This paper describes a population health status index for health services research and planning purposes. The H-index uses data on average life expectancy at birth and percent of the population free from disability, however defined. It is useful in comparing the health status of health services areas relative to that of the more healthy areas selected to serve as the norm.
The statistical procedure used in deriving the H-index is centour analysis, by means of which the Euclidian distances of the service areas in the study sample in two-dimensional space to the centroid of the normative areas are reflected in the H values computed. The farther away from the centroid, the less resemblance the service area has to the norm and the lower its health status. A computational example with seven normative states and 10 states in the study sample is given.
In his monumental work entitled, Improvements in the Quality of Life: Estimates of Possibilities in the United States, 1974-1983, (Tedeckyj, 1975), the author stated the objectives of the health component of the quality of life to be: (1) long life, and (2) life characterized by physical and mental wellbeing. Long life is indicated by average life expectancy at birth, whereas physical and mental welibeing is reflected in the percent of the population with disabling conditions due to physical and/or mental disorders. While life expectancy and freedom from disabling conditions may not satisfy the definition of health by the World Health Organization as "a complete state of physical, mental and social weUbeing, not merely the absence of disease or infirmity", 0Vodd Health Organization, 1957), they are certainly two of the crucial, and more importantly, easily obtainable indicators of any population health status.
The necessity of considering both average life expectancy at birth and the percent of population free from disability in an overall population health status lies in the fact that the two health status indicators appear to measure discrete aspects of health that are not or only moderately associated. In fact, our data show only a correlation of 0.41 between the two indicators for the 17 states used in the computational example. Undoubtedly some of the states, particularly the normative states, have both high average life expectancies at