Risk Management

What can physicians do to improve diagnostic accuracy?

ANNE M. MENKE, RN, PhD, OMIC Risk Manager

When we presented the data from our study of diagnostic error at the 2015 OMIC Forum, we asked the ophthalmologists in the audience to vote on what factor contributed the most to diagnostic errors. The choices were: atypical presentations (patient factors), physician’s cognitive process (physician factors), failure to follow up on test results (system issues), or poor communication among healthcare providers (system issues). Most ophthalmologists voted for the two types of system issues. However, when we analyzed factors in claims where the experts felt the standard of care was not met, we found instead that physician factors were the main force driving these claims (87%). System issues figured in only 13% of the claims, and patient factors in none. The correct answer to our audience response question was, therefore, the physician’s cognitive process.

Q What do we know about the cognitive process?

A There have been many studies on decision-making in general and a growing number on clinical reasoning. Researchers have identified two different ways we reason called System 1 and 2, or Fast and Slow Thinking.¹ Fast thinking draws upon our experience and is intuitive and automatic, while slow thinking is deliberative and rational. When first learning a new skill, we use mostly slow thinking and then rely upon fast thinking once the skill is mastered. A common example is learning how to drive, which becomes more and more automatic but needs to be deliberative in bad weather.

Q Is fast thinking effective during the diagnostic process?

A Without being able to move expeditiously through the diagnostic process, physicians would have difficulty seeing patients as scheduled. It is reassuring to know that fast thinking works well much of the time. Experts feel that physicians can safely rely upon it when attempting to diagnose patients with common conditions that present in typical, easily recognized ways. The very cognitive shortcuts and biases that make the process so efficient, however, can lead physicians astray. They may quickly arrive at a diagnosis and forego a more extensive exam or review of systems. Or they may rely upon an earlier thorough exam when following patients for a known condition. This happened in a number of glaucoma cases when physicians failed to regularly examine the optic nerve or compare current IOP measurements to earlier ones to check for slow changes over time. Memories of recent similar cases may also adversely influence the decision-making process. One physician in our study who had recently diagnosed giant cell arteritis gave that same diagnosis to a patient who ended up having a retinal detachment.

Q How can I tell when I need to switch to slow thinking?

A Slow thinking is obviously needed for complex presentations, and physicians readily engage it when they are puzzled by findings or unsure of a diagnosis. The prevalence of diagnostic errors indicates that physicians cannot easily determine when a more deliberative approach is needed. There are quick ways to check to see if you need to conduct a more thorough evaluation or change your diagnosis. The first is to take a “diagnostic time out” to get a second opinion from yourself. Asking yourself, “Could this be something else?” prompts you to seek alternative explanations and develop a differential diagnosis. “If I’m wrong, what don’t I want this to be?” helps take the worst case scenario into account. Challenge your diagnosis by checking for unexplained findings, test results that are surprising, or patients whose condition is worsening despite treatment. There will always be uncertainty. The goal is to establish a working diagnosis that allows you to move the diagnostic process and treatment forward and then to revise the diagnosis based upon tests results and the patient’s course.

Q How can I involve patients in the diagnostic process?

A The Institute of Medicine recently published a book-length analysis titled “Improving Diagnosis in Health Care.”² It identified the patient as key to the diagnostic process by defining diagnostic error as “the failure to (a) establish an accurate and timely explanation of the patient’s health problem(s) or (b) communicate that explanation to the patient.” The report made some suggestions. Clarify that arriving at a diagnosis is a process that takes time. Explain to the patient what you think is causing the condition and then ask if your explanation makes sense to the patient. Share uncertainty. This lesson was hard-earned by one physician in our study who did not determine the cause of vision loss. He reported after his lawsuit was settled that he now lets patients know when he has not yet found a cause for the vision loss and makes sure to discuss unexplained vision loss with colleagues.

1. 1. See, for example, Daniel Kahneman. “Thinking: fast and slow.” Farrar, Straus, and Giroux. 2011.
   2. National Academies of Sciences, Engineering, and Medicine. “Improving diagnosis in health care.” Washington DC: The National Academies Press. 2015.