So as we talked about it, it's not just enough to be measuring the average levels,
but we also want to be looking at variation.
And particularly, equity or equality.
So one of those most commonly used measures of
equity is called the concentration index.
So let's take a look at how this works.
So I'm going to take one measure of health inequities using children that
are underweight as the measure of health status.
And what you see at the bottom is the percent of the population or
percent of the sample ranked by income, from the poorest to the 25%,
to the 50th and by the time you hit 100, that's the whole population.
And you think of them listed in a row according to how poor they are from
poorest to richest.
And on the top you see what cumulative proportion of that sample includes
children that are underweight.
Now looking at this graph, there is a line that goes diagonally right across,
that 45 degrees.
Now, if things were perfectly equal, if there's a equal probability of
a child having underweight by each income group, you'll follow that line.
So in other words, the poorest 25% of the sample will have
25% of the underweight children, 50%, 50% and so on.
What you see here is the slopes for two different periods of time.
The green one was in 1987 and you see a bit of a bulge that goes above the line.
And then the orange one is in 1994, where the bulge went even further away.
So how do we use this to measure inequity?
Well, what we do for the concentration index is we take that area between
that curve and the 45 degree line.
The green one in 1987 and the orange in 1994,
double that area and that becomes the measure.
The measure goes from minus 1 to plus 1, and
by convention if the concentration index is put at less than 0 when
the variable is higher, are more common among the poor.
So what you see here is, in first case,
it's a number that is going to be less than 1 in 1987 and
then by 1994 it's going to be even larger, but in a negative direction.
So that number is going to be even closer to minus 1.
What you see means that in terms of proportional children that
are underweight, more of them become concentrated among the poor.
So that this particular measure of underweight children has become more
inequitable over time.
Now, that doesn't tell you whether the total number of children are less or more,
because in fact in this program, there are fewer children that were underweight.
They just happened to be more likely among the poor than they were in 1994,
than they were in 1997.
So the concentration index is a common measure.
Here's another view of a concentration curve,
this time showing four different kinds of health services.
This is a study that we did in India back in 2001.
The red diagonal line is the equal line and
what you see here is that the immunizations in outpatients care through
primary healthcare is above the line.
In other words, using those services is more likely to be
used by the poor than by the rich or by wealthier populations.
Whereas on the other hand, outpatient care at the hospital is most likely to be
used among wealthier populations and the inpatient hospital care
is the next one in terms of likely to be used by wealthier patients.
So in this case, you would say that because the benefits in this case of
using these services is a good thing, having a negative concentration index for
use of services is a good thing.
Immunizations and outpatient care through primary
health services would be pro-poor with a concentration indices less than 1.
And the other services would be pro-rich in terms of its distribution of use and
would have concentrations in the indices above 1.
Now there's nothing inevitable about the concentration index and
whether a health system is more pro-poor or not.
Here what we've done is a study showing concentration indices of public
financing of health, but it's sometimes called benefit incidence analysis.
Which looks at all of the public financing in use of health services in a country.
And each one of these bars represents a country in the green.
The brown bars actually represent states in India.
And what you see is a large distribution,
because benefit in this case is a good thing.
You'll see that being above 1 would be pro-rich and
being minus 1 is pro-poor, for taking public funding for health.
And what you see is a large variation, some of which are pro-poor in this
distribution and some of which are pro-rich and even within India,
the average one there is actually shown in yellow-brown.
But even in India, the states show very different levels of equity.
So again, this is a way of looking at inequity across a health system and
you can apply this type of analysis to any kind of indicator where
you have data on both the outcome as well as the distribution among wealth or
incomes of the population.