next up previous
Next: An approximation of the Up: Interest of such an Previous: Interest of such an

A consistent public health indicator


Sullivan indicator is a health state measure of a collection of persons which is independent of its age structure. So, if the same definitions are used both for health states and age groups, it offers a possibility to compare health states of the entire population between two dates in spite of a modification of its age composition (provided that the sizes of the age groups are not too large). The comparability is also increased by the calculation of this indicator separately for males and females and if necessary (in order to adjust for structural differences) for other categories.

But the Sullivan indicator has also two essential advantages:

  1. It is not very sensitive to the size of the age groups which are used provided an adequate calculation of the tex2html_wrap_inline1022 (see section 3.2.3). For that reason, the use, for example, of an abridged life table gif, instead of a complete life table, doesn't change essentially the results if values of tex2html_wrap_inline1022 for the age intervals are calculated by summing values of person years on single age intervals, tex2html_wrap_inline1186 , from the unabridged life table [7]. Generally, it is preferred to use five (sometimes ten)-year age intervals because the sample of most of the surveys used to derive the prevalence ratios are too small for allowing smaller age intervals.
  2. Because of the use of life tables (which derive from measures of age-specific death rates on the entire population), the value of this indicator has often a good statistical precision [18] [7]. Moreover, this precision is rather easy to calculate (see example in annex).

Eric Hauet
Fri Apr 25 22:40:35 DFT 1997