February 21, 2022 | News | Interview

Why We Should Monitor Disparities in Old-Age Mortality with the Modal Age at Death

Viorela Diaconu and colleagues want to introduce new measures to gauge key demographic issues in aging societies, one of them being the modal age at death. In this interview she talks about the advantages of this measure and her findings after applying it to Finnish register data.

Dr. Diaconu, over the last five decades mortality has been decreasing markedly at older ages in high-income countries. Researchers thus started to be interested in knowing how these decreases are distributed across social classes. You propose a new measure to do so. Why?

Usually, disparities in old-age mortality are assessed using indicators of mortality that are based on a fixed old-age threshold. Age 65 is generally set as the age distinguishing the young from the old because at this age rates of sickness and death begin to show a marked increase. However, the threshold age of 65 today does not have the same meaning as it had five decades ago. And it is different for a person from a lower socioeconomic background than for someone from a higher socioeconomic background. My colleagues Alyson van Raalte, Pekka Martikainen, and I thus want to introduce a new measure called the modal age at death for monitoring social disparities in mortality; the reason being that it is not based on an arbitrary selection of the old-age threshold and that it is not fixed over time.

You have applied the measure to Finnish register data. What are your findings?

We show that the modal age at death, i.e., the age when most deaths in a birth cohort occur, has been increasing for all occupational classes and for both sexes in Finland since the 1970s. While the uncertainty surrounding the timing of death of older people has been decreasing. These results suggest that great progress has been made in survival improvements at older ages and in reducing the uncertainty about the timing of death. This applies to all occupational classes in Finland.

Was the pace of improvements similar across socioeconomic backgrounds?

Yes, in general it was. But it was slightly faster among the lower occupational class compared to the upper occupational classes. This suggests that once the lower occupational classes reach older ages, they have mortality improvements on par with the higher occupational classes. Throughout 1971 to 2017, the upper non-manual classes lived longer than the manual classes, such as reflected by occupation-specific trends in the modal age at death and in life expectancy at age 65 and 75. However, the gap between the non-manual and the manual classes in survival chances to older ages evolved differently, depending on which mortality indicator was used. Trends in the modal age at death by occupational class revealed that there were relatively stable disparities over the 1971-2017 period. But trends in life expectancy at age 65 and 75 revealed that the disparities widened.

Do you have an explanation for the different development of disparities over time?

Well, indicators based on a fixed old-age threshold such as 65 or 75 have some limitations: They do not consider that today’s older people are not the same as yesterday’s, they compare groups that have different mortality risks over age and over time, and they capture changing cause-of-death compositions at older ages over time. But there is no inherent reason why the modal age at death and conditional life expectancy indicators should agree on the direction of inequality. Different mortality indicators used to measure inequalities often lead to different conclusions.

Did you get results that surprised you?

Yes. The most surprising result is that the proportion of individuals surviving to the modal age at death was stable throughout the observation window, for more than 45 years, for all classes and both sexes. This was in contrast to the trends observed in the proportion of individuals surviving to age 65 and 75. Over the study period the proportion of individuals surviving to these ages nearly doubled across occupational classes. This suggests that, unlike conditional life expectancy, the modal age compares individuals with similar survival chances over time and across occupational classes. It can thus be considered as a measure of aging that reconceptualizes age based on the characteristics of individuals, in this case their survival characteristics.

Original Publication

Diaconu, V., van Raalte, A., Martikainen, P: Why we should monitor disparities in old-age mortality with the modal age at death. PLOS ONE (2022). DOI: 10.1371/journal.pone.0263626