BACKGROUND: This is an Impairment Study Capture System (ISCS) study of contemporary
diabetes mellitus mortality among insured lives. Because the diagnosis and treatment of
diabetes has changed during the last 15 years, many applicants may be expected to exhibit more favorable outcomes than in the past. The study covers policy-years durational experience extending to only 10 years. METHODS: We analyzed the total mortality experience of 41,972 insurance
policies. The
policies were issued at standard or substandard premium rates between 1989 and 2002
policy anniversaries. The number of
policies terminated by death (actual deaths) is compared with expected deaths using the 2001 Valuation Basic Table (2001 VBT). Main outcome measures are expressed as
mortality ratios (MR %) and excess
death rates/1000 (EDR/M). Poisson
confidence intervals are used to test the
statistical significance of
mortality ratios at the 95%
confidence limit. RESULTS: The total experience is based on 103,104 policy-years exposure: males 57,888 policy-years (56%) and females 45,216 policy-years (44%). There were 495 policy-deaths 284 male and 211 female. Substandard risks represented the majority of the total exposure, 76,658 policy-years in both sexes combined (male 56%, female 44%). The
mean duration of substandard exposure was 2.3 years. Total mortality for all insured age-groups and risk categories combined was 187%. The
mortality ratios for
policies rated standard had
confidence intervals that were consistent with 100% of the 2001 VBT. The
mortality ratios for
policies rated substandard had
confidence intervals that were above 100% of the 2001 VBT.
Mortality ratios varied with the type of treatment. They were lowest in those treated with diet alone and highest in individuals treated with diet plus
insulin. CONCLUSION: A clinical diagnosis of
diabetes continues to demonstrate evidence of increased, but improving, mortality in insured individuals. The underwriting risk appraisal process effectively categorizes the risk, especially for the substandard classes where the ratings assigned to
policies were consistent with the mortality results. The lack of significant differences in the
mortality ratios between males and females as well as between
nonsmokers and smokers indicate that the early duration variations by gender and
smoking status in the 2001 VBT account for these differences in early duration
diabetes mortality. Subsequent follow-up studies containing longer durations may show these differences emerging. Results must be interpreted with caution because of the small
data set, limited number of ISCS participating companies, and durational experience extending to only 10
policy years.
