Symbolic logic analysis of cause of dealth in humans: Application to 108 patients after coronary artery bypass graft surgery

Over the past two decades there has been increasing interest in the development of an objective, or formalized "medical logic", and many authors have employed classical symbolic logic as a part of their approach. On the other hand, it has become clear that certain patterns of reasoning which are commonplace in evaluating patients clinicopathologically are awkward to handle in classical symbolic logic. The present paper proposes an extension of classical symbolic logic which addresses three problems in medical reasoning: (i) the problem of provisional diagnosis, (ii) the problem of inaccessible data, and (iii) the problem of the adequate discharge summary. It is proved mathematically that with a suitably constructed logic, the system "complains" until all questions involving threats to the patient's health are either answered or shown to be unanswerable because of inaccessibility of data. To illustrate this method, the cause of death was studied in 108 patients who had been autopsied at The Johns Hopkins Hospital after coronary artery bypass surgery. The analysis disclosed that 46% of patients suffered a fatal complication which could be attributed to events in the perioperative period; in 15% of patients the cause of death was unexplained by the analysis. Computerized symbolic logic analysis is a useful supplement to intuitive reasoning in assigning cause of death to patients with complex medical histories.