Morbidity, mortality and other health-related decrements
Decrement framework
Health business is rarely a single-decrement problem. Morbidity incidence, recovery (termination), mortality and lapse compete, so a multiple-decrement table is needed. Rates may be central or initial, and exposed to risk should be measured in policy-years, not policy counts; recent years are understated until late-reported claims emerge. Distinguish select from ultimate experience, and cohort from cross-sectional data.
Drivers of the rate
Occupation class and earnings (income protection), age, sex, socio-economic group, region, underwriting basis, claims management (early intervention shortens duration), medical advances, diagnostic coding changes, and pandemic effects on both incidence and reporting.
Worked example
A portfolio shows 480 claims over 24,000 policy-years - a crude rate of 20 per 1,000. But 10,000 of those policy-years sit in a higher-risk class expected to run at 1.5 times the standard rate. Weighted exposure = 14,000 + (10,000 × 1.5) = 29,000, so the standardised rate is 480 / 29,000 = 16.6 per 1,000. Comparing the crude 20 with last year's 18 would falsely suggest deterioration; the book mix has shifted towards the riskier class.
Traps
Using population or NHS rates for insured lives, which ignores selection; confusing prevalence with incidence; ignoring claim run-off when measuring recent incidence, or fitting chain-ladder development factors to a triangle distorted by a claims backlog; assuming a fall in recovery rates and a rise in incidence are interchangeable - they hit duration and cost differently; treating upcoding or a definition change as genuine morbidity movement.