Healthy life expectancy (cont.)
Stages in the process of determining healthy life expectancy
These stages are illustrated on the following page.
A current (period) life table is required.
A second data source is required, which gives the age specific prevalence of health problems by age groups which match the life table groupings. These data are usually provided from a cross sectional survey, or from a census. The questions asked will determine what sort of healthy life expectancy can be estimated. Examples would be general health, disability, mental disorders, cancer.
The prevalence of the disorder under investigation - usually expressed as a percentage - represents the percent of people suffering this disorder at any particular time, in the stated age group. It can also be viewed as the percent of person-years spent with this disorder in that age group.
Since the nLxcolumn of the life table can also be viewed as a population structure or a statement of person-years we can take this column and discount it - that is remove from it the percent of years spent with the disorder, as given by the prevalence survey figures, by age group.
Once this is done and we have a discounted nLx column we can cumulate the new column to form a Tx column and then calculate a new ex column in the normal way, ex=Tx / lx.
This ex column is an estimate of life expectancy without the time spent with the disorder.
This is an estimate of healthy life expectancy of one form or other. It is simply derived and intuitive. This process is known as Sullivan's Method and is well recognised and widely applied. It is an approximate method but seems to work well. There are other methods but they tend to be more technical, require more data and are computer-intensive, sometimes requiring standalone programs.
Since life expectancy is a rather speculative measure anyway, and healthy life expectancy even more so, the simple Sullivan estimation method is usually adequate.