Clinical reasoning is the process of using patient information and medical knowledge to decide what may be wrong, what needs attention first, and what to do next. It is more than matching symptoms to a diagnosis: urgency, age, pre-existing conditions, clinical experience, and the consequences of an action can all change a judgment. section 2.2 section 8.0
A practical sequence for a patient encounter is:
- Address emergencies first, such as active blood loss. section 8.0
- Check an obvious diagnosis with a decisive test, then provide the necessary treatment. section 8.0
- If the answer is unclear, investigate likely explanations with affordable, safe, noninvasive tests before considering more burdensome measures. section 8.0
Choosing the next test
For a stable patient who reports a broken arm after a fall, the likely diagnosis can be confirmed with a decisive test. In a different case, if a simple throat culture can test a suspected diagnosis, it may be a more reasonable first step than an expensive, uncomfortable MRI. The choice depends on what question the test can answer. section 8.0
Turning this judgment into software is difficult. Rules must remain coherent across different patients and circumstances, yet much of a clinician’s expertise is implicit and hard to express as if-then logic. Doctors also use experience, analogies, interpersonal cues, and procedures to revise hypotheses. IBM Watson’s struggles with medical diagnosis illustrate why success at answering trivia questions does not establish an ability to handle clinical reasoning. Computerized support should therefore be judged against the complexity of the patient-care decisions it is meant to assist. section 2.2 section 10.3
