Probabilistic Medical Reasoning and Clinical Decision-making
Medicine is the science of uncertainty and art of probability.
Medicine is full of uncertainty to the point that there is no absolute certainty in all diagnoses from clinical, procedural and laboratory diagnostic (blood tests, imaging, etc.), histopath, operative, and even post-mortem.
Medicine is full of uncertainty to the point there is no absolute certainty in predicting the life span of a patient; when a patient will die; course of a disease and treatment outcome.
Medicine is really an art of probability in problem-solving and decision-making.
There are two kinds of probabilities that one use in problem-solving and decision-making in the practice of medicine.
These are subjective probability and frequential probability.
Subjective probability is the probability derived from an individual’s personal judgment about whether a specific event or outcome is likely to occur. It has no formal calculations and only reflect the individual’s opinions and past experience. This is the more common type probability used by physicians in predicting an event or outcome as there are still lots of medical conditions with no documented or unpublished calculated frequential probability. The main reason for this lack is the difficulty and tediousness in getting accurate frequential probability.
Frequential probability is the probability derived from a formal study noting the relative frequency of occurrence of an event or outcome observed in a number of repetitions of the experiment. The frequential probability is more reliable than subjective probability as this is derived from a formal study of large series of incidents or cases. As mentioned earlier, however, there are still lots of medical conditions with no documented or unpublished calculated frequential probability.
Probability, whether subjective or frequential, is used by physicians in arriving to a clinical diagnosis after history and physical examination; pretreatment diagnosis after paraclinical diagmostic procedures are done; prediction of the outcome of treatment; and prediction of course and prognosis of a disease.
Because clinical data are imperfect and outcomes of treatment are uncertain, the physicians often are faced with difficult choices. Physicians often use probabilistic medical reasoning to deal with uncertainty inherent in medical decision.
More often, physicians use subjective probability more than frequential probability as often times, it is difficult to assess probability from “hard” data. One, frequential probability data are absent or not readily available. Second, frequential probability data if present do not match the requirements of the patients.
Challenge: is there a systematic way of calculating subjective probability? Am looking for one.