Table 1.
Common approaches for operationalizing health inequities in program evaluation
| Measure of health inequity | Description | Indicator of inter-group or inter-individual inequalities | Indicator of health inequalities in relation to socioeconomic position | Able to reflect the difference principle (sensitive to the absolute transfer of measures of health or income/wealth between groups) | Indicator of health inequalities in relation to other population groupings of interest (race, sex and geographical location) | Example papers |
|---|---|---|---|---|---|---|
| Before and after measures within only the disadvantaged population | Comparison (difference or ratio) of the measure of health within only one group to demonstrate impact of a program on the measure of health | None | No | Does not reflect transfer of either health or income/wealth | No | Case (2001). Does money protect health status? Evidence from South African pensions (37) |
| Range | Comparison (difference or ratio) of the measure of health between population groups. The comparison group could be any of the groups. Also, include shortfalls which compare measure for each group to a norm before inter-group comparison. | Can be operationalized for both inter-group and inter-individual differences. It takes no account of the sizes of the groups being compared when used as a measure of inter-group difference | Yes | Could be used to reflect transfer of both health and income/wealth | Yes | Townsend and Davidson (1982). Inequalities in health: The Black report (43) |
| Lorenz curve and Gini coefficient | The Lorenz curve is the plot of cumulative proportions of the population ranked by health (from the sickest person to the healthiest) against the cumulative proportion of health. The Gini coefficient is twice the area between the Lorenz curve and the diagonal. It ranges from 0 to 1 (i.e. from complete equality to when all the health is concentrated in the hands of one person). | Inter-individual indicator | No | Does not reflect transfer of either health or income/wealth | Yes (when combined with stratified analysisa) | Le Grand and Rabin (1986). Trends in British health inequality (15) |
| Pseudo-Lorenz Curve and pseudo-Gini coefficient | Plot of cumulative proportions of population grouped by social class and ranked by their mean health status (from the group with lowest mean health status to the group with the highest mean health status) against the cumulative proportion of health. The pseudo-Gini coefficient is twice the area between the pseudo-Lorenz curve and the diagonal. It ranges from 0 to 1 (i.e. from complete equality to when all the health is concentrated in the hands of one group) | Inter-group indicator. It accounts for the population sizes of the groups being compared | No | Could be used to reflect transfer of health only | Yes | Leclerc, Lert and Fabien (1990). Differential mortality: some comparison between England and Wales, Finland and France (44) |
| Index of dissimilarity | The sum of the absolute difference of share of total population health and share of population for each group divided by two. | Inter-group indicator. It accounts for the population sizes of the groups being compared | No | Could be used to reflect transfer of health only | Yes | Pappas (1993). The increasing disparity in mortality between socioeconomic groups in the United States, 1960 and 1986 (45) |
| Concentration curve and index | The concentration curve plots the cumulative proportions of the population ranked by their socioeconomic position (beginning with the most disadvantaged to the least disadvantaged) against the cumulative proportions of health. The concentration index (ranges from −1 to +1; i.e. from health being concentrated in the hands of the most disadvantaged to the least disadvantaged) is twice the area between the concentration curve and the diagonal. | Can be operationalized for both inter-group and inter-individual differences. It accounts for the population sizes of the groups being compared when used as a measure of inter-group. difference | Yes | Could be used to reflect transfer of health. It does not typically reflect the transfer of income/wealth unless there is a related change in the income/wealth ranking | No | Wagstaff, Paci and van Doorslaer (1991). On the measurement of inequalities in health (41) |
| Slope and relative indices of inequality (SII and RII) | SII is the slope of the regression line of the mean health status for each group against the relative rank of socioeconomic status (not health status) beginning with the most disadvantaged. It assumes a linear relationship between the two properties. Both properties are weighted by the square root of the group size. RII is SII divided by the mean health status for the entire population. | Inter-group indicator. It accounts for the population sizes of the groups being compared | Yes | Could be used to reflect transfer of health. It does not reflect the transfer of income/wealth unless there is a related change in the income/wealth ranking | No | Mackenbach et al. (2008). Socioeconomic Inequalities in health in 22 European countries (46) |
| Other regression techniques | Regression of the health status measure against the measure that defines the population groupings (e.g. socioeconomic status). Assumes a linear relationship between the two properties. | Inter-group indicator. It accounts for the population sizes of the groups being compared when weighted least square regression is used and weights are defined by group size | Yes | Could be used to reflect transfer of both health and income/wealth | Yes | Valkonen (1989). Adult mortality and level of education: a comparison of six countries (47) |
a
Stratified analysis in this case implies computing Gini coefficient of health within each strata of the population grouping.