American medicine runs more than 14 billion tests a year. While some tests can be lifesaving, many are used at the wrong time or on the wrong patient and are useless or even harmful.
The medical industry has spent enormous effort making more advanced tests but expended little effort learning how to use tests correctly. There is a science for how to better use tests, diagnostic stewardship, but most doctors have never heard of it.
As an infectious disease doctor, I see tests being misused daily. For example, urine cultures are one of medicine’s most overused tests. After a positive urine culture, patients without any symptoms of a urinary tract infection (UTI) usually start unnecessary and potentially harmful antibiotics. As we age and develop chronic illness, it becomes normal for bacteria to grow in the urine without causing problems. The unnecessary antibiotics that are used feed the bigger threat to our patients: antibiotic resistance, which means common antibiotics won’t work when we need them.
A research project in Michigan attempted to fix this problem in 46 hospitals across the state. They tried education first, but it had no impact. Doctors couldn’t be trained to stop using antibiotics for patients who had a positive test but no symptoms of UTI. I can understand why. When I see patients in this situation, I have to explain to them why the test doesn’t mean they have a UTI and why the antibiotics could do more harm than good. This takes time and experience. It is much easier to order the antibiotic.
What did work in Michigan, remarkably well, was stopping unnecessary testing using diagnostic stewardship. Diagnostic stewardship is changing the ordering, laboratory processing, or reporting of tests to improve the treatment of patients. In Michigan, the hospitals required doctors to enter true symptoms of a UTI to be able to order a urine culture or had the lab perform cultures only if the urine sample had evidence of inflammation. They found a drop from around 29% of patients with a positive culture being unnecessarily treated to less than 17%.
The Michigan results point to something structural. Diagnostic stewardship works because the intervention happens at a central location — say, changing how tests can be ordered in the electronic medical record or that the lab always performs tests in the best order.
Because of this, a single change can then impact hundreds of thousands of patients in a system. Diagnostic stewardship works to change how a doctor orders tests with ideas from behavioral economics, the science around influencing choice without outright denial. For example, a hospital may require a doctor to enter patient symptoms that are consistent with UTI to order a test, or a laboratory may not process a urine culture if there’s no evidence of inflammation in the sample from a different test (a prerequisite for true infection).
Critically, doctors can always get around these changes by calling the lab or entering more information for patients with a special need. But the simplest pathway, the default approach, is changed for the better.
Diagnostic stewardship isn’t limited to urine cultures but has been shown to improve how we use more costly molecular tests for influenza or Covid, new genetic blood tests, and even CT scans. Despite their groundbreaking technology, these tests still require diagnostic stewardship to use correctly, or their results can be useless, confusing, or even harmful.
Patients experience the consequences of wrong testing every day: An unnecessary CT often identifies benign lesions that trigger a surgical biopsy with complications like bleeding and infections. A misleading DNA test can treat the wrong pathogen while missing the real one. Or it can flag harmless viral fragments and delay necessary surgery. What’s more, hospitals that adopt diagnostic stewardship also reduce costs.
AI inherits this problem. Trained on human text, it replicates human patterns of over-testing. AI has potential to improve testing, but researchers at Microsoft have shown that diagnostic stewardship principles must be embedded into AI systems when they are designed, not added as an afterthought.
We all need to demand better from medicine. Doctors’ payments should reward diagnostic stewardship and penalize indiscriminate testing. Legislation could require new tests to use diagnostic stewardship from the start. The Centers for Disease Control and Prevention has called on hospitals to build diagnostic stewardship programs. Without a confirmed director, the CDC lacks the authority to promote them.
The patient who developed a resistant infection from an unnecessary antibiotic, the patient whose surgery was delayed, or the patient who had surgery after a CT false alarm: None of them needed a new drug or a new machine. They needed a smarter system. We have the science to improve testing and diagnosis. Now we need the will to use it.
Daniel Morgan, M.D., M.S., is an epidemiologist and physician in Baltimore, and vice president of the Society for Healthcare Epidemiology of America (SHEA), which advocates for diagnostic stewardship.
Source: www.statnews.com
