Blog

Can we avoid diagnostic errors without overtesting?

One of the biggest barriers to reducing overtesting is the pressure on clinicians not to miss anything. On the one hand, we tell doctors they have to reduce overuse, but at the same time we chide them for missing a rare diagnosis. Is there a way to find a balance between overtesting and undertesting — or is this even the right question to ask?

In a new piece in the Annals of Internal Medicine, Associate Director of Brigham and Women’s Center for Patient Safety Research and Practice Dr. Gordon Schiff and others argue that there isn’t a “simple linear tradeoff” between underdiagnosis and overdiagnosis, but that these problems are two sides of the same coin – the lack of cautious and careful approaches to diagnosis.  

Both underdiagnosis and overdiagnosis stem from the lack of cautious and careful approaches to diagnosis.  

In the article, a working group of clinicians, health policy experts, and educators provide recommendations for transforming our diagnostic process to reduce both missed diagnoses and overtesting. Here are a few of their recommendations:

Focus more on listening, less on testing

Patients and clinicians both perceive testing as the best way to get diagnostic information, and see testing as a demonstration of clinicians’ concern for patients’ symptoms. However, there are better ways for clinicians to show patients that they care, and to guide treatment decisions. Taking a detailed patient history and conducting a physical exam “often provide more valuable information than multiple radiologic or chemical tests,” the authors write.

How can simply listening to patients provide more information than a scan or blood test? Dr. Stephen Martin, associate professor of Family Medicine and Community Health at the University of Massachusetts Medical School, and one of the authors of the paper, explained this in an interview: 

“Listening to patients is where we learn a great deal about the arc of the symptoms they’re having, the context and severity of the symptoms, what patients are concerned about and why. You can notice and examine symptoms like rashes, ask detailed questions about where patients traveled recently, ask them about other potential stressors in their lives. A CT scan doesn’t have the answers to these questions.”

Of course, taking a detailed history and having a conversation takes more time than simply ordering a lab test, which is a barrier to careful diagnosis for clinicians who are under pressure to see as many patients as possible each day. It also helps for the clinician to have a relationship with the patient over time, so they can evaluate symptoms in context with how the patient feels normally.

Embrace uncertainty

When we’re experiencing symptoms, we want to know the cause and the cure, as soon as possible. Unfortunately, many symptoms are not connected to a particular diagnosis – for example, an acute cough or headache, in the absence of other symptoms, could be caused by a variety of things. In these cases, it might be better to offer treatments for the symptom and follow up in a few weeks if the problem isn’t better, rather than run tests to try and figure out the cause.

“It helps to know that many symptoms can self-resolve,” said Martin, “The majority of symptoms improve over 4-12 weeks, usually regardless of medical intervention.”

Uncertainty can be scary, for both patients dealing with symptoms and anxiety, and for clinicians who are responsible for patients’ health. However, using diagnostic tests to assuage anxiety from uncertainty can cause patients harm, and may not provide definitive diagnostic answers anyway. As Dr. Dan Morgan, associate professor of epidemiology at the University of Maryland School of Medicine, points out in an op-ed in The Washington Post, many doctors vastly overestimate the accuracy of diagnostic tests and the prevalence of false positives:

“Doctors especially fail to grasp how false positives work, which means they make crucial medical decisions — sometimes life-or-death calls — based on incorrect assumptions that patients have ailments that they probably don’t,” writes Morgan. “When we do this without understanding the science of risk and probability, we unacceptably increase the chances of making the wrong choice.”

Understanding the potential for harm from testing, including false positives, overtreatment, stress, and financial burden, is essential. At the same time, we need to acknowledge that our bodies are complex and we may not know the cause of every symptom, and foster honest communication about uncertainty between doctors and patients.  

“There are often unexplained symptoms, and being humble about this is an important part of being a clinician,” said Martin. 

Know the potential harms of testing

The idea that we should do tests “just to be safe” leads to unnecessary testing and discounts the potential harm of diagnostic tests. The authors point out that testing has potential direct harms such as complications from invasive tests and harms from radiation, as well as indirect harms such as treatment of false positives or cascades or additional tests. 

Additionally, diagnosing patients unnecessarily can lead to anxiety, stigmatizing labels, and lost time from medical appointments and procedures. Before conducting diagnostic tests, clinicians and patients should think through how the test results (or the diagnosis itself) would affect their treatment. In many cases, a diagnosis is not needed to select among treatment options.  

In a culture that expects doctors to have an answer and treatment for everything, recommending watchful waiting rather than conducting diagnostic tests is difficult. But having a long-term, trusting relationship with patients is extremely valuable. With a strong doctor-patient relationship, patients are more likely to trust that they could get better over time, and know that their doctor is not ignoring their concerns by not running a test. 

Additionally, if clinicians know they can follow up regularly with the patient, they can better track the patient’s symptoms over time, and will have more information to make a diagnosis or run a test if the symptoms continue. 

Quoting from the paper, “The careful use of time is a powerful incubator for diagnosis,” said Martin.