The Cognitive Error Paradigm

Krisstina Gowin, DO

As fellows in oncology, we are constantly required to assimilate data, interpret diagnostic tests, decide on appropriate treatment regimens, and employ compassionate care in an increasingly truncated allotment of time. We must simultaneously juggle these responsibilities with each patient, and as we progress in our training, we hardly even notice we have 12 balls in the air at once.

Our thought processes undoubtedly become somewhat automated as we develop our thinking algorithms and we rely more and more on our gestalt assessments of patients. We are trained to treat patients based on evidence-based medicine, a term defined as “the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients.”¹

Yet, delivery of better medical care requires more from us. It requires a higher level of thinking. In fact, to be the best requires thinking about our thinking. Here, I share 3 cases that exhibit physician errors in cognition and how questioning our thinking may lead to the delivery of better care.

Recently, a 56-year-old male came to me for consultation regarding his newly diagnosed metastatic prostate cancer. He was an anxious marketing associate from the East Coast who presented to his primary care physician with increased urinary frequency, particularly at night. His prostate-specific antigen (PSA) level was found to be elevated at 4.0. He was referred to his local urologist, who recommended a transrectal ultrasound with biopsy, and shortly thereafter the patient was diagnosed with Gleason grade 6 prostate cancer. Prior to starting therapy, he was staged with a positron emission tomography—computed tomography (PET-CT) scan; although not supported by National Comprehensive Cancer Network guidelines, this is a common practice in the community.

The patient was found to have “numerous fluorodeoxyglucose—avid lesions within his osseous skeleton consistent with metastatic prostate cancer.” He presented to my office with the pathology slides, imaging report, and records from another physician noting his diagnosis of metastatic prostate adenocarcinoma. It was a busy day in the clinic, and I was behind schedule. Two patients were already anxiously awaiting their appointments. Initially, I was relieved by what seemed to be an easy consult. It was a cut-and-dried case of prostate cancer. My plan was to start androgen deprivation therapy and move on to the next patient. However, something seemed amiss. He was only 56 years old. He only had Gleason grade 6 prostate cancer. He had no family history of cancer. And yet he presents with widely metastatic disease?

I promptly called the radiologist to review the images. Interestingly, all of his “lesions” were in joint spaces and more typical of degenerative joint disease. This was only arthritis! My instinct had been correct and this patient was misdiagnosed initially. I was able to change his terminal diagnosis of metastatic prostate cancer to a case of local and potentially curable prostate cancer. The patient and his family were ecstatic, but wisely inquired, “How could this have happened?” Where had the medical decision-making gone wrong?

The difficulty in providing an explanation lies, in part, in that no true medical error occurred. It was an error in medical thinking. The radiologist viewed the images under the clinical premise that the patient was recently diagnosed with prostate cancer. He was looking for evidence of prostate cancer and he believed he found it. The radiologist had made an error in medical cognition, namely an “attribution” error. He saw lesions in the skeleton and wrongly attributed them to prostate cancer because he did not question the diagnosis and think, “What else could this be?”

In another case, an 80-year-old grandfather came to me for a consult on the biochemical recurrence of his prostate cancer. He had been diagnosed 5 years earlier and had been successfully treated with brachytherapy. His urologist had trended his PSA and after a long period of being undetectable, it had increased over the preceding 4 months with a doubling time of only 2 months. A CT scan of his abdomen revealed diffuse lymphadenopathy, particularly within the obturator nodes, a pattern typical of metastatic prostate cancer. He was referred to me for treatment of his metastatic disease.

Upon my evaluation, I noted that a serum protein electrophoresis performed 2 years prior showed a mild monoclonal gammopathy that had not been addressed. I diligently performed my evaluation, including work-up of his monoclonal gammopathy, and given the questionable diagnosis of metastatic prostate cancer, elected to refer him for excisional lymph node biopsy. Several days later I was paged by the hematopathologist, who relayed the diagnosis of lymphoplasmacytic lymphoma, or Waldenström’s macroglobulinemia. When the news had been communicated to him, the patient questioned, “How could I be mislabeled as having stage 4 cancer?” How could his trusted urologist make such a mistake?

Our 80-year-old grandfather’s urologist fell into the cognitive error of “anchoring,” or latching on to an already established diagnosis. The urologist opened our patient’s medical record and immediately read the first line of the last note entitled “prostate cancer follow-up.” When his PSA was elevated in the presence of CT-demonstrated lymphadenopathy, his first thought was prostate cancer. “Of course!” the urologist thought. He failed to challenge his own thinking and consider an alternate diagnosis.

A final case demonstrates an “availability” error, when an unusual case creates a sharp bias in medical decisionmaking. A 60-year-old female with newly diagnosed diffuse large B-cell lymphoma underwent routine outpatient chemotherapy utilizing the well-known regimen of R-CHOP (rituximab, cyclophosphamide, hydroxydaunomycin, oncovin, and prednisone). On day 4 of her chemotherapy, she developed tumor lysis syndrome, acute renal failure, and arrhythmias. She died in the intensive care unit later that day.

The attending physician, extremely distraught by the outcome of this patient, subsequently developed a new policy requiring that all of his patients undergoing initial chemotherapy with R-CHOP be admitted to the hospital. He is acutely aware of the lack of evidence to drive his decision and the additional cost, inconvenience, and potential risk of hospital-acquired infections to the patient. His personal experience has driven him to develop a sharp bias in the treatment of his patients. His poignant past experience is cognitively available to cloud future decisions.

These 3 cases demonstrate how cognitive errors can affect medical practice. As Atul Gawande, MD, MPH, writes, “There is science in what we do, yes, but also habit, intuition, and sometimes plain old guessing.” We are human, and by definition, imperfect in our nature. The acceptance of imperfection is our first step toward avoidance of the cognitive error traps. We must constantly be on guard. We must question our conclusions and thought processes. We must continually evaluate our emotions and contemplate how they may affect our assessments.

The errors of attribution, anchoring, and availability are only as strong as our negligence of them. The recognition of the potential for error and resilience to a cavalier mindset is our only insurance. To be better doctors now, and many years from now, we must continue to think about how we think.

Reference

1. Sackett DL, Rosenberg WM, Gray JA, Haynes RB, Richardson WS. Evidence based medicine: what it is and what it isn’t. BMJ. 1996;312(7023):71-72.