Date created: | Last Updated:

Category: Project

Description: Traditional medical inference requires explicit determination of a patient’s ailment prior to deciding the best treatment option. Evidence-based medicine informs these decisions from the outcomes of previous patients with similar ailments who received similar treatments. Precision medicine seeks to expand this nomothetic framework with many more potential diagnoses, the total possible number of which is inherently limited by the number of similar previous patients available for reference. The concept of making patient-care decisions using idiographic information unique to a particular patient may be a familiar concept to practicing clinicians, but it has no formal role within evidence-based medicine. The collective result is constrained inferential capacity of the dominant medical philosophy, leading to limited effectiveness of individual patient treatment decisions. A means of combining nomothetic and idiographic inference to optimize individual treatment outcomes would be a welcome addition to the modern medical armamentarium. Novel idiographic search algorithms informed by nomothetic prior beliefs can construct predictive models about individual patients that provide rigorous clinical decision support. This advanced medical inference framework exploits modern machine learning to generalize the concept of diagnosis and to make effective treatment decisions with or without definitive etiological understanding of a patient’s ailment.