Life expectancy at birth is a commonly used method of assessing health, improvements in health over time, and differences in health between different groups (defined on the basis of gender, time, geography, deprivation, social class, etc.). A common misconception is that life expectancy at birth measures the expected or average duration of life of a newborn; it does not. It is a measure of life expectancy assuming that the current age-specific mortality rates continue throughout an entire lifetime. Advances in healthcare, changes in political and social circumstances, wars and natural disasters, changes in the prevalence of risk factors and changes in diseases and medical conditions (such as acquired immunodeficiency syndrome (AIDS), pandemics, diseases resistant to antibiotics, etc.) and many other factors which influence health and life expectancy cannot be anticipated, so it is not possible to predict mortality rates for each age group in the future. This is particularly the case for newborns, as the future mortality rate in 90 years’ time when that person is 90 years of age would need to be estimated. Life expectancy at birth is frequently used, but it is possible to calculate life expectancy at any age. For example, life expectancy at age 65 years could be calculated. This will tend to be closer to the true or actual duration of life than life expectancy at birth would be for a newborn. This is because changes in mortality rates will tend to be reasonably gradual, so that the current rates of mortality (on which life expectancy calculation is based) might be a reasonable prediction of mortality rates in the next 20 years or so.

Life expectancy at birth is calculated for an arbitrary 100,000 males or females and is the average of how long they will live based on current actual age-specific mortality rates. In most cases the calculation is undertaken for grouped ages (5-year age bands, e.g. 0, 1-4, 5-9, 10-14, 15-19, …, 90+) rather than for single years of age.

For example, if the mortality rate within the first year of life was 5 deaths per 1,000 live births, then this would equate to 500 deaths in our arbitrary 100,000 population prior to one year of age. Their ‘contribution’ to the life expectancy calculation would be 0.1 years each (most infants who die within the first year of life die within the first seven days of life so their contribution is relatively low compared to other age groups). These 500 ‘individuals’ contribute a total of 50 years to the life expectancy calculation. The contribution of the remaining 99,500 alive at the end of the first year would be one year each so 99,500 years in total. So the total contribution to the life expectancy calculation would be 99,550 years.

If the mortality rate in the second age group (1-4 years) is 40 deaths per 100,000, then there would be around 40 deaths (among the remaining 99,500 individuals) and their ‘contribution’ would be around 20 years (half a year on average) multiplied by four as there are four years for the age group (1-4 years) so they contribute 80 years in total. The remaining 99,460 individuals would contribute a full four years and thus contribute 397,840 years in total to the life expectancy calculation. So the contribution for the second age group would be 397,920 years, and the total contribution in the first two age groups would be 497,470 years.

These calculations continue for each age or age group with the ‘contribution’ in the final (open ended) age group (e.g. 90+ year age group) calculated in a slightly different method.

The ‘contributions’ for each age group are summed and divided by 100,000 (the starting number of individuals) to obtain the average life expectancy at birth over these arbitrary 100,000 individuals.

It is not possible to calculate life expectancy where the population or the number of deaths in the oldest (90+ year) age group is zero. Locally this occurs for Kingswood ward which had a younger population and is the least deprived ward in Hull, although it is possible that the Office for National Statistics may have underestimated the population in Kingswood due to considerable new housing development in the last decade. For cases, where the population is zero in the 90+ year age group, the population has been changed to one. For cases, where the number of deaths is zero in the 90+ year age group, the number of deaths has been assumed to be the same as the Hull average (so the same percentage of deaths in that area as Hull for the same gender and period of time). For instance, if the population and deaths of a particular area were both zero then the population would be changed to one, and if the number of deaths in Hull for that same gender and time period was 140 and the population was 400 giving a 35% mortality rate for Hull, then it would be assumed that the number of deaths in that particular area would be 0.35.

Healthy life expectancy or disability-adjusted life expectancy can also be estimated. This is a modelled estimate of the average number of years a person would expect to live in good health or disability-free based on current mortality rates and prevalence of self-reported good health / free from disability. For healthy life expectancy, good health is based on the response to the question “How is your health in general; would you say it was very good, good, fair, bad or very bad?”. The responses “very good” and “good” were categorised as “good health”. For instance, if only 10% of people in the 70-74 year age group classify their health as “good” then the ‘contribution’ to the life expectancy calculation would be 10% or one-tenth that for overall life expectancy calculation. The disability-adjusted life expectancy calculation works in a similar way in that only a percentage of the population alive (those who are disability-free) contribute ‘years’ to the calculation.

Also see: Disability Adjusted Life Expectancy, Disability Adjusted Life Years, Quality Adjusted Life Years and Years of Life Lost.