ReCent Actuarial News

Cause of claims in life insurance products

August 2024

The World Health Organisation reports that “the world’s biggest killer is ischaemic heart disease, responsible for 16% of the world’s total deaths [1]”.

What does this statistic mean? Several questions arise for the life insurance industry:

  • How does this affect insurance portfolios?
  • Is this population proportion applicable to insured lives as well?
  • If so, is the impact similar for different insurance products, different target markets or different distribution channels?
  • How would this affect product design?

These questions, and others, make it important to collect and analyse cause of life insurance claims. The following sections explore the analysis of cause of death and disability statistics, including considerations of some of the important drivers.

Cause of claims in life insurance products

The causes of claims are an important feature of life insurance. A basic example of this relates to double indemnity death policies that pay out double the sum assured if a death is due to accidental causes. In this example, to determine the correct payout level requires for the cause of claim to be known. Requirements such as the one above has led to the collection of cause of claim details by most life insurance companies. This is done as part of their claims processes. Cause of claims statistics are also collected by many state agencies and regulators around the world.

Applications

One main area of application of cause of claims statistics is in the design and pricing of life insurance products. This covers the pricing and determination of claim eligibility in products such as those described below:

  • Double/multiple indemnity. These are products where double (or other multiples) of the sum assured are paid if the claims cause is accidental. In these cases, the main requirement is just to be able to identify whether the claim events are due to natural causes or due to accidental causes. Such products are designed because accidental causes of claims are perceived to be less prone to anti-selection and moral hazard risks. This enables higher sums assured to be offered with little or no underwriting.
  • Initial exclusion (natural causes). Another design feature that is used to mitigate anti-selection risks is to only pays for natural causes of claims after an initial exclusion period – typically 1 to 2 years.
  • Cause specific products.These products provide insurance cover against one or several specified causes. This feature is more commonly used in living benefits insurance (e.g., disability or critical illnesses). Critical illness products covering cancer causes only, particularly targeted at women, are common in various markets in the world. The main motivation to offer these products is to create affordable policies that cover the risks that specific groups of people would be concerned about.
  • Tiered/multiple/specified cause payments. These are products designed where the payment and size of payout depends on the cause and severity of a claim cause. The most well-known examples are in critical illness insurance products where cancer diagnosis leads to a claim payout and the amounts paid are linked to the clinical staging of the cancer diagnosis. In disability and personal accidents policies, claim triggers could be linked to very specific claim events such as the loss of one eye.

Cause of claim information is also used in other operational and risk control parts of life insurance companies. Examples of this include the following:

  • Enforcing exclusion clauses. In general, life insurance policies are likely to have clauses that exclude self-inflicted causes of events (e.g., death from overdose of illegal drugs) from eligibility for claims.
  • Monitoring non-disclosure. The occurrence of some causes of claims at early durations of a policy is often used as a trigger for review of possible non - disclosure at time of sale.

The discussion above shows that the collection and use of high-quality claims information is important for life insurance. This significance of cause of claims information, in some instances, influenced the way in which causes of events are recorded. Examples of this include the following:

  • The reluctance by medical professionals to state HIV as a cause of death on certificates during periods when HIV was excluded from claim payments
  • The certification of deaths as being COVID-19 related during periods when governments were offering financial and other support to COVID-19 victims

Patterns in cause of deaths – illustration using country population statistics

The analysis of cause of claim statistics is done in many settings but the interpretation of the results requires careful considerations. It is important to set up credible baselines to which subsequent analyses are compared. These comparisons should consider the impact of the main drivers of causes.

Three main attributes influence causes of claims in life insurance. These are:

  • Age
  • Gender
  • Population subgroup. This is determined by the insurance target market (location, socio-economic groups), the type of product, sales process or other factors that determines the population under study.

This paper provides illustrations of these influences. Later we provide an analysis of observations within a data set of insured lives in Australia, allowing for the features discussed above.

In Figure 1 the proportion of all deaths due to cancer and circulatory (including cardiovascular) causes are shown for three developed countries: UK (England and Wales), USA and Australia. All other causes of claims are grouped together. The three countries exhibit broadly similar proportions of circulatory deaths with the USA showing a lower proportion of cancer deaths.

Figure 1: Distribution of deaths in populations of three countries for 2021(all ages).[2]

The cause of claims analysis is useful but should not be confused with the rates of deaths by cause. Considering the UK and Australia as examples, even both countries have the proportion of cancer of about 30%, the rate of death due to cancer is different in the two countries - at levels of 250 and 200 deaths per 100,000 population per year, respectively. The following table illustrates this point, using the rounded deaths and population values. The observation discussed above arises primarily from that the UK has a higher mortality rate than Australia.

Table 1: Comparison of cancer death rates for three countries for 2021 (all ages)

Age

Causes of claims vary significantly by the age of the individuals. In Figure 2 and Figure 3 the distributions of cancer deaths and circulatory deaths by age are shown for the three countries. It is apparent from these line graphs that the proportion vary by age. The following illustrative examples of these variations are taken from Figure 2 and Figure 3:

  • For UK, cancer deaths in the age group 65 – 69 is 43% compared to 14% in the 20-24 age group. Australia exhibits higher proportions at older ages than the UK and lower proportions at younger ages.
  • For ages 60 – 64, the cancer death proportion for USA is 30% compared to the 48% for Australia within the same age group.
  • For all three countries, the proportion of circulatory deaths continue to rise into old ages but the proportion of cancer deaths is at a peak between ages 60 and 74.

Any aggregate results (such as in Figure 1) will be masking these age and country differences.

Figure 2: Distribution of cancer deaths by age for three countries – total population

Figure 3 Distribution of circulatory deaths by age for three countries – total population.

Analyses of causes of claims statistics should always consider the age profile of the persons, especially where the results are used for comparison purposes.

Gender

Cause of death proportions differ by gender. To explore this, the population data for the three countries is restricted to the age group 50 to 54. Figure 4 shows distinct gender differences with females showing a higher proportion of deaths due to cancer and males showing a higher proportion of circulatory deaths. This observation is consistent for the three countries.

Figure 4: Distribution of claims in populations of three countries in 2021 – by gender for ages group 50 to 54

Other considerations

The claim cause distributions will be influenced by factors other than age and gender. Any factors that differentiate a group of persons from another has the potential to have an impact on claim causes. Within an insurance context this includes aspects such as whether someone has insurance or not, the type of insurance product (e.g., death or disability), the target market, level of underwriting, duration since underwriting, size of policy or premium. These factors are driven by socio-economic attributes or policyholder behaviour.

Calendar time impacts are also an important consideration due to the impact of aspects such as:

  • Medical advances
  • Introduction, changes, and cessation to government health initiatives such as screening programs
  • Changes in the economic levels of countries and the emergence of pandemics

Insured lives analysis - observations in some Australian insured lives

This section outlines observations based on insured lives in Australia and New Zealand. These results are presented within the context of the considerations provided in the previous sections. The analysis is based on approximately 12,000 death and 5,000 total and permanent disability (TPD) claims that occurred within the five-year period 2019 to 2023 and for ages between 30 and 65 years [3]. TPD policies payout a lump sum on disability.

The overall proportion of cancer and circulatory causes of death is similar between the insured lives and the population, as shown in Table 2. These two causes contribute a lot less to the disability causes of claim due to the dominance of mental, nervous, and musculoskeletal causes in disability insurance.

Table 2: Distribution of deaths and claim causes in lives aged between 30 and 64 in Australia. [3]

The time trends of causes are shown in Figure 5. The causes of death claims have remained stable over the five years (small uptick in cancer claims in 2023 largely due to the fact that cancer claims are generally reported sooner; in Australia terminal illness claims will be paid out prior to the death of a policyholder based on medical advice). The disability claims, show a steady rise in cancer claim proportions with a corresponding drop in contribution from other causes.

Figure 5: Calendar time trends in claim causes in lives aged between 30 and 64 in Australia.

Cancer deaths and mental disability claims are discussed in more detail below.

Cancer death claims

The contribution of cancer to all deaths is significantly higher in females (approximately half of females’ deaths) compared to males (approximately one third of males’ deaths). In this data set, breast cancer and lung cancer were the leading single causes of deaths in females. There is a marked increase in the proportion of lung cancer deaths as age increases. Breast cancer proportion is stable for most part of the 30 to 64 age range – as shown in Figure 7.

Figure 6: Composition of cancer deaths in females by cancer type - top causes

Figure 7: Proportions of lung and breast cancer in all cancer deaths for females - by age group

It is noted that in July 2025, Australia will be starting a nationwide lung cancer screening programme for people aged 50 and above.

Mental and nervous disability claims

The high proportion of TPD mental and nervous claims is of concern within the Australian insurance industry. Stress, anxiety, and depression constitute approximately 50% of these mental and nervous claims. It Is important to monitor these causes of claims, especially for income protection policies, as they tend to have longer recovery times than other claim causes. This has implications on the pricing of the product and on claim expense resources required to manage the claims. In this data set, there is no evidence that this proportion is increasing over calendar time.

In Figure 8, degenerative illnesses constitute about a fifth of the claim causes and there is observation of marked increase in these causes for males as age increases (see Figure 9).

Figure 8: Causes of mental and nervous disability claims in disability insurance.

Figure 9: Proportions of stress and degenerative claims in all mental claims - by age group

Various other analyses are ongoing on this data including the following, among others:

  • The development of the results as the late reported claims are registered
  • The potential impact of the COVID-19 pandemic on both disability and death claims
  • The potential impact of occupation on mental health claim causes
  • The relationship between disability causes of claim and death causes of claim

These output from the analysis is an important component of the ongoing enhancements of products and processed to assist life insurance companies to increase value to policyholders.

Conclusion

It has been discussed that the analysis of claim causes should be done within a specific context. The context should take into consideration aspects such as policyholder attributes, socio-economic and temporal attributes as well as product or channel attributes. This is important when the analysis includes comparisons across different groups of lives.

If comparisons over time are being considered, then it may be useful to establish a base line on which subsequent analysis are based.

The analysis has shown broad consistency between the population and the set of insured lives that was analysed. This creates opportunities to apply results obtained from population data to some insured lives portfolios. This suitability depends on the type of benefit under consideration. The difference between death benefits and disability benefits (living benefits) was shown in the different contributions of causes to claims. We have considered lump sum disability policies in the examples above, but income protection policies (proving regular income payments) and critical illness policies are likely to have different distributions of causes. Gender and age impacts on insured lives claims are very clear and should be considered in analytic work.

This paper has not considered data on different target markets or different distribution channels but, at a minimum, the differences in cause profiles by gender, age and product that we discussed would impact target markets, distribution and insurance processes that are influenced by these factors.

There are opportunities to create and understand further claims insights and trends using the individual claims data. Expansion of insights beyond just the cause, to include the lifestyle journeys that lead to the claims, may help improve customer outcomes in product and underwriting design.

Author

Chessman Wekwete

Head of Data Insights

Hannover Life Re of Australasia

References

  1. WHO Available under https://www.who.int/news-room/fact-sheets/detail/the-top-10-causes-of-death. Viewed on 15/05/2024
  2. Analysis done by Hannover Re. Circulatory category covers ICD codes I00 to I99. Cancer includes ICD codes C00 to D49. USA data is extracted from Centers for Disease Control and Prevention, National Center for Health Statistics. National Vital Statistics. UK data is from Source: Office for National Statistics licensed under the Open Government Licence. Australia data from General Record of Incidence of Mortality (GRIM) books - Dataset - data.gov.au. For UK and USA data the “U” ICD codes, that we assumed relate to Covid-19 deaths, are excluded in the data.
  3. Data source from Hannover Re

If you need further information or would like to send us feedback, please feel free to get in touch.

Contact forms

The information provided in this document does in no way whatsoever constitute legal, accounting, tax or other professional advice. While Hannover Rück SE has endeavoured to include in this document information it believes to be reliable, complete and up-to-date, the company does not make any representation or warranty, express or implied, as to the accuracy, completeness or updated status of such information. Therefore, in no case whatsoever will Hannover Rück SE and its affiliated companies or directors, officers or employees be liable to anyone for any decision made or action taken in conjunction with the information in this document or for any related damages.

Hannover Re is the registered service mark of Hannover Rück SE.

Hannover Rück SE © 2024

Imprint | Disclaimer | Data privacy | www.hannover-re.com