Ischemic Heart Disease Morbidity Rates Estimation using Hidden Markov Model Regression

Oflaz Z., Yozgatlıgil C., Kestel S. A.

23rd International Congress on Insurance: Mathematics and Economics (IME), Munich, Germany, 10 July 2019, pp.53

  • Publication Type: Conference Paper / Summary Text
  • City: Munich
  • Country: Germany
  • Page Numbers: pp.53
  • Middle East Technical University Affiliated: Yes


The precise estimation of mortality and morbidity tables is critical for optimal pricing of the life insurance and the health insurance products. Especially, unexpected high costs regarding the critical illnesses require understanding the factors in the likelihood of such an event. The number of occurrences of a certain illness in a population is an indication of its frequency. However, the characteristics of the illness alter the definition of the probability of an event. The data on such cases includes usually non-zero claim cases which makes difficult to determine the first time to be diagnosed. For this reason, it gains importance to depict the unforeseen factors.

In this study, we aim to model the possible factors and their states based on the number of occurrences/claims for a certain critical illness. For this purpose, we propose a Poisson hidden Markov model with explanatory variables to estimate morbidity rates of chronic disease. The hidden states will be determined based on a real-life case study focused on ischemic heart disease. The implication of states on the premium pricing is discussed.