[Clinical research XXI. From the clinical judgment to survival analysis]

Decision making in health care implies knowledge of the clinical course of the disease. Knowing the course allows us to estimate the likelihood of occurrence of a phenomenon at a given time or its duration. Within the statistical models that allow us to have a summary measure to estimate the time of occurrence of a phenomenon in a given population are the linear regression (the outcome variable is continuous and normally distributed -time to the occurrence of the event-), logistic regression (outcome variable is dichotomous, and it is evaluated at one single interval), and survival curves (outcome event is dichotomous, and it can be evaluated at multiple intervals). The first reference we have of this type of analysis is the work of the astronomer Edmond Halley, an English physicist and mathematician, famous for the calculation of the appearance of the comet orbit, recognized as the first periodic comet (1P/Halley's Comet). Halley also contributed in the area of health to estimate the mortality rate for a Polish population. The survival curve allows us to estimate the probability of an event occurring at different intervals. Also, it leds us to estimate the median survival time of any phenomenon of interest (although the used term is survival, the outcome does not need to be death, it may be the occurrence of any other event).