Diagnoses by general practitioners: Accuracy and reliability

Only 10% of the results of consultations in primary care can be assigned to a confirmed diagnosis, while 50% remain "symptoms" and 40% are classified as "named syndromes" ("picture of a disease"). Moreover, less than 20% of the most frequent diagnoses account for more than 80% of the results of consultations. This finding, confirmed empirically during the last fifty years, suggests a power law distribution, with critical consequences for diagnosis and decision making in primary care. Our results prove that primary care has a severe "black swan" element in the vast majority of consultations. Some critical cases involving "avoidable life-threatening dangerous developments" (ALDD) such as myocardial disturbance, brain bleeding, and appendicitis may be masked by those often vague symptoms of health disorders ranked in the 20% most frequent diagnoses. The Braun distribution predicts the frequency of health disorders on a phenomenological level and reveals the "black swan" problem, but is not a tool by itself for arriving at accurate diagnoses. To improve predictions and enhance the reliability of diagnoses we propose standards of documentation and a systematic manner by which the risk facing a patient with an uncertain diagnosis can be evaluated (diagnostic protocols). Accepting a power law distribution in primary care implies the following: (1) primary care should no longer be defined only by "low prevalence" properties, but also by its black-swan-incidence-problem. This includes rethinking malpractice and the requirements of malpractice litigations; (2) at the level of everyday practice, diagnostic protocols are tools to make diagnoses more reliable; (3) at the level of epidemiology, Braun's system of classification is useful for generating valid information by which predictions of risks can be improved.