Although clinicians are often unsure of diagnoses, they tend to underestimate the rate of diagnostic errors and frequently fail to recognize how diagnostic testing affects patients, a study published online May 15 in the Journal of General Internal Medicine shows.
“Perhaps the most striking finding was that physicians often feel unsure of diagnoses, regardless of setting or experience,” write Thilan P. Wijesekera, MD, MHS, from Yale University School of Medicine, New Haven, Connecticut, and colleagues.
In an interview with Medscape Medical News, David E. Newman-Toker, MD, PhD, professor of neurology, ophthalmology, and otolaryngology, Johns Hopkins University School of Medicine, Baltimore, Maryland, emphasized that diagnostic uncertainty is a fact of life in medicine and that diagnostic errors cannot be completely avoided.
“Nevertheless,” he said, “it is critically important that physicians not use this reality as an excuse to be nihilistic about improving diagnosis. There is ample evidence that we could (and should) be doing better than we are today in making accurate and timely diagnoses.”
Diagnosis is one of the most important tasks performed by clinicians, and the “concept of diagnostic uncertainty has gained increasing attention as providers navigate the challenges in medical decision making,” the authors say.
In 2015, the National Academy of Medicine (NAM) published its landmark report, Improving Diagnosis in Health Care. In that report, diagnostic errors were identified as common problems relating to patient safety.
With this in mind, Wijesekera and colleagues conducted a survey of residents (n = 196) and attending physicians (n = 70) from nine Connecticut internal medicine training programs to gain insight into how clinicians make diagnoses and deal with diagnostic error.
The NAM report highlighted five key factors that have a negative effect on physicians‘ ability to make correct diagnoses. Respondents in the current study identified time constraints (n = 178; 70%) during the diagnostic process as the factor that most hindered their ability to make a diagnosis.
Almost half of physicians in both inpatient (49%) and outpatient (41%) settings indicated feeling diagnostic uncertainty every day.
Despite the challenges of the diagnostic process, most respondents (inpatient, 67%; outpatient, 61%) reported that they did not consistently think about the advantages and risks of diagnostic testing for their patients.
Also, despite the high rate of diagnostic uncertainty among clinician respondents, most believed that diagnostic errors were uncommon. The majority thought they occurred once a month or less frequently (inpatient, 54%; outpatient, 60%).
This is in stark contrast with findings in the NAM report, which indicated that diagnostic errors arise in 10% to 15% of patient encounters, the authors of the current study note.
The new study found that clinicians regarded history taking (38%) and assessment (28%) as the most common origins of diagnostic error. Change in a patient’s status (attendings, 45%; residents, 34%) most commonly alerted clinicians to a diagnostic error.
“Future quality improvement and medical education interventions should be directed at improving efficiency, increasing high-value care, and emphasizing clinical skills in patient care,” Wijesekera and colleagues conclude.
Discussing the reported frequency of diagnostic uncertainty, Newman-Toker said he suspects that, “if pressed, most physicians would admit there is diagnostic uncertainty in the vast majority of encounters where a new symptom or problem is being evaluated, though that uncertainty may be resolved through the course of a hospitalization, for example.”
For more than half of the physicians to indicate that diagnostic errors occur only once a month or less frequently indicates they are not aware of the errors, Newman-Toker stressed. “This is not surprising, because we and others have reported on the problem of lack of feedback resulting in miscalibration about both one’s individual performance as well as the general group performance.
“Some of this has to do with the roles of the specific physicians surveyed,” he explained. “If one surveyed specialists or subspecialists (who tend to see patients down the line after errors are detected), one would find that almost all saw patients daily who had suffered diagnostic errors. In my subspecialty practice focused on dizziness, it was a relatively rare exception to encounter a patient in clinic who had not been misdiagnosed. Of course, specialists have the opposite problem — they don’t see the cases diagnosed correctly and treated effectively who never make it to their offices.”
Bedside clinical assessment and decision making (including history and physical examination, as identified in the current study) have been identified as the most common causes of diagnostic error in most physician surveys, malpractice claims analyses, and focused studies of error, said Newman-Toker. “I suspect this would be uniform across populations surveyed,” he said.
Overall, the new results indicate the need for raising awareness among outpatient and inpatient internal medicine physicians about both diagnostic uncertainty and diagnostic error, Newman-Toker stressed.
Although the lack of time as a risk factor for errors is a common complaint, especially among primary care physicians, he noted that there is little evidence that giving physicians more time than they currently have would reduce the diagnostic error rate. “This is an area that merits further study before implementing practice changes,” he said.
He believes the problem of inadequate history taking and physical examinations leading to diagnostic error must be addressed through a systematic set of changes to medical education and clinical practice.
Newman-Toker said that his team at Johns Hopkins recently completed a study, which is yet to be published, in which 9 hours of simulation-based training in diagnosis of one common symptom (dizziness) led first-year internal medicine interns to outperform the graduating third-year senior internal medicine residents.
“This suggests that the 2 years of internal medicine training were not terribly efficient in improving diagnosis of dizziness. I suspect that, to varying degrees, the same is true for some other common symptoms,” he said. Improving education through simulation training may thus be a fruitful avenue for improving diagnostic skills, he noted.
“It is incumbent upon all of us to heed the NAM’s call to action: ‘Improving the diagnostic process is not only possible, but it also represents a moral, professional, and public health imperative,’ ” Newman-Toker concluded.
The authors have reported no relevant financial relationships. Newman-Toker has received grants the Society to Improve Diagnosis in Medicine (SIDM); the Gordon and Betty Moore Foundation; the Patient-Centered Outcomes Research Institute; and Coverys. He has served as director for the Armstrong Institute Center for Diagnostic Excellence at Johns Hopkins University, as an unpaid editorial board member for the journal Diagnosis, and as a paid consultant in which he has reviewed medicolegal cases for plaintiff and defense firms related to misdiagnosis of neurologic conditions, including dizziness and stroke. He is the founding board member of the SIDM and receives honoraria for speaking engagements at academic institutions. He has been loaned research equipment related to the diagnosis of dizziness and stroke by GN Otometrics and Interacoustics. GN Otometrics has optioned a license for Johns Hopkins technology (diagnostic decision support algorithms related to diagnosis of stroke in patients with dizziness invented by Newman-Toker).
J Gen Intern Med. Published online May 15, 2019. Abstract