We make decisions every day both for ourselves and our patients, yet we seldom think about how we make these decisions. In medicine, we consider data and evidence, leaving aside emotion and influence in favor of rationality. Cognitive bias, a systematic error in thinking, can sneak in and influence us.
Cognitive bias is relatively unstudied and under-reported. In a review of cognitive bias affecting physician decisions, including 38 articles evaluating clinical vignettes or real case scenarios, at least one type of cognitive bias was present in every study evaluated [1]. Framing effect and overconfidence were the most commonly studied biases, while anchoring, premature closing, and confirmation bias (see table) were the most commonly occurring, with a frequency of approximately 80% each [1, 2]. Five of seven articles (71%) found an association between cognitive bias and therapeutic or management errors, but too little evidence was present to determine if these errors had an impact on patient outcomes [1].
A recent case highlights these issues. The patient was an otherwise healthy 44-year old male with a history of appendectomy for perforated appendicitis 10 years prior in Nigeria. He was referred after several emergency department visits for various abdominal complaints; a CT scan showed a 4.4 cm round solid-cystic mass in the right lower quadrant within or abutting the mesentery of the right colon and appeared to be arising from distal ileum; this was reported by the radiologist as suggestive of a GIST (Figure 1). His exam was unremarkable outside a well-healed lower midline incision. His case was reviewed at our multidisciplinary tumor board, where the discussion was focused on the characteristic radiologic features that suggested a GIST; other diagnoses were considered such as, neuroendocrine tumor or a dilated Meckel’s diverticulum, but none were consistent with the imaging and further serology was negative. He was taken for resection where the mass appeared to arise from the distal ileal mesentery with dense adhesions to the ascending colon wall, necessitating a right hemicolectomy. The mass was firm without a clear capsule and no appreciable adenopathy or other abnormal lesions. His post-operative course was uneventful, and he was discharged 4 days later. His pathology report was released shortly thereafter: foreign body granuloma with no malignancy. The slides were reviewed in conference, where the pathologist declared, “The final path is gauze,” (Figure 2).
Since he was symptomatic, surgical removal was indicated even had the correct diagnoses been considered from the beginning, yet such a simple diagnosis wasn’t considered. The common biases discussed above affected us all. Over-reliance on the initial radiologic impression of a malignancy prevented considering a foreign body, even though his surgical history was known. Other alternative diagnoses were considered, but the differential wasn’t broad enough. Furthermore, diagnostic momentum and framing bias in our oncology-based conference propelled us even further down the path.
The risk for bias is everywhere, and personal factors (fatigue, cognitive loading), patient factors (complex histories, incomplete information) and system factors (workflow designs, insufficient time) can all raise the risk [2]. Physicians can employ several strategies to help counteract biases, but the most important is recognizing its existence and our susceptibility by discussing events in conferences to increase awareness and teach others [2, 3]. Forced reflection on a diagnosis, by using checklists to expand differential diagnoses or asking the question, “How else can this be explained?” can be helpful counteract diagnostic momentum, anchoring bias and premature closing. Using a systematic method for presenting information and taking your own history can minimize framing bias and controlling the environment by minimizing distractions and fatigue can be helpful overall [2, 3].
The part not yet discussed concerns what happened 10 years ago – the retained foreign body. These can happen at any institution, not just overseas – one prestigious U.S. hospital reported a rate of 1:5,500 operations, with sponges being the most common [4]. Such patients were less likely to have had surgical counts performed, more likely to have had emergency surgery or an unplanned change in operation, and a higher body-mass-index [5]. While surgical counts can help detect retained objects and are a simple, cost-effective strategy, many patients can still have retained material even in the setting of a correct count [4]. Robust processes, deployment of technology and an emphasis on just culture can help understand and prevent these issues in the future.
In summary, cognitive bias can influence our medical decisions and the care we provide for patients, without even our overt knowledge. Staying true to your principles, interviewing patients personally, considering a broad differential, and discussing cases openly, can help counteract and limit the impact.
References:
- Saposnik G, Redelmeier D, Ruff CC, Tobler PN. Cognitive biases associated with medical decisions: a systematic review. BMC Med Inform Decis Mak. 2016;16:138-51.
- Quick Safety: an advisory on safety and quality issues [Internet]. The Joint Commission: c2019. Cognitive biases in healthcare; 2016 Oct 28 [cited 2019 July 8]; [about 3 screens]. Available from: https://www.jointcommission.org/assets/1/23/Quick_Safety_Issue_ 28_Oct_2016. pdf
- Bhatti A. Cognitive bias in clinical practice: nurturing healthy skepticism among medical students. Adv Med Educ Pract. 2018;9:235-7.
- Cima RR, Kollengode A, Garnatz J, Storsveen A, Weisbrod C, Deschamps C. Incidence and characteristics of potential and actual retained foreign object events in surgical patients. J Am Coll Surg. 2008;207(1):80-7
- Gawande AA, Studdert DM, Orav EJ, Brennan TA, Zinner MJ. Risk factors for retained instruments and sponges after surgery. N Engl J Med. 2003;348(3):229-35.
- Becker’s Hospital Review: integration and physician issues [Internet]. ACS Communications; c2019. How 4 types of cognitive bias contribute to physician diagnostic errors, and how to overcome them; 2017 June 9 [cited 2019 July 8]; [about 2 screens]. Available from: https://www.beckershospitalreview.com/hospital-physician-relationships/how-4-types-of-cognitive-bias-contribute-to-physician-diagnostic-errors-and-how-to-overcome-it.html
Paul Dyrud
Publications Committee – Reviewers of the Year!