A Look at the Survivor Behaviour of Long Term Care Claimants
Survival analysis data for individuals who need long term care (LTC) insurance is usually scarce, but I had the rare opportunity to study a European life insurance company’s data on nearly 3,000 long term care claimants.
For the study, a person was deemed as disabled and in need of care when unable to perform a set number of activities of daily living or when considered cognitively impaired. The corresponding care level was fairly severe.
It is well-known that females have higher life expectancy than males and the same holds true for disabled lives. My analysis revealed females live 1,030 days on average following the point of claim, compared to 695 days for males.
I used separate statistical models for female and male disabled lives to take account of their differing survival behaviour. This meant I could draw conclusions on mortality rates beyond the observation time and analyse the influence of certain risk factors on survival.
While the lifetime distribution isn’t always of interest, how it varies between certain values of the risk factors is.
Once a person needs care, his or her survival is influenced over time by risk factors such as age at claim and pathology. I used models that accommodate differences in survival behaviour between certain groups; for example, cancer patients vs. non-cancer patients or patients of different ages at the point of claim, as well as the evolution of these differences over time.
The data showed cancer patients have higher mortality than non-cancer patients, especially at the time of claim. The impact is stronger in females and it takes some years for the mortality rates for both groups (cancer vs. non-cancer) to converge.
In the group of all non-cancer patients, the relative risk of death for people with a neurological condition or dementia is also fairly time-dependent. The data showed these individuals have a relatively small risk of death at the start of care, which increases over time, and results even out in higher mortality.
I also studied the influence of age at claim on disabled lives mortality. After identifying significant age bands, the analysis showed that younger lives have higher mortality than older ones in the first year of LTC. This could be explained by pathology at point of claim.
The results of the analysis go beyond merely revealing survival behaviour of LTC claims from data that is hard to obtain; the results can help insurers to make even better the development and pricing of LTC products. You can get more technical details about the analysis in my articles: “How to Handle With Care Data on Disabled Lives Mortality – A Statistical Approach” and “Long Term Care – How Does the Survival Behaviour of Claimants Differ?”