Namely, the data on disease occurrence in the population, eg, prior probabilities, strongly influence the calculation. However, this would be a mistake! The likelihood that the positively tested patient really has the disease in this case is less than 1%.
One can intuitively conclude that this probability is 99%. If we know that the accuracy of the test is 99% and that the disease appears in 1 out of 10 000 people, we can determine the probability that the positively tested patient is ill. And second, we need to know the occurrence of the disease in the population. First, we need to know the accuracy of the testing method. If we want to calculate the likelihood that one positively tested patient has the disease, one must know different expectations. The best example for explaining Bayesian statistics may be diagnostic tests. In this easily understandable and intuitive example, Alzheimer disease is an event and age is a condition associated with this event. The age information strongly changes the likelihood that these symptoms occurred due to Alzheimer disease.
Furthermore, it makes a big difference if this patient is 16 years old or 75 years old. For example, if the patient has difficulty remembering recent events and has mood swings and loss of motivation, how likely can we suspect Alzheimer disease? It is not easy to answer just on the basis of these symptoms. The basis of Bayesian statistics is Bayes' theorem, which describes the probability of an event occurrence based on previous knowledge of the conditions associated with this event. The number of published medical articles using Bayesian statistics in the period from 1995 to 2018 (, February 2019).