Is Misdiagnosis Inevitable?

Leigh Page


March 28, 2016

In This Article

The Need for Feedback

Researchers think physicians are often not fully aware of diagnostic errors because they don't get enough feedback about them. "We don't have routine autopsies anymore, and nothing has replaced the autopsy in providing feedback," Dr Graber says.

Emergency physicians, hospitalists, laboratory pathologists, and diagnostic radiologists don't follow their patients. Even the primary care physicians who do follow them are often in the dark. "When patients discover your diagnosis was wrong, they may not tell you," Dr Graber says. "They may just go elsewhere, and you won't hear from them again."

One way to provide feedback to physicians is to create a voluntary error reporting system, which is what the Maine Medical Center in Portland has done. Robert Trowbridge, MD, a hospitalist at the medical center who helped create the program, says the program gets about two reports a week. Clinicians can anonymously report errors, and most of the reports come from someone other than the person who made the error, he says. The report is confirmed by a review of the patient's medical records.

Dr Trowbridge says setting up the program took a lot of work—"we had to convince clinicians that this was an important thing to do"—but once it was up and running, doctors accepted it. Errors are dealt with in "a nonpunitive manner," he says. "I haven't really found that doctors are defensive about this. They want to know if they've made an error."

By reviewing the cases, he says, physicians can better understand diagnostic pitfalls, and hospital officials can make system changes, such as making sure that hospitalists have access to the patient's ambulatory medical record.

Dr Trowbridge says the program takes steps to keep the information from plaintiff's attorneys in malpractice cases. After the cases are analyzed, patient and physician identifiers are removed and medical record numbers are discarded, he says.

However, on the basis of national estimates, the program captures only a small percentage of diagnostic errors. Many reformers want to create a more robust database. One method is to survey patients, but Dr Trowbridge says that sending out and collecting survey forms would be too expensive. Furthermore, patient reports can be misleading, researchers say. Some patients might be unaware that an error took place, whereas others might wrongly assume that an error was made.

Another problem is voluntary and patient-based reporting may uncover errors weeks or even years after the fact. That may be too late to avert fast-moving conditions, such as heart attack, stroke, and many cancer cases.

Instead, researchers would like to use electronic health records (EHRs) to create a robust, real-time misdiagnosis database, using electronic searches of patient records. A search algorithm, called a "trigger," identifies patients who had been treated and then had the same symptoms later—suggesting that the diagnosis had been inaccurate. Then physicians perform detailed chart reviews on these patients to confirm a misdiagnosis.

David E. Newman-Toker, MD, associate professor of neurology at Johns Hopkins University, has been using this technique to study patients with dizziness who were treated and released by emergency departments and then had a major stroke requiring hospitalization. He says he is getting closer to having a reliable measurement that can be used to monitor misdiagnoses. "These approaches are hard to build," he says, "but then are easy to maintain and use."

When the trigger identifies a case, researchers follow up with chart reviews to confirm a diagnostic error. However, physicians reviewing the same chart may disagree on what constitutes an error, so such instruments as the Safer Dx framework are being developed to standardize these reviews.

Researchers say that using triggers and chart reviews is still very much a work in progress. "Triggers will need to be developed further, because the accuracy is still quite low," says Dr Wachter at UCSF. He thinks the process could take 5-10 years, if not longer.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.