Shortcuts: How Bias Can Undermine Diagnosis and Treatment

Posted Jun 3, 2020

Robert Morris, MD, CCHP-P

The concept of "bias" is often misunderstood as simply meaning prejudice against a person, group, or thing. In fact, biases involve an array of brain processes – shortcuts – that can lead to erroneous health care decisions and affect diagnosis and treatment. Generally, biases are unconscious, which makes them more potentially damaging.

Unconscious bias takes many forms. One of those, confirmation bias, has a particularly insidious effect on diagnostic accuracy. Confirmation bias occurs when we reach conclusions based on what we expect to find by selectively accepting some information and ignoring other information that does not fit our expectations. It’s a shortcut in thinking that can lead to disaster. Availability bias fuels confirmation bias by accepting an idea – in this case, a diagnosis – based on how “available” it is, or how easily it comes to mind. To make matters worse, if an “expert” makes a diagnosis, the rest of us tend to accept it without question, a phenomenon called expert certainty.

Corrections is especially ripe for mistakes due to bias. Many of our patients misuse alcohol and drugs; some exaggerate pain, malinger, and scheme. Unfortunately, we can come to believe this applies to all our patients.

There is a relatively new technology that can amplify the persistence or “stickiness” of an erroneous diagnosis: the electronic health record. Once a diagnosis appears in the record it tends to stay there, whether or not it is correct, and mislead the next clinician if the patient is not reevaluated.

The Case of the 'Depressed' Teenager

A young man in clinic carried the diagnosis of depression and, much to the consternation of the staff, refused referrals for depression. I was asked to talk to him and convince him to go to therapy. As we talked, he said, “I'm not depressed. I’m angry and I want help with that. But everyone ignores my request.” He lived in a dangerous neighborhood and had suffered multiple attacks at school and on the street. He was anxious and angry, and engaged in body building so he would be ready to attack anyone he perceived as a threat.

Despite that information, the resident physician continued to record in his notes that the patient was depressed and refused treatment. That was what was in the chart, so it “must be correct,” despite my request that it be removed.

I did succeed in getting “anger issues” added to the problem list. After we worked with him to get his high school degree and enrolled in a college outside his neighborhood, his “depression” – his anger and anxiety – subsided.

Several types of bias are illustrated in that example. The previous diagnostician had diagnosed depression and, rather than listen to the patient who contradicted that diagnosis, the resident continued it (expert certainty). "Everyone" knows depression is common in young people (availability bias). The resident ignored information offered by the patient, relying instead on what was in the chart (confirmation bias).

More Complexity Can Mean More Harm

More complex presentations offer a greater opportunity for bias to lead to a misdiagnosis. A 17-year-old boy in detention complained of back pain and leg weakness. His initial neurologic exam was normal, although it was difficult to perform because he was on the floor of his room. He said he couldn’t move, and he was too heavy to lift. A back X-ray was also normal. The fact that he didn’t appear to be stressed by his ailment added to a conclusion that he was looking for secondary gain. After a few weeks without improvement he was seen by the facility neurologist, who proclaimed that he was faking, despite the fact that he was now leaking urine. Staff members also reported that they had seen him standing and walking, so he was surely faking.

Several weeks later he was sent to the ER that served the facility, where an MRI revealed that an aneurysmal bone cyst in a vertebral body had eroded the posterior wall and was pressing intermittently on his spinal cord, leading to the waxing and waning of symptoms. Surgery removed the cyst and over several months his symptoms resolved. He later told us that he hadn’t been worried because he knew we were in control. His apparent lack of concern about his condition revealed expert certainty on his part!

In How Doctors Think, Jerome Groopman, MD, reminds us to beware of expert opinion when common sense would indicate that something else is going on. He suggests:

• Make a list of alternative explanations.

• Throw out previous conclusions and readdress the case.

• Be ready to get another opinion.

• Finally, beware that biases of all types can unconsciously creep into our daily clinical work.

– Robert E. Morris, MD, CCHP-P, is the 2020 chair of NCCHC’s board of directors and board liaison of the Society for Adolescent Health and Medicine. This column first appeared in the Spring 2020 issue of CorrectCare, Vol. 34, Issue 2.