Abstract
Given the increasing incidence of stroke, we design and price a commercial stroke insurance using National Health Insurance Research Database（NHIRD）for insurance companies and customers in Taiwan. We collect 99,165 people over 40 years old during the period from 1996 to 2012 to perform the empirical study. In the study, the Panel Logit Model and Panel Negative Binomial Model are employed to estimate the probability of stroke and the frequency of outpatient visits or inpatient days due to stroke respectively. We find that age, gender, monthly income, place of residence, occupation, and disease history have significant effects on the probability of stroke and the frequency of outpatient visits or inpatient days due to stroke.