Age dependent decline of relative risks in life insurance medical underwriting

INTRODUCTORY: Life insurance medicine focuses on mortality hazards. People are free to insure themselves for small or large amounts and for short or long-terms. This freedom makes it necessary for life insurers to assess and select the mortality risks in a medical underwriting process. Medical underwriting guidelines are based on company statistics, population surveys following (clinical) epidemiological principles and clinical studies. Mortality of potential life insurance applicants is compared to life tables of insured populations, or to adjusted life tables of the general population. Because many risk determinants have higher normal values at higher ages, it is reasonably to assume that the relative hazards (RHs) or mortality ratios calculated for these risk determinants should be age dependent. This is also common use in underwriting guidelines, and can have much influence on the accessibility of life insurances for (chronically) diseased people. A proof of principle is therefor warranted.

This population-based cohort study uses NHANES- datafiles from the Third National Health and Nutrition Examination Survey (NHANES III) and the NHANES Linked Mortality Files 2010. Only participants aged 20 to 69 that were examined in mobile examination centers, without a history of some prevalent high risk diseases were included. The observed mortality was compared to the expected mortality in a Generalized Linear Model (GLM) with Poisson error structure with two reference populations, which theoretically both can serve as preferred reference for life insurers: The United States Life Tables 2008 and the 2008 Valuation Basic Tables based on the insured population of 35 US life insurers. The age dependency was assessed of the values and the RH s of the systolic blood Pressure (SBP), aspartate aminotranseferase (ASAT), lactate dehydrogenase (LDH), serum albumin and albuminuria, with correction for ethnicity, household income, history of diabetes mellitus, BMI and serum cholesterol.

All 5 continuous risk determinants had age dependent values in the comparison between ages 20-54 and 55-69 (Mann-Whitney U P < 0.001). Graphical inspections using age at time of interview revealed only for the SBP an increase with age. In the GLM again only SBP had a significant interaction term with age at time of interview. It made no difference which life tables were used for the calculation of the expected mortality.

Age dependency of RHs of risk determinants can be assumed if the risk determinants themselves are age dependent on statistical and graphical inspection. In other cases age dependency might not be significant, or cannot be modelled with some form of linear function as customary in many underwriting guidelines. The RHs or mortality ratios in current medical underwriting guidelines for life insurances should be checked for age dependency by analysing the underlying data statistically and graphically and by using GLM and appropriate life tables.