The Complexities of Diagnosis – Understanding How and Why we Fail
There is one additional element of medical and nursing education reform that we have not yet considered and that is the complexity of diagnostic processes. Transforming healthcare will require increased emphasis on diagnostic risks and professional vulnerabilities1, 2 especially as we confront the challenges that deterioration in population health and an increase in illness burden will present in the future.
When patients seek our counsel, they bring their concerns and share their symptoms, their past, family and social histories. In a busy practice, where physicians and nurses see 25 or more patients daily, this is a boatload of information. Add to this the clinical data provided by physical examinations, laboratory and imaging investigations, all to be evaluated during a hectic and challenging day, and it becomes easier to appreciate the many opportunities for errors.
Physicians and nurses have adopted strategies to rapidly sort through this morass of information in order to quickly arrive at tentative working diagnoses. Unfortunately, each of these strategies poses risks for error and harm.
Case Study:
A teenage boy presented to an emergency department with anorexia, nausea, anxiousness and lethargy, having been brought in by friends after collapsing at a party where alcohol and amphetamines were being consumed. The emergency department physician, now in the 11th hour of a hectic shift, noted the lethargy and a peculiar odour on the teenager’s breath, which he concluded was due to alcohol mixed with tobacco and pizza residua. He performed a cursory neurologic exam and a rendered a working diagnosis of alcohol intoxication, r/o coincident amphetamine abuse. Not unreasonable…
Notably, the patient’s blood pressure was 180/120 and his pulse was 90 bpm, both of which the doctor attributed to amphetamines. IV fluids were started, blood work drawn for toxic screen and other labs, and the patient was put to the side while other, presumably more serious, patients were addressed. The patient’s blood alcohol was 0.19 (legally intoxicated) but no amphetamine was detected.
Three hours after arrival in the emergency department the patient suddenly experienced a 20-minute left sided seizure that responded to medicinal management. At this time, it was noted that he had asymmetric pupils, with the right pupil being very large, and a scalp contusion over his right hemi-cranium, beneath his thick hair. A thorough review of his laboratory studies revealed dramatic elevations in urea nitrogen and creatinine, suggesting advanced kidney failure. A routine urinalysis had not been ordered.
The correct diagnoses, subsequently confirmed, included end-stage renal failure secondary to chronic pyelonephritis and polycystic kidney disease, a right-sided subdural hematoma subsequent to head trauma and uremic bleeding diathesis, and alcohol intoxication. The odour of his breath was due to uremic poisoning. His subdural was evacuated, dialysis begun and he was discharged three weeks later.
This patient’s diagnosis was delayed because of the lack of thoughtful attention to detail. The physician’s judgment was impaired by several common knowledge-based errors in conscious thinking, i.e., in cognition. First, the physician was affected by memory bias, having seen presentations similar to this in the past and therefore rapidly concluding that this was just another alcoholic teenager. Second, he was affected by confirmation bias, seeking out lab studies to confirm his suspicion and not even checking those that would have suggested otherwise. Furthermore, he was subject to an availability heuristic, a tendency to affirm the first diagnosis that comes to mind, “grabbing the low hanging fruit on the vine” and finally, he was guilty of overconfidence, the tendency to believe in the validity of a chosen course of action while discounting or not considering contradictory evidence2.
Errors of cognition are what Professor James Reason has called rule-based and knowledge-based errors3. Rule-based errors occur when the wrong rule is applied to deductive reasoning or when the correct rule is misapplied. Knowledge-based errors are even more complex because often we do not recognize when they occur. In both instances, these kinds of errors are more likely to occur when physicians are psychologically stressed, fatigued, sleep deprived, task saturated or otherwise distracted.
Prevention begins with awareness of vulnerabilities and the preconditions that set the stage for these types of errors. Teaching awareness of diagnostic vulnerabilities and risks for personal failure is a moral imperative and a quintessential element in medical and nursing education reform.
Refs:
- Leape L., et.al. Transforming Healthcare: a Safety Imperative. Qual Saf Health Care 2009;19:424-428
- Leape L. Error in Medicine JAMA 1994;272:1851-57.
- Reason J. Human Error. Cambridge, Mass; Cambridge University Press, 1990.