A prognostic score is generated for each individual in the study by determining the probability that the individual will fall into a particular category of the exposure variable.
A prognostic score can sometimes be used to describe the probability of experiencing a negative outcome.
A prognostic score for an individual is sometimes determined by summing the number of risk factors possessed by that individual.
It is never possible to obtain a direct estimate of an individual’s probability of experiencing the event of interest: instead, we must categorise all the prognostic scores and determine into which category each individual falls (e.g. low, moderate of high risk of experiencing the event).
If a regression model is used to generate prognostic scores, a model that explains a large proportion of the variability in the data must be good at predicting which patients will experience the event.