Dr. Mark Graber and his colleagues published an interesting article in the July 11, 2000 issue of the Archives of Internal Medicine involving diagnostic medical errors. Diagnostic error was defined for purposes of the study as a diagnosis that was unintentionally delayed,wrong, or missed. Errors were categorized as no fault, system related errors, or cognitive errors.
Cognitive errors were either due to faulty knowledge, faulty data gathering or faulty synthesis. Examples of system related errors included 1) biopsy report of cancer not communicated to the patient who missed a clinic appointment 2) consultation request for work up of pulmonary nodule lost 3) radiologists not available to read ER x-rays. The most common cognitive error was faulty information synthesis usually manifest as "premature closure" which is the tendency to stop considering other possible diagnoses after a diagnosis was reached. This is also referred to as a type of anchoring heuristic error.
The authors state that defective knowledge as a cause of error was rare and more commonly reflected problems with synthesis of available information.This refers to a formulation of how humans solve problems namely by searching for an explanation that best fits and then the search stops. Physicians at all levels of training and experience do this.
Are there take-home messages here? With premature closure, the suggestion is to make a conscious effort to not completely stop after you reach a diagnosis but ask " what alternatives should be considered?" This could be done initially and from time to time as the clinical case plays out. Some times things seem so obvious we do not do that but that practice seems like a good mental back up mechanism to minimize errors.
System problems can be varied and the particulars and the particular solutions may vary locally but one problem seemed prominent;radiologists not interpreting films in real time particularly in a emergency setting. Non-radiologist clinicians have been complaining about this since forever.
Personally, I have found this more of a problem in academic settings than in private practice, though it occurs there as well. With digitalization of imaging and broad band internet much of the interpretation could be done at home by the radiologist on call ( and this happens in some settings already). If the clinical doc is awake at 2 in the morning trying to unravel a clinical problem so should the radiologist.
The mechanism(s) of what cognitively occurs with missed diagnoses is not just of academic interest. Dr. Donald A Redelmeier wrote an excellent article entitled " The Cognitive Psychology of Missed Diagnoses" in the Annals of Internal Medicine, Jan 18, 2005. The full text is available by subscription. Cognitive psychology deals with how humans-including physicians-think and he presents basic concepts from that field in the context of a real life clinical case which give them practical significance.
The Graber article dealt with cases collected from 5 academic, tertiary case centers in which there was probably reasonable time to see patients and often with faculty oversight. With physicians in some managed care settings seeing patients every 5-10 minutes, we may well need to develop a revised taxonomy of error generation. Or perhaps most of the errors would be placed in the systems error section under "not enough time spent to figure out the problem."