unnecessary tests may simply be the ones who keep up with the literature, without one driving the other.
Even so, the message is hard for administrators to dismiss. If up-to-date clinical knowledge travels with less waste, investing in physician education doubles as a cost-containment strategy. And for the roughly 92,000 doctors already answering LKA questions every quarter, that experiment is, in a sense, already under way.
Source: Radiology Business
