Introduction
Review
Methodology of measuring horizontal equity
Method of concentration curve and concentration index
Concentration curve
Concentration index
Method of Lorenz curve and Gini-coefficient
Lorenz curve
Gini-coefficient
Theil index
Atkinson index
Chi-square value method
Index of dissimilarity
Conclusions
Measure | Definition | Complexity of calculation | Application | Required data | Benefits | Caveats |
---|---|---|---|---|---|---|
Concentration curve, concentration index | Calculate and compare cumulative percentage of population (ranked by socioeconomic factors) and health resource. CI , COV(X,H) = E(XH)-E(X)E(H) | Complex but aided by statistical software | Systematic assessment, and can be a rough estimation on equity of differences between different regions. | Income of individual, health resource of individual | -not only represent overall inequity, also reflect accurately which social classes allocated with more resource and which less via positive or negative CI value | -incapable of considering the other variables, especially the resource delivery itself. |
-socioeconomic factors are taken into consideration when measure the inequity. And which is very sensitive to different social classes | ||||||
-simple to calculate | ||||||
-simple to interpret when combine with corresponding curve | ||||||
-the concentration index must be interpreted with the curve | ||||||
-does not allow for within or between income group comparisons | ||||||
Lorenz curve, Gini index | Calculate and compare cumulative percentage of population (ranked by how much resources obtained) and health resource. | Complex but can aided by statistical software | Systematic assessment, and can be a rough estimation on equity of differences between different regions. | Health resource of individual, total health resource, population of area | -a graphical representation of allocation inequality that can be compared over time and between geographic areas | -incapable of showing different kinds of inequality represented by various shapes of Lorenz curves [34] |
-simple to calculate | ||||||
-data readily available | ||||||
-can be calculated for individual and household level data | ||||||
-easily interpreted when combine with Gini coefficients | ||||||
-does not emphasize inequalities in the top or bottom of the spectrum (polarization) | ||||||
-shows the direction of allocation redistribution but does not indicate where the redistributions are occurring | ||||||
-does not allow for within or between income group comparisons | ||||||
-overlook socioeconomic factors | ||||||
Theil index | Calculate the equity of health resource by population (area) in each region. | complex | Measure equity of the allocation of health resources between different regions or the units. | Population of units or regions, total population , health resources in units or regions, total resource | -shows the contributions to inequality by within group and between group components | -complex to calculate and interpret. |
-varies greatly when the distribution varies regardless of the change in distribution occurs at the top, middle or bottom | ||||||
-high sensitivity to the efficiency of health resource allocation | ||||||
-resource redistributions will impact the calculation irrespective of whether the redistribution takes place between top and bottom or top and middle | ||||||
-cannot directly compare populations with different sizes as calculation is dependent on number of individuals in the population or group | ||||||
Atkinson index | Calculate the health resources of ith region and the proportion of population in which people get the resources. ,if ε ≠ 1,,if ε = 1 | complex | Assess the inequity of allocation, address needs of inequity assessment in health benefits analysis | Health resource of ith region, the proportion of population in ith region who get the resource, inequality aversion ε | -incorporates a sensitivity parameter directly into the equation. | -sensitivity parameter means that a subjective judgment has been made about inequality |
-not intuitive | ||||||
Chi-square Value Method | Calculate the actual and theoretical frequency of health resources. | Easy when analyze aided with statistical software | Assess the difference between actual allocation of health resource with the expected allocation | Actual resources in ith region, the total resource, expected frequency of health resource allocation | -sensitive to reflect the inequity of allocation | -always need to standardize the data, otherwise may influence the results |
-reveal the trend of equity over time | ||||||
-the judgment is subjective when based on a certain significance level α | ||||||
index of dissimilarity | Calculate the health resources and population in each socioeconomic level (region). | easy | assess the differences of resource allocation in different economic level(region), and calculate the degree of variance | Resource in jth region(or in a certain socioeconomic level), the population in jth region | -can know the differences between the situation of health resource allocation in each region (level) and the proportion of the population in relative region (level) | -can’t reflect the socioeconomic status influence on health resource allocation. |
-not intuitive |
Example: measuring the equity of health resource allocation in Chongqing (China)
Year | Allocation by population | Allocation by area | ||||
---|---|---|---|---|---|---|
Beds | Doctor | Nurse | Beds | Doctor | Nurse | |
1998 | 0.3300 | 0.2373 | 0.4407 | 0.5709 | 0.4934 | 0.5906 |
2002 | 0.3456 | 0.2728 | 0.4099 | 0.5832 | 0.5273 | 0.6389 |
2007 | 0.3071 | 0.2725 | 0.4141 | 0.5640 | 0.5326 | 0.6352 |
2012 | 0.2389 | 0.2843 | 0.3715 | 0.5019 | 0.5494 | 0.6049 |
Year | Allocation by population | Allocation by area | ||||
---|---|---|---|---|---|---|
Beds | Doctor | Nurse | Beds | Doctor | Nurse | |
1998 | 0.0778 | 0.0438 | 0.1149 | 0.2559 | 0.1935 | 0.3195 |
2002 | 0.0847 | 0.0537 | 0.0980 | 0.2675 | 0.2160 | 0.3362 |
2007 | 0.0670 | 0.0543 | 0.1108 | 0.2531 | 0.2209 | 0.3315 |
2012 | 0.0415 | 0.0579 | 0.0859 | 0.1982 | 0.2408 | 0.3013 |