Text S1: Measures of catastrophic health expenditure and poverty
We measured the catastrophic healthcare expenditure, as well as the poverty related to
healthcare spending, using indicators and techniques adapted from O’Donnell et al. (1).
Catastrophic healthcare expenditure (CHE)
We measured the incidence, intensity, and distributional fairness (across income quintiles)
of CHE. We used household monthly income as a denominator in calculating the CHE
instead of the usually recommended household consumption expenditure, or non-food
expenditure, because we did not have data on the latter.
Headcount
Headcount is the percentage of households whose monthly OOP expenditure, as fraction of
monthly household income, for outpatient care (for chronic conditions) exceeded certain
threshold. Most commonly accepted and used threshold in literature has been 10% at which
households are usually forced to cut down their subsistence needs (2). We calculated the
headcount at four different thresholds i.e. 5%, 10%, 15%, and 20% using the following
formula,
1
𝐻𝑒𝑎𝑑𝑐𝑜𝑢𝑛𝑡 = (𝑁) ∑𝑁
𝑖=1 𝐸𝑖
where 𝐸 is an indicator equal to one if 𝑇𝑖 ⁄𝑋𝑖 > 𝑧 and zero otherwise, 𝑇𝑖 is the OOP
expenditure by a household 𝑖, 𝑋𝑖 is the income of a household 𝑖, 𝑧 is the given catastrophic
threshold, and 𝑁 is the sample size.
Overshoot
Headcount only suggests the percentage of households that spent OOP beyond a particular
threshold but does not give an idea on how far (intensity) they spent beyond the threshold.
Overshoot measures the degree by which an average OOP expenditure (in entire sample)
crossed the given catastrophic threshold. We measured the overshoot by using the
following formula,
1
𝑂𝑣𝑒𝑟𝑠ℎ𝑜𝑜𝑡 = (𝑁) ∑𝑁
𝑖=1 𝑂𝑖
where the excess payment of household 𝑖 is defined as 𝑂𝑖 = 𝐸𝑖 ((𝑇𝑖 ⁄𝑋𝑖 ) − 𝑧).
Mean positive overshoot
Unlike the overshoot that uses all the households as denominator, the Mean Positive
Overshoot (MPO) uses only those households, which have actually experienced CHE, as
the denominator. Hence MPO measures the degree by which the average OOP expenditure
by households that have experienced catastrophe has exceeded the given catastrophic
threshold. We measured the MPO by using the following formula,
𝑀𝑒𝑎𝑛 𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒 𝑜𝑣𝑒𝑟𝑠ℎ𝑜𝑜𝑡 (𝑀𝑃𝑂) = 𝑂𝑣𝑒𝑟𝑠ℎ𝑜𝑜𝑡/𝐻𝑒𝑎𝑑𝑐𝑜𝑢𝑛𝑡
Hence if household 𝑖 experienced the CHE, it would have spent (𝑀𝑃𝑂𝑖 + 𝑧) percentage of
the household income on healthcare.
Concentration curve & index
Concentration curve and index help to understand the distribution of CHE across the
income quintiles. Concentration curve above the 45-degree line (line of equality) leads to
negative value for concentration index and suggests disproportionately higher concentration
of catastrophe among the poor households and vice versa. When the concentration index is
zero, it suggests the absence of the income-related inequalities in distribution of CHE.
Concentration index has been calculated using the following formula,
𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 𝑖𝑛𝑑𝑒𝑥 (𝐶𝐼)
= (𝑝1𝐿2 − 𝑝2𝐿1) + (𝑝2𝐿3 − 𝑝3𝐿2) + (𝑝3𝐿4 − 𝑝4𝐿2) + (𝑝4𝐿5 − 𝑝5𝐿4)
where 𝑝 is the cumulative percentage of households ranked by their monthly income, 𝐿 is
the cumulative percentage of households experiencing catastrophe for the corresponding 𝑝.
Numbers (1 to 5) suggests the relevant income quintile.
Poverty measures
Poverty headcount ratio
It is the percentage of individuals in the sample population whose gross per-capita monthly
income is less than the poverty line. We also measured this ratio using the per-capita
monthly income net of the OOP payments. This way we can get the percentage of
individuals who fell below the poverty line due to OOP payments for outpatient care. We
have considered the poverty line (PL) of INR 599.66 per capita per month as specified for
the urban Karnataka (3). Poverty headcount ratio gross of the OOP expenditure was
calculated using the following formula,
𝑔𝑟𝑜𝑠𝑠
∑𝑁
𝑖=1 𝑠𝑖 𝑝𝑖
𝑃𝑜𝑣𝑒𝑟𝑡𝑦 ℎ𝑒𝑎𝑑𝑐𝑜𝑢𝑛𝑡 𝑟𝑎𝑡𝑖𝑜
=
∑𝑁
𝑖=1 𝑠𝑖
𝑔𝑟𝑜𝑠𝑠
where 𝑝𝑖
= 1 𝑖𝑓 𝑥𝑖 < 𝑃𝐿 and 0 otherwise, 𝑥𝑖 is the per capita income of the household 𝑖,
𝑠𝑖 is the size of the household 𝑖, and 𝑁 is the number of the households in the sample that
incurred OOP expenditure.
𝑔𝑟𝑜𝑠𝑠
Similarly, the poverty head count ratio net of the OOPE was calculated by replacing the
𝑔𝑟𝑜𝑠𝑠
𝑝𝑖
with 𝑝𝑖𝑛𝑒𝑡 = 1 𝑖𝑓 (𝑥𝑖 − 𝑇𝑖 ) < 𝑃𝐿 (and 0 otherwise) in the above formula. Here 𝑇𝑖 is the
per capita OOP expenditure by the household 𝑖.
Mean poverty gap
Mean poverty gap measures how far (intensity of poverty) individuals fell below the poverty
line. It uses the entire sample as the denominator and hence provide an average extent (in
currency) by which individuals fell below the poverty line. Mean poverty gap gross of the
OOP expenditure was calculated by using the following formula,
𝑔𝑟𝑜𝑠𝑠
∑𝑁
𝑖=1 𝑠𝑖 𝑔𝑖
∑𝑁
𝑖=1 𝑠𝑖
𝑔𝑟𝑜𝑠𝑠
= 𝑝𝑖
(𝑃𝐿 − 𝑥𝑖 ).
𝑀𝑒𝑎𝑛 𝑃𝑜𝑣𝑒𝑟𝑡𝑦 𝑔𝑎𝑝 𝑔𝑟𝑜𝑠𝑠 =
𝑔𝑟𝑜𝑠𝑠
where 𝑔𝑖
Similarly mean poverty gap net of OOPE was calculated by using the above formula but
𝑔𝑟𝑜𝑠𝑠
replacing 𝑔𝑖
with 𝑔𝑖𝑛𝑒𝑡 = 𝑝𝑖𝑛𝑒𝑡 (𝑃𝐿 − (𝑥𝑖 − 𝑇𝑖 )).
Mean Positive Poverty Gap (MPPG)
Unlike the mean poverty gap that uses entire sample as denominator, the MPPG uses only
those individuals who are (or fell) below the poverty line as the denominator. We measured
MPPG by using the following formula,
𝑀𝑒𝑎𝑛 𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒 𝑝𝑜𝑣𝑒𝑟𝑡𝑦 𝑔𝑎𝑝 = 𝑀𝑒𝑎𝑛 𝑝𝑜𝑣𝑒𝑟𝑡𝑦 𝑔𝑎𝑝⁄𝑃𝑜𝑣𝑒𝑟𝑡𝑦 ℎ𝑒𝑎𝑑𝑐𝑜𝑢𝑛𝑡 𝑟𝑎𝑡𝑖𝑜
We calculated the mean positive poverty gap gross as well as net of the OOP expenditure
by using respective values of the mean poverty gap and the poverty headcount ratio.
In order to enhance comparability of these measures across the studies using different
poverty lines and/or currency units, these measures were normalised over the poverty line
(by dividing them with the poverty line).
References
1.
O’Donnell O, Van Doorslaer E, Wagstaff A, Lindelow M. (2008) Analyzing health
equity using household survey data: A guide to techniques and their implementation.
Washington DC: The World Bank. 220 p.
2.
Van Doorslaer E, O’Donnell O, Rannan-Eliya RP, Somanathan A, Adhikari SR et al.
(2006) Effect of payments for health care on poverty estimates in 11 countries in Asia:
an analysis of household survey data. Lancet 368(9544): 1357–1364.
3.
Planning Commission (2007) Poverty estimates for 2004-05. New Delhi: Press
Information Bureau, Government of India. 1–2. Available from:
http://pib.nic.in/newsite/erelease.aspx?relid=26316 Accessed 2012 Apr 1.