Logical commentary:
How does the clinician limit her differential diagnoses? What criteria does she use to narrow the hypothetical diagnostic space? The clinician will look carefully for features in the patient’s history, perform a physical examination, and pursue initial testing. Some details will be retained; others rejected or held in abeyance. She will listen to tone of voice; try to assess anxiety bearing in mind that the patient has received heretofore 30 mg of morphine. She will examine not only quantitatively but also qualitatively—observe the state of the manifestation of his pain, anxiety, and the look in his eyes. She must proceed with a rapid assessment to rule out referred pain—e.g. from acute pancreatitis, from another abdominal organ. Some of her observations are salient at the time of the first examination. It is not uncommon for some physicians, in certain clinical circumstances, to delay an hour or more before an acute abdomen diagnosis is firmly established by re-examination. But in this case there is no time to delay, as the pain seems to be continuous even with morphine analgesia. Some of the observations fit into a picture of myocardial damage, others do not fit.
The emphasis must be on reforming the observations into an emerging, bringing-into-focus picture, on filling in the pieces as expeditiously as possible to include less common conditions, even redundantly reviewing the initial observations so as not to overlook subtle clues—e.g. persistent 7/10 chest pain partially relieved by higher and more frequent narcotic administrations and a soft holosystolic murmur across the precordium.
In the case at hand the physician is faced with conflicting observations and testing that is equivocal. Although the history of non-exertional chest pain seems counter-intuitive for stress-induced coronary artery disease, the negative enzyme measurements and ECG tracings should be evidence against the diagnosis of cardiac injury.
But there is strong clinical and epidemiologic background information from population studies and clinical disease presentations that are atypical and do not fit the textbook picture based on the population under consideration. Despite the negative tests, the suspicion (prior probability) of cardiac injury is high enough to persist in testing the clinical presentation with routine tests or even to perform angiography or tests to show calcification of the coronary arterial system.
The clinician must not miss the commonest diagnosis; she must not prematurely close the consideration of hypotheses. Perhaps the presenting picture needs refocusing and another observer must review the data? This may be prudent as the case develops in atypical fashion i.e. borderline enzymes, ECG non-diagnostic, pain relief minimal. At the same time, she must not miss a life-threatening condition either. The standard protocol calls for repeating cardiac enzymes; and it is standard practice to continue to monitor the ECG and to observe the patient closely for signs of decompensation, while at the same time reviewing the differential diagnoses so not to miss or confuse this case with another lookalike disease, a mimic of cardiac injury. If the physician were working in or with another situation where prevalence would differ, the list of probable diagnoses might change—e.g. in South America she would need to consider Chagas disease, uncommon in the U.S.A.
Now the work of the physician is to complete, as well as she can, the picture of the specific etiology of chest pain. She may need to repeat the initial tests, review the clinical history, and hypothesize other explanations in a timely fashion. This is another challenge. She may call for another pair of eyes; sometimes this is a fallback position, other times it grows out of being confused by the picture or even dissatisfaction with her own thought process. She may doubt her diagnostic considerations and require additional testing to secure an accelerated pathway to a more assured diagnosis. This is common especially during training.
Saliency emerges based on experience—i.e. recall of the association of clinical findings (history of non-exertional chest pain, negative enzymes and ECG, continuing pain relieved only with narcotics) with previous disease states. She chooses those features that functioned in the past to anticipate an outcome—a diagnosis. Discrimination based on previous experience gives her the ability to create a full picture and to recognize that this clinical picture resembles, in all or most of the details, pictures she has seen before. This may be called pattern recognition. Patterns that clarify the way to proceed towards diagnosis selection and treatment emerge from experience. In short, successful diagnosis requires careful, judicious, and experienced observation for pattern recognition.11For an exposition of how this process works semiotically—that is, by the creation of mental signs such as pictures, diagrams, schematta, and so on—see Silveira 2005.