## Summary judgement

In previous posts we have looked at problems with the quality and reliability of cause-of-death data and a list of hurdles for mortality projections based on such data.  One other issue is that of detail.  While cause-of-death data is spread over literally thousands of individual causes, important detail is lost on the most important mortality risk factor of all.  Oeppen (2008) states the problem:

"deaths are often tabulated by 5 year age groups and the open age interval into which the deaths of the oldest-old are aggregated is often de fined at a relatively young age such as 85. Unfortunately, it is at these high ages where most of the temporal dynamics are occurring."

By way of illustration, Table 1 shows the level of detail available for a leading cause of death for males of retirement age in England and Wales.  The data in Table 1 are at exactly the level of detail provided, i.e. without any further summarization.

Table 1. Male deaths for ICD-9 code 4850 (Bronchopneumonia, organism unspecified) in England and Wales in 2000. Source: 20th Century Mortality.

AgesDeaths
50–54175
55–59216
60–64379
65–69731
70–741,463
75–792,821
80–843,653
85+7,380

How important is this loss of age-related information?  The probability of a male aged 65 exceeding the age of 85 is 43.1% according to the pensioner table S1PA.  Indeed, most current actuarial bases for pension schemes assume that over half of males aged 65 exceed this age.  Grouping data into a category of "age 85 and over" is therefore a material loss of detail when one is interested in pensioner mortality.

Information is always lost when data are summarized.  This need not be an insurmountable problem if the degree of summarization is modest.  However, losing all age-related details above age 85 is a major drawback for cause-of-death methods if they are to be applied to questions of pensioner mortality.

Assume we have a random variable, $$X$$, with expected value ... Read more