Public Retreat,Private Expenses,and Penury

Abstract

Health care can be expensive for the un-insured, often constituting a potential poverty trap. Urban India is particularly vulnerable to this possibility given the greater demand for health, absence of a structured health care system, overburdened public institutions, ubiquitous, and unregulated private health care market and the generic paucity of public funds. Using nationally representative household level data for two points of time and a suitable alteration of an existing methodology, this article computes the degree and depth of impoverishment from out of pocket medical expenses, and its variation across states and select socioeconomic characteristics. Roughly 6 percent of the urban population or about 18 million people face impoverishment entirely due to out of pocket medical expenses in India. There are substantial inter-state and inter-group variations in the incidence of this burden. The findings are potentially crucial as India prepares to embark on its journey toward universal health coverage.

Keywords

health financing poverty urban out-of-pocket

Introduction

Public policy in India witnessed a reorientation after 1991 with the setting in of the structural adjustment program. One of its tenets was fiscal austerity, which resulted in declining public expenditure on social services (Dev & Mooji, 2002). Public finances over the years clearly reveal the government’s tendency to withdraw these services often under the garb of public–private partnerships (PPP) or through outright privatization. It may be argued that the state’s withdrawal manifested more unambiguously in urban India than in rural India. There are two main reasons for this manifestation. First, in the early 1990s, only a quarter of the total population resided in urban areas and a fraction of them voted; this meant that the electoral backlash from the urban populace on sensitive issues of public service delivery was largely insignificant. Second, the relatively higher per capita income in urban areas created the perception that urban inhabitants did not deserve public subsidies on these services. Urbanization, on the other hand, continued unabated as an inevitable outcome of the faster rates of growth to which the economy transited in the post-liberalization period. Apart from this “pull” factor, the “push” of a traditional rural sector also contributed to it. The net result was a huge demand–supply gap in basic services, which the private sector was allowed to fill. While this has arguably kept services available, keeping these services affordable has emerged as a crucial policy challenge.

The health sector has possibly borne the greatest brunt of this perceived shift. India currently spends around 4 percent of its GDP on health. Public (central, state, and local governments combined) spending on health, however, accounts for just 1 percent of GDP; private and external sources spend the remaining 3 percent. The share of public expenditure in total health expenditure is around 20 percent, while households account for another 70 percent of total health spending, almost all of which is in the form of out-of-pocket (OOP) expenses.¹ Only those employed in the formal sectors – a minority in India – can avail reimbursement in any form.² Such high OOP spending by households has certain adverse implications. While access to health care is reduced considerably for some,³ others who opt for treatment face a catastrophic burden of health care expenditure. Standard measures of poverty, however, cannot capture this aspect. They might inaccurately categorize a household as non-poor simply because high medical expenses (financed often through borrowing and distress sale of household assets) raise its total spending above the poverty line while pushing spending on food, clothing, and shelter below subsistence levels. Another dimension to the financial burden of illness extends beyond the movement of a household in and out of poverty in the very short run. Treatment costs are not necessarily paid out of current income; a household might have to dis-save, borrow, accept contributions, sell assets, reduce other non-discretionary expenditures like food, etc. to finance treatment cost. Thus, indebtedness and distress sales of assets are some of the other potential outcomes accruing to households from a health financing system that relies excessively on private OOP expenditure.

Against this backdrop, the article attempts an estimation of treatment cost-induced impoverishment in urban India. Given the large socioeconomic variations intrinsic to the country, the analysis is extended to determine the inter-state as well as inter-group (religion, caste, etc.) variations in the incidence of this phenomenon. The entire analysis is carried out for two points of time – separated by a decade – to understand the inter-temporal variations in the financial burden of morbidity in urban India. This is particularly important since the urban health system in India has undergone substantial changes in the past decade or so with potential implications for health care financing and its aftermath.

The article has seven sections and is organised as follows. Section 2 attempts to unravel the complexity of the urban health sector in India and argues that the phenomenon of illness-induced impoverishment might be more prevalent in urban areas. Section 3 reviews the literature on health financing by households and its effects. Section 4 discusses the advancements and modifications in this article with respect to the data and methodology as compared to previous studies. Section 5 elaborates on the data and methodology adopted. Section 6 presents and discusses the degree of illness-induced impoverishment disaggregated across states and select socio-economic characteristics while Section 7 concludes with a policy perspective.

Health Financing in Urban India – Public Apathy and Private Choices

The financial burden of health expenses is invariant of the place of residence (rural or urban). However, the urban health scenario in India today is arguably more complex because of both supply-side and demand-side characteristics. On the supply side, there is no urban equivalent of the standard three-tier health system that exists in rural India. Urban India has a mixed health system where ministries of health, private insurance, social insurance, and targeted schemes coexist to serve different sections of the population. For example, in a large urban centre like Delhi, health planning, and service delivery is done independently by the Municipal Corporation of Delhi (MCD), the New Delhi Municipal Corporation (NDMC), the Government of Delhi, Employees State Insurance (ESI) dispensaries, Central Government Health Scheme (CGHS) clinics, hospitals run by the Ministry of Railways, defense etc. If we add to these large public hospitals like the All India Institute of Medical Sciences (AIIMS), numerous private hospitals and clinics, and independent NGOs, what emerges is an extremely complex system that requires an unlikely synergistic effort to achieve homogeneity in the provision of health care to the urban populace at standardized costs. Such a system tends to fragment, increase administrative costs, limit pool sizes, and undermine both equity, and efficiency objectives. Urban private providers generally have better access to technology but unregulated prices and quality of services. In addition, government-run hospitals and clinics in urban areas that are starved of public funds have been charging citizens more for medicines, diagnostic procedures, and surgical aids in recent years, and therefore user charges and other effective charges to consumers have increased concomitantly even in the public health system.

On the demand side, while the rural mass in India – and especially the poor – are more likely to forego treatment because of the lack of awareness and illness perception and dearth of medical facility in the vicinity, urban India tends to place a premium on health – due to higher education levels and mass media penetration that generates awareness. This results in even poor households willing to spend to ensure minimal health care. While this ensures the demand for medical care in some sense, the issue of financial burden looms large on poor households. Other exogenous factors like the higher cost of living, increased exposure to accidents, lifestyle issues and poor environmental condition makes the urban mass increasingly vulnerable to indisposition and the financial burden of health care. The discussed demand and supply characteristics of the urban health therefore constitute a strong case for a detailed investigation of the phenomenon of illness-induced impoverishment, with a special focus on urban India.

Financial Burden of Illness – A Review of Literature

In recent years, uncovered health expenses have emerged as one of the predominant reasons for both short-run and long-run alterations in the standard of living of households in India. The acknowledgment of treatment cost as a potential producer of poverty has inspired researchers to study the “poverty ratchet” (Chambers, 1983) or the “medical poverty trap” (Whitehead, Dahlgreen, & Evans, 2001) in different country settings and under dissimilar health systems. The existing literature on the financial implications of health care has largely used a couple of proxy measures to compute this burden – catastrophic health expenditure and health care cost-induced impoverishment. The concept of catastrophic payments is defined as the circumstances when OOP payments cross some threshold share of household expenditure and is considered a major concern in the health financing system of any country (Berki, 1986; CMH, 2001; Kawabata, Xu et al., 2002; Meesen & Zang, 2003; OECD & WHO, 2003; Pradhan & Prescott, 2002; Wagstaff & van Doorslaer, 2003; Whitehead et al., 2001; Wyszewianski, 1986; Xu, Evans et al., 2003). It is acknowledged that the choice of threshold commonly chosen – 10 percent of total expenditure – is somewhat arbitrary (Pradhan & Prescott, 2002; Ranson, 2002; Wagstaff & van Doorslaer, 2001). The rationale is that this is an approximate threshold at which the household is forced to sacrifice other basic needs, sell productive assets, incur debt or be impoverished (Russell, 2004). A recent WHO article, using survey data from 89 countries, finds that 3 percent of households in low-income countries, 1.8 percent of households in middle-income countries and 0.6 percent of households in high-income countries incur catastrophic health expenditures (Xu et al., 2007).

Soaring health care expenditure often affects the magnitude and pattern of household consumption. When a member falls ill, the household faces several different costs (treatment cost, transportation cost, opportunity cost of care giving, etc.) and adopts diverse strategies to finance them. While the OOP expenses set in a “real time” reduction in the standard of living, the coping strategies very often turn out to be potential poverty traps. The chain of events has often been termed as the “poverty ratchet” (Chambers, 1983) or the “medical poverty trap” (Whitehead et al., 2001). Gertler and Gruber (2002) studied the impact of health shocks on households’ consumption patterns in Indonesia, providing evidence that illness reduced labor supply and household income. Similarly, Wagstaff (2005) finds evidence that health shocks are associated with a reduction in consumption in Vietnam, in particular for the uninsured. Dercon and Krishnan (2000) show that in Ethiopia the consumption risks associated with health shocks are not borne equally by all household members. In addition, estimates of the financial burden of illness are available for at least six Latin American countries (Baeza & Packard, 2005), China (Lindelow & Wagstaff, 2005), Thailand (Limwattananon, Tangcharoensathien & Prakongsai, 2007), and 14 Asian countries and territories (van Doorslaer, O’Donnell, & Rannan-Eliya, 2007).

Studies on India have found average expenditure on medical care rising invariably with monthly per capita consumer expenditure or income of the household⁴ (NSSO, 1992, 1998; Rajaratnam, Abel, Duraisamy, & John, 1996; Visaria & Gumber, 1994). However, medical expenditure as a proportion of total resources at the household’s disposal was much lower for the rich (Krishnan, 2000). Estimates show health expenditure as a percentage of annual income varying from 3 percent in the richest 20 percent of the households to 12 percent in the bottom 20 percent of the households (Gumber, 2002). A study of 35 villages in Rajasthan found that health and health expenses were one of the main causes of 85 percent of all cases of impoverishment. One-half to two-thirds of all poor households mentioned ill health and health expenses as a contributory cause (Krishna, 2004). Such impoverishment is of even greater concern given the evidence from another detailed study in Rajasthan that shows that health care purchased is often of poor quality – even harmful (Banerjee et al., 2004). In India, more than 37 million people went below the poverty line in 1999–2000 as per the $1 norm because of OOP payments (van Doorslaer et al., 2005) – in addition to those already below the poverty line and pushed further into acute poverty because of OOP payments. Two other studies (Bonu, Bhushan, & Peters, 2007; Garg & Karan, 2005) estimated that roughly 3.25 to 3.5 percent of the population became poor because of health care payments. A more recent study with NSSO data reports that after adjusting for the sources (borrowings, contributions and sale of assets etc.) of OOP expenditure, 63.22 million individuals or 11.88 million households were impoverished due to healthcare expenditure in 2004 (Berman, Ahuja, & Bhandari, 2010). Based on a case study of slums in Delhi, Chowdhury (2011) shows that even the cost of outpatient treatment impoverished 13 percent of the sampled households. Two more studies with nationally representative data from India find that close to 4 percent of the population fall below the poverty line because of health care payments (Ghosh, 2012; Shahrawat & Rao, 2012). The latter study also found that OOP payments aggravated the prevalence and intensity of poverty in India between 1993–1994 and 2004–2005.

Points of Departure from Earlier Studies

This article departs from earlier studies on at least three aspects related to data, methodology, and the level of analysis.

First, most studies of the financial burden of illness in India have used the NSSO consumption expenditure survey (CES) data. This study uses the NSSO unit record data on Morbidity and Health Care (MHC) because this data detail the parameters on treatment cost extensively – unlike the CES data. The MHC round also collects information on the socioeconomic characteristics of households – including consumption expenditure – which allows us to calculate poverty ratios from the same data.

Second, the methodology of computing illness-induced impoverishment by simply subtracting OOP expenses from consumption expenditure and comparing it with the existing poverty line ignores the fact that the poverty line itself may include some non-food expenditure, no matter how inadequate. In other words, some amount of health expenditure is implicit in the poverty line that needs to be deducted and then compared with the household expenditure net of OOP payments to get a more robust estimate of the phenomenon (O’Donnell, van Doorslaer, Wagstaff, & Lindelow, 2008).

Third, India – especially urban India – displays large variations across states and in the socioeconomic profile of its population. In view of this, the article looks into the differential incidence of the financial burden of OOP expenses across states as well as socioeconomic categories for the two points of time.

Data and Methodology

Unit record (household-level) data from two successive National Sample Survey rounds (1995–1996 and 2004) on morbidity and health care form the database for the study. These are thin (small sample) rounds separated by a decade, which provide the opportunity to examine the impact of possible changes in the health system of the country on financial implications of morbidity. Table 1 compares certain aspects of the two rounds.

Table 1.

A Summary of the Data Sources

	NSS 52nd Round, July 1995–June 1996	NSS 60th Round, January–June, 2004
Number of urban blocks surveyed	4991	2668
Number of urban households surveyed	49658	26566
Percentage reporting ailment in last 15 days	5.4	9.9
Percentage of ailing who were treated during last 15 days	9.1	8.9
Percentage of treated ailments receiving non-hospitalized treatment from government sources	20	19
Average total expenditure (₹) for non-hospitalized treatment per ailment	175	306
Percentage of persons hospitalized during a reference period of 365 days	2	3.1
Percentage cases of hospitalization treatment from government sources	43.1	38.2
Average total expenditure (₹) per hospitalization	3921	8851

Source: Compiled from Report No. 441 (1998) and Report No. 507 (2006) of National Sample Survey Organisation.

Note: The annual average exchange rates were 33.4498 in 1995–96 and 44.9315 in 2004–05 relative to the US Dollar, according to the Reserve Bank of India, available at http://rbidocs.rbi.org.in/rdocs/Publications/PDFs/72784.pdf (last accessed on October 13, 2014).

Morbidity levels were generally higher in 2004 than in the previous round. This was true for hospitalization as well as non-hospitalized ailments. There has been a visible decline in the utilization of government sources for treatment. Treatment cost has nearly doubled in nominal terms within the decade. These observations presumably bear essential implications for the issue of the financial burden of illness. Tables A1 and A2 in the appendix present a disaggregated summary of OOP health expenses for the two points of time. The share of OOP health expenditure in total consumption expenditure of the household was estimated at around 8 percent, nearly double its share in the 1990s.

The CESs of the NSS place the health share of household budget in urban India at around 5 percent. Therefore, one is tempted to say that regular CESs underestimate the health care costs of the households and subsequently the resultant economic burden. The distribution of the health share in household budget was pro-rich in that households in the higher expenditure quintiles spent a smaller proportion of their total resources on health care – despite the higher perception of ailments among high-income groups and the predominance of lifestyle diseases, which necessitates expensive treatment and a general preference for private sources of treatment within this group that involves higher costs.

The methodology is an adaptation of Wagstaff and van Doorslaer’s (2003) attempt to estimate illness-induced impoverishment for Vietnam with a relevant modification.

Consider a household “i.”. Suppose,

‘S_i’ = size of the ith household.

‘MPC_i’ = monthly per capital total consumption expenditure of the ith household,

H_i = monthly per capita health expenditure of the ith household.

Also let “L” be the poverty line that the household faces. In order to measure poverty gross of health care payment, we define

\begin{matrix} P_{i}^{gross} 1 if {MPC}_{i} < L \\ 0, otherwise \end{matrix}

(1)

Now if N is the number of households in the sample, an estimate of poverty headcount ratio gross of health payments is given by,

{HC}^{gross} \frac{_{i 1}^{N} S_{i} P_{i}^{gross}}{_{i 1}^{N} S_{i}}

(2)

Again, individual poverty gap gross of health payment is given by,

G_{i}^{gross} P_{i}^{gross} (L {MPC}_{i})

(3)

The mean of this gap in rupee terms is given by,

G^{gross} \frac{_{i 1}^{N} S_{i} G_{i}^{gross}}{_{i 1}^{N} S_{i}}

(4)

Figure 1 illustrates the discussed methodology. The case displayed in the figure assumes implicitly that the relative position of households in the gross and net of OOP expenditure distribution does not change. In the standard case (pre-payment), headcount is H^gross and poverty gap is equal to the area “A”. In the special case (post-payment), poverty headcount increases to H^net and the gap is now given by the sum of “A,” “B,” and “C”. Area “B” represents the increase in the intensity of poverty due to health care payments for those households who were already poor based on pre-payment monthly per capita expenditure (MPCE). Similarly, area “C” stands for the addition to the poverty gap due to new entrants into poverty after paying for health care. The value of (H^net – H^gross) corresponds to the fraction of households considered non-poor despite their MPCE net of OOP payments for health care being below the poverty line.

However, the Indian poverty line already contains an allowance for non-food components – no matter how inadequately it represents the true state of affairs. For methodological accuracy, we need to adjust the poverty line by subtracting the health component of this non-food allowance from the existing poverty line. Suppose households⁵ in and around the poverty line incur an average per capita health expenditure denoted by X.

The adjusted poverty line is then given by, L* = L – X

In order to estimate poverty net of health payments we first define $P_{i}^{net}$ such that,

\begin{matrix} P_{i}^{net} 1 if {(M P C}_{i} H_{i}) < L * \\ 0, o t h e r w i s e . \end{matrix}

(5)

Finally, the head count net of health payments is obtained by replacing $P_{i}^{gross}$ in equation (2) with $P_{i}^{net}$ such that,

{HC}^{net} \frac{_{i 1}^{N} S_{i} P_{i}^{net}}{_{i 1}^{N} S_{i}}

(6)

The individual poverty gap net of health payments is given as,

G_{i}^{net} P_{i}^{net} {L * {(M P C}_{i} H_{i})}

(7)

One of the central objectives of the current analysis is to make inter-state comparisons in impoverishment induced by treatment cost. However, state-specific poverty lines being different, a meaningful comparison would require normalization of the poverty gap, which is done by dividing the gap by the respective adjusted poverty lines.

Thus normalized poverty gap,

{NG}^{g r o s s / n e t} \frac{G^{g r o s s / n e t}}{L *}

(8)

So the relevant modification to the original methodology (Wagstaff & van Doorslaer, 2003) that is made in this article is to deduct the health expenses corresponding to the expenditure class that contains the poverty line, from the existing poverty line to arrive at an adjusted poverty line. The rationale for such an adjustment is that the poverty line, as defined in India, already includes an allowance for health expenses, albeit a meagre one. The distribution of consumption expenditure net of OOP payments is then compared with this adjusted poverty line to estimate the degree of impoverishment.

Figure 1.

Source: World Bank (2008).

Methodological Caveat

A particularly sensitive issue with the methodology illustrated above is that of downward adjustment of the poverty line. The reason why the downward revision might be a subject for disagreement is that poverty lines in India as given by the Planning Commission are already considered low (Patnaik, 2006; 2007). It has been rightly suggested that these lines do not adequately reflect the minimum expenditure required for a decent standard of living – so much so that they have often been termed as “starvation lines” (Guruswamy & Abraham, 2006). Therefore, any estimate of poverty headcount based on these lines is bound to underestimate the true extent of the poverty prevailing in India. The basis of this criticism is embedded in the methodology of construction of poverty lines. The current poverty line is simply a price-updated version of the poverty line in 1979. It is not based on the newer capability theories, nor has it been profoundly revised (Sen, 2005; Alkire, 2008). None of the governments that have come to power has shown the political will to redesign the consumption basket of the 1970s, incorporate contemporary consumption patterns and re-draw the poverty line and subsequently poverty ratios based on the consumption expenditure data from the current quinquennial rounds of the NSS.

However, the poverty line of 1979 was not merely based on subsistence food requirements; it contained a non-food component too, which was the expenditure on non-food items of the group of people who met the nutrition norms (i.e., 2400 kcals and 2100 kcals for rural and urban areas respectively). Thus, in a sense, the average total expenditure of households just satisfying the normative nutritional requirements has been used as the poverty line. Therefore, implicitly, this line considered the expected spending on health care of those in the region of food poverty. Although the health care needs of a household are highly stochastic and a seriously ailing person faces health care expenses well above the average age. Any robust attempt to recalculate poverty ratios net of health care payments must therefore be preceded by a readjustment of the poverty line by deducting the mean OOP health expenditure of the MPCE class within which the poverty line lies from the original poverty line. Such an adjustment of the poverty line, however, might push some households out of poverty if their spending on health care is less than the mean health expenditure, that has been deducted from the poverty lines. This ethical issue notwithstanding, one must recognize that the indicators of interest in the current context are the post-payment head count and gap and – more importantly – its deviation from the pre-payment figures that would indicate illness-induced impoverishment.

Table 2 shows the poverty line adjustment for the years 1995–1996 and 2004 respectively. Columns 2 and 5 display the original state specific poverty lines.⁶ Columns 3 and 6 give the total per capita OOP expenditure on the treatment of ailments for households belonging to the MPCE class that contains the poverty line for respective states. The adjusted poverty lines form the basis of the subsequent analysis on illness-induced impoverishment.

Illness and Impoverishment: Trends and Patterns

Table 3 displays the extent and depth of illness-induced impoverishment among the urban population of 15 major states at two points of time. The poverty head count under the “pre-payment” column implies the proportion of individuals in each state who are poor based on the adjusted poverty line. The “post-payment” poverty headcount column gives the proportion of individuals impoverished when their total health expenditure is netted out of total consumption expenditure. Therefore, the “difference” roughly gives the percentage of individuals who are impoverished exclusively due to the burden of OOP payments on health care; this is the indicator of interest in the current analysis. The same applies to the case of the poverty gap that calculates the depth of poverty gross and net of health care payments. However, since poverty lines are state-specific, it renders a comparison of poverty gaps impractical. Hence, a normalized poverty gap is calculated and expressed in percentage terms that make inter-state comparison possible.

Table 2.

Poverty Line Adjustments, 1995–1996 and 2004

State	1995–1996						2004
State	Original Poverty Line (₹) (2)	Monthly Per Capita Health Expenditure Of the MPCE Class Containing Poverty Line (₹) (3)	Adjusted Poverty Line (₹) (4) = (2) – (3)	Original Poverty Line (₹) (5)	Monthly Per Capita Health Expenditure Of the MPCE Class Containing Poverty Line (₹) (6)	Adjusted Poverty Line (₹) (7) = (5) – (6)
Andhra Pradesh	334.18	8.87	325.31	542.89	64.83	478.06
Assam	257.07	1.97	255.1	378.84	41.25	337.59
Bihar	282.21	10.19	272.02	435	51.94	383.06
Gujarat	358.37	12.82	345.55	541.16	50.79	490.37
Haryana	315.1	11.35	303.75	504.49	30.69	473.8
Karnataka	373.17	11.6	361.56	599.66	31.49	568.17
Kerala	354.91	20.32	334.59	559.39	109.85	449.54
Madhya Pradesh	374.09	20.86	353.23	570.15	30.88	539.27
Maharashtra	397.29	18.02	379.27	665.9	67.13	598.77
Orissa	354.82	17.6	337.22	528.49	52.56	475.93
Punjab	294.71	29.41	265.29	466.16	50.12	416.04
Rajasthan	335.11	12.45	322.66	559.63	55.78	503.85
Tamil Nadu	369.08	24.72	344.35	547.42	24.07	523.35
Uttar Pradesh	306.53	28.4	278.13	483.26	57.37	425.89
West Bengal	297.53	7.06	290.47	449.32	96.15	353.17
All India	338.15	26.14	312.01	538.6	45.08	493.52

Source: Computed from unit record data of NSS 52nd Round and 60th Round.

Out-of-pocket medical expenses entirely impoverished 18 million individuals or about 6 percent of the urban population – disconcertingly, a higher proportion than a decade ago (3.5 percent). However, the direction and degree of inter-temporal variation is along expected lines, since the quantum of public health services has reduced significantly and high-cost, private sources of treatment have proliferated in recent years. This has arguably been assisted by higher health-seeking behavior among urban masses.

If we look into the state-wise incidence of this burden, Kerala, Uttar Pradesh, and West Bengal displayed the highest difference between pre-payment and post-payment head count ratios in 2004. Apart from Haryana, all other states demonstrate an increase in the magnitude of “medical impoverishment” between 1995–1996 and 2004. The increase was highest for Kerala and West Bengal. Kerala is unique in terms of its extremely good health infrastructure – both public and private. The high literacy rates resulting in augmented morbidity perception have ensured a robust demand for health care. It is, therefore, uncertain whether the high level of medical impoverishment in Kerala is due to the composition of the health sector or the nature and quantum of demand for medical services, much of which might be perceptional rather than symptomatic. This is certainly not the case with other states like West Bengal, where the quality of publicly provided medical care has declined significantly in the past decade or so and consequently more and more people have been opting for expensive and burdensome private treatment.

Table 3.

Inter-state and Inter-temporal Variations in Illness Induced Impoverishment in Urban India, 1995–1996 and 2004

Sates	Poverty Head Count (%)						Poverty Gap (Rs)						Normalized Poverty Gap
	1995–1996			2004			1995–1996			2004			1995–1996			2004
	Pre-Payment (%)	Post-Payment (%)	Difference (%)	Pre-Payment (%)	Post-Payment (%)	Difference (%)	Pre-Payment (₹)	Post-Payment (₹)	Difference (₹)	Pre-Payment (₹)	Post-Payment (₹)	Difference (₹)	Pre-Payment	Post-Payment	Difference (%)	Pre-Payment	Post-Payment	Difference (%)
Andhra Pradesh	31.93	35.54	3.61	13.76	20.14	6.38	23.02	32.08	9.06	13.82	46.74	32.92	0.071	0.099	2.8	0.029	0.098	6.9
Assam	7.42	9.39	1.97	1.85	6.34	4.49	2.95	8.73	5.78	0.44	12.36	11.92	0.012	0.034	2.3	0.001	0.037	3.5
Bihar	30.94	34.89	3.95	19.24	24.52	5.28	18.58	26.21	7.63	21.23	48.96	27.73	0.068	0.096	2.8	0.055	0.128	7.2
Gujarat	18.57	21.33	2.76	6.23	11.15	4.92	12.59	20	7.41	5.29	27.31	22.02	0.036	0.058	2.1	0.011	0.056	4.5
Haryana	12.01	17.62	5.61	9.24	14.78	5.55	7.11	28.37	21.25	6.02	26.3	20.28	0.023	0.093	7	0.013	0.055	4.3
Karnataka	39.52	41.56	2.04	33.1	37.05	3.95	36.51	43.03	6.51	42.55	58.01	15.46	0.101	0.119	1.8	0.075	0.102	2.7
Kerala	26.33	30.97	4.64	11.09	22.67	11.58	17.52	25.49	7.97	9.96	53.43	43.47	0.052	0.076	2.4	0.022	0.119	9.7
Madhya Pradesh	42.08	45.84	3.76	42.62	49.06	6.44	32.59	52.73	20.14	54.59	89.26	34.67	0.092	0.149	5.7	0.101	0.166	6.4
Maharashtra	28.28	31.19	2.91	22.78	29.78	7	27.2	34.49	7.29	37.32	67.44	30.12	0.072	0.091	1.9	0.062	0.113	5
Orissa	38.47	41.4	2.94	24.79	31.54	6.75	31.91	39.84	7.93	30.61	49.4	18.79	0.095	0.118	2.4	0.064	0.104	3.9
Punjab	3.49	6.09	2.6	7.35	14.15	6.8	1.83	5.66	3.83	5.85	60.28	54.43	0.007	0.021	1.4	0.014	0.145	3.1
Rajasthan	21.47	23.68	2.22	24.28	28.69	4.41	10.69	15.07	4.38	17.06	41.26	24.2	0.033	0.047	1.4	0.034	0.082	4.8
Tamil Nadu	34.56	38.12	3.56	23.8	28.87	5.07	27.19	33.37	6.17	21.94	57.33	35.39	0.079	0.097	1.8	0.042	0.11	6.8
Uttar Pradesh	21.69	27.18	5.49	22.56	31.36	8.8	11.21	26.32	15.11	19.2	49.36	30.16	0.04	0.095	5.4	0.045	0.116	7.1
West Bengal	21.35	24.25	2.9	7.36	14.57	7.21	12.22	17.13	4.91	4.89	28.08	23.19	0.042	0.059	1.7	0.014	0.08	6.6
All India Urban	23.52	27.01	3.49	19.35	25.47	6.12	15.31	23.88	8.57	21.73	49.17	27.44	0.049	0.077	2.7	0.044	0.1	5.6

Source: Computed from unit record data of NSS 52nd Round and 60th Round.

The difference between the pre-payment and post-payment poverty gaps indicates an increase in the depth of poverty because of health care payments. After paying for health care, the poverty gap in urban India increased by ₹ 8.57 in 1995–1996 but by ₹ 27.44 in 2004. Therefore, the depth of poverty increased threefold in nominal terms after paying for medical care between the two periods. The normalized poverty gap is obtained by expressing the poverty gap as percentage of state-specific poverty lines. In percentage terms, too, the difference in poverty gap pre- and post-health care payment almost doubled for urban India between the two points of time. Haryana, Madhya Pradesh, and Uttar Pradesh demonstrated a higher percentage increase in the depth of poverty due to OOP expenses in 1995–1996. In 2004, Uttar Pradesh was also among the states with higher poverty-deepening effects of treatment cost along with Kerala, Tamil Nadu, and West Bengal.

The incidence of the financial burden of illness cannot be uniform across the population – especially in a country like India. This is because aspects like religion, caste, class, etc. play a dominant role in the access to medical services as well as the ability to pay for these. These aspects have been found closely associated with an individual’s position in the socioeconomic ladder, which in turn partially influences his illness perception, choice of service provider and overall household resilience to shocks. This, therefore, makes a strong case for looking into the incidence of illness-induced impoverishment across these characteristics.

Religious and Social Groups

Table 4 shows the incidence of the phenomenon of “medical poverty” across religious and social groups for the two points of time. The 52nd Round of the NSS on Morbidity and Health care did not collect data on religious groups. Also, among social groups, OBCs were absent since they were included in NSS rounds only since 1999–2000 (55th Round). In 2004, Muslims were the most vulnerable to the “medical poverty trap” as evident from a much higher poverty head count after deducting private health payments. The head count ratio increased by more than 8 percentage points among Muslims due to OOP payments for health care, an observation that supplements the findings of the Sachar Committee report.⁷ The poverty gap, however, was found to be higher for Christians implying that though relatively lesser persons in this group were pushed into poverty, their ability to pay for other non-discretionary expenditures was reduced by the highest margin. The lowest increase in the depth of poverty after paying for health care accrued to the Hindus. Among the social groups, scheduled caste (SC) households experienced the highest increase in incidence as well as depth of poverty in both the years. The proportion of individuals moving into poverty because of OOP health payments almost doubled between 1995–1996 and 2004 among the urban SC population.

Table 4.

Illness-induced Impoverishment across Religious and Social Groups

	Poverty Head Count (%)						Poverty Gap (₹)
	1995–1996			2004			1995–1996			2004
	Pre-payment	Post-payment	Difference	Pre-payment	Post-payment	Difference	Pre-payment	Post-payment	Difference	Pre-payment	Post-payment	Difference
Religious Group
Hindu				17.78	23.46	5.68				20	45.7	25.7
Muslim				28.7	37.33	8.63				31.71	64.53	32.82
Christian				10.38	14.65	4.27				8.59	45	36.4
Others				17.03	22.62	5.59				23.5	58.58	35.07
Social Group
SC	39.03	42.79	3.77	30.23	36.91	6.68	31.08	40.21	9.12	36.5	64.62	28.12
ST	45.63	48.64	3.02	30.03	33.15	3.12	42.74	46.3	3.56	40.74	56.42	15.68
OBC				24.41	30.57	6.16				26.32	53.55	27.23
Others	22.95	26.2	3.25	11.01	17.07	6.06	16.38	25.3	8.92	11.95	39.94	27.99
All India	23.52	27.01	3.49	19.35	25.47	6.12	15.31	23.88	8.57	21.73	49.17	27.44

Source: Computed from unit record data of NSS 52nd Round and 60th Round.

Household Characteristics

Though analysis based on religion and caste throws some light on the relatively vulnerable groups and is important from a policy perspective, one must realize that medical impoverishment per se may not be a direct outcome of these attributes. Apart from the nature of illness and health service utilization, there are certain household-level factors that often designate relative vulnerability. It might be possible, therefore, to map the degree of illness-induced impoverishment with selected household characteristics. Two such factors are the gender of the household head and the type of household on the basis of their main occupation. The gender of the household head is often an important factor when dealing with health service utilization mainly because female-headed households are generally found to be economically more vulnerable than their male counterparts. Unfamiliarity with the job market, lack of specific skills and very often downright gender discrimination in payments place female-headed households in a relatively worse-off position leading to economic vulnerability. On the other hand, households with a female head habitually display better health-seeking behavior, more so if the female head is educated and informed. Several studies have proven that empowerment of the mother through education, income generation or simply enhanced status, which is typical of certain societies go a long way in securing better health especially for her children and other family members.

Given the current scheme of health financing in India where treatment cost is substantially paid out of pocket,⁸ the household’s ability to pay is also an important determinant of health-seeking behavior. This vindicates the inclusion of type of household (determined by the source of the household’s income during the 365 days preceding the date of survey) as a variable across which the degree and intensity of “medical poverty” may vary. For example, a regular wage or salaried household might display high health spending simply because of formally assured reimbursement from the employer. On the other hand, a casual labor household might display a distorted demand for medical care because of the lower ability to pay for treatment as well as the probable loss of man-days owing to the ailment or its treatment. Another important categorical variable in analyzing illness-induced impoverishment might be the consumption expenditure quintile (CEQ) to which a household or an individual belongs. The pertinent issue here is to explore whether even the higher quintiles demonstrate cases of medical poverty trap. Therefore, we explore the extent of illness-induced impoverishment across sex of household head, type of household and CEQs (see Table 5).

Table 5.

Illness-induced Impoverishment across Select Household Characteristics

	Poverty Head Count (%)						Poverty Gap (Rs)
	1995–1996			2004			1995–1996			2004
	Pre-Payment	Post-Payment	Difference	Pre-Payment	Post-Payment	Difference	Pre-Payment	Post-Payment	Difference	Pre-Payment	Post-Payment	Difference
Sex of household head
Male	25.85	29.2	3.36	19.23	25.34	6.11	19.03	27.82	8.79	21.44	49.13	27.69
Female	27.88	30.75	2.88	20.61	26.82	6.21	23.03	31.86	8.83	24.74	49.59	24.85
Type of Household
Self-employed	27.18	30.84	3.66	20.12	27.02	6.9	18.76	28.52	9.76	20.63	43.44	22.81
Regular wage/salaried	16.79	19.85	3.05	11.34	16.42	5.08	10.95	18.07	7.12	12.29	37.51	25.21
Casual labour	54.5	57.49	2.99	42.72	48.75	6.03	49.3	58.79	9.5	54.88	85.3	30.42
Others	20.63	24.51	3.88	13.88	21.4	7.52	16.63	30.56	13.92	17.54	85.44	67.9
Consumption Expenditure Quintiles
Poorest	100	100	0	100	100	0	74.78	84.64	9.86	85.13	121.1	35.97
Poor	20.39	27.93	7.54	7.37	19.37	12	4.37	16.51	12.13	2.2	28.47	26.27
Middle	0	3.69	3.69	0	5.24	5.24	0	6.09	6.09	0	22.3	22.3
Rich	0	1.51	1.51	0	2.9	2.9	0	4.6	4.6	0	13.41	13.41
Richest	0	1.1	1.1	0	2.29	2.29	0	10.79	10.79	0	40.52	40.52
All India	23.52	27.01	3.49	19.35	25.47	6.12	15.31	23.88	8.57	21.73	49.17	27.44

Source: Computed from unit record data of NSS 52nd Round and 60th Round.

No significant difference is noticeable between male-headed households (MHH) and female-headed households (FHH) in terms of impoverishment due to health payments. In 1995–1996, male-headed households had a slightly higher difference between pre-payment and post-payment head count, indicating a higher proportion of impoverishment cases as against female-headed households. In 2004, however, the situation was reversed with female-headed households displaying a marginally higher difference in head count on incorporating health care costs. With respect to type of households, the highest incidence of poverty induced by treatment cost was among individuals belonging to “others” type of households. This was true for both points of time. Households belonging to this category are predominantly headed by renters and pensioners. The depth of poverty as given by the difference between pre-payment and post-payment poverty gaps was also high for this group followed by the casual labor households. The highest increase in illness-induced poverty over time was noted for the casual labor households for whom the difference in pre- and post-payment headcount doubled between 1995–1996 and 2004.

The poorest CEQ though economically the most vulnerable did not register a single incident of “medical poverty.” The interpretation of this observation, however, is purely technical. The poverty line was above the cut-off representing the poorest CEQ, and therefore, entry into poverty for this group was out of the question – they were already poor. The poverty gap however increased after payment. In nominal terms, the post-payment poverty gap in 2004 was four times that in 1995–1996. The incidence of poverty was highest for the “poor” CEQ. This was largely because this category predominantly comprised households located just above the poverty line. As such, movements into poverty were more prevalent within this category. In 2004, health care payments impoverished almost 12 percent of the total population in this group. Government anti-poverty programs generally exclude them on the pretext that they are technically above the poverty line. As such, they constitute the most vulnerable group as far as health shocks are concerned. Unwarranted health events in the family are found to drive even richer households into poverty, at least in the short run. This is apparent from the observation that there were cases of illness-induced impoverishment even among persons belonging to the higher expenditure quintiles. However, they might be relatively at ease with this burden largely due to their corpus of savings and as such may not have to adopt drastic coping measures, which are inevitable in the case of poorer households.

The impoverishing impact of OOP health expenses per se, is essentially a “realtime” phenomenon when measured by a household’s current consumption expenditure. The duration of its poverty status is crucially dependent on the sources of financing. In other words, high cost incurred on a particular ailment episode might push a household into poverty real time, but a liberal corpus of saving might come to its rescue and eventually pull it up. Poorer households, without the cushion of financial savings, might resort to borrowing from friends, relatives, formal and informal moneylenders or even distress mortgaging and selling of assets. This is particularly true in the Indian context where financial inclusion is heavily biased against the poor. Thus, the sources of financing OOP health expenses do in a way determine the intensity and duration of the burden.

Conclusion

Illness and its treatment is a potential producer of penury. There is substantial variation in the incidence of the burden of health care expenditure, particularly in urban India, where there is vast heterogeneity in socioeconomic characteristics and the prevailing health system. In this context, the article attempts to explore the degree and distribution of this phenomenon among the urban populace in India. Roughly 6 percent of the urban population or about 18 million individuals were impoverished entirely due to OOP medical expenses in India. Although there were substantial inter-state variations in the incidence of this burden, most states demonstrate an increase in the degree of “medical impoverishment” between 1995–1996 and 2004. The poverty gap before and after health care payment almost doubled for urban India between the two points of time. Urban Muslims were the most vulnerable to the “medical poverty trap”; their head count ratio increased by more than 8 percentage points in 2004 due to OOP payments for health care. Among the social groups, SC households experienced the highest increase in the incidence and depth of poverty in both years. The proportion of urban SC individuals impoverished because of OOP health payments almost doubled between 1995–1996 and 2004. No significant difference is noticeable between MHHs and FHHs in terms of impoverishment due to health payments. Households headed by renters and pensioners had the highest incidence of poverty induced by treatment cost at both points of time. The depth of poverty was also high for this group followed by the casual labor households. The highest increase in illness-induced poverty over time was noted for the casual labor households for whom the difference in headcount before and after payment doubled between 1995–1996 and 2004. With 12 percent of the total population in the group facing impoverishment, the lower middle quintile was easily the most vulnerable lot as far as health shocks are concerned. Unwarranted health events in the family are found to drive even richer households into poverty, at least momentarily.

In India, the state has clearly failed to deliver quality public health services at affordable cost to its citizens and is also reluctant to revamp the system with a judicious mix of financing, regulation, monitoring and implementation. Recent initiatives like introducing cash transfers instead of the public distribution system, etc. indicates the state’s desire to change its historically contemplated role of a provider of public services to that of a facilitator of these services. This has resulted in weak lower-tier public health institutions and, consequently, a huge pressure on specialty hospitals, and institutes of research in urban areas. The urban mass, who cannot afford the long waiting time in public institutions, opt for private for-profit providers. Incidentally, the government provided many private hospitals in urban areas land at a concessional rate on the condition that these hospitals provide a certain proportion of beds free to the poor. However, these institutions have been found to flout these conditions regularly. While public insurance schemes, such as the Rashtriya Swasthya Bima Yojana (RSBY) does acknowledge the phenomenon of health care-induced impoverishment, their coverage is restricted to hospitalization episodes. Although hospitalization entails higher treatment costs, non-hospitalized morbidity is generally the more prevalent form of indisposition and, therefore, potentially more debilitating for a poor household – notwithstanding the lower cost of inpatient treatment. Moreover, while the scheme provides financial protection to the poor in some way, it does not ensure quality of service.

The best way forward would be to direct more financial and human resources into the overtly ailing public health sector. Apart from the cost aspect, this would go a long way in ensuring quality of service, which currently seems to elude the urban poor. A well functioning public health system involving preventive as well as curative health care can also reduce the indirect costs of illness that are largely hidden or are manifested in terms of choice of service provider. The unchecked growth of the commercial private sector must be restrained – if not stopped. Strict observance of standard guidelines for medical and surgical intervention and use of diagnostics and standard fee structure should be made obligatory. In view of the variation in treatment-seeking behavior of the urban populace, sufficient support should be provided to traditional systems of medication too so that they can emerge as a low cost but equally effective alternative to the urban poor. In other words, from a policy perspective, we need to target the reasons for impoverishment rather than the poor per se. This work has argued that morbidity and its treatment thereof is the key event affecting household economic solvency in the short run with potential indebtedness in the longer run for the urban households in India. Therefore, even in the context of counting the poor, this work is an appeal to explicitly incorporate health shocks and their aftermath in the existing poverty lines for an accurate representation of the marginalized sections of society.

Footnotes

Appendix

Table A2.

Share of Health (OOP) in Total Household Consumption Expenditure

	Share of OOP Health Expenditure in household consumption expenditure, 1995–1996 (%)						Share of OOP Health Expenditure in household consumption expenditure, 2004 (%)
	Consumption Expenditure Quintile						Consumption Expenditure Quintile
	States	Poorest	Lower Middle	Middle	Upper Middle	Richest	All	Poorest	Lower Middle	Middle	Upper Middle	Richest	All
Andhra Pradesh	3.2	7.6	4.4	5.2	3.2	4.7	11.0	8.0	6.2	8.4	9.6	8.7
Assam	3.2	4.0	2.3	8.0	5.3	4.5	6.4	19.8	2.6	8.6	2.5	7.6
Bihar	3.7	3.9	7.4	2.7	4.1	4.2	7.4	13.9	7.1	6.1	3.4	8.3
Chhattisgarh							5.3	5.1	7.3	7.5	3.2	6.0
Delhi	5.1	2.5	3.2	2.4	3.0	2.9	1.2	1.2	1.3	1.7	1.1	1.3
Gujarat	6.0	2.9	3.2	2.7	3.6	3.4	11.0	5.4	7.0	4.8	10.7	6.9
Haryana	3.7	4.7	6.5	14.3	10.7	9.1	4.9	8.4	7.0	5.3	9.9	6.8
Himachal Pradesh	1.1	1.9	3.4	4.1	1.5	2.5	6.3	5.3	4.9	5.0	0.9	3.9
Jharkhand							6.6	6.6	6.5	6.4	4.0	6.1
Karnataka	2.9	2.8	2.6	3.8	3.3	3.1	4.0	5.9	4.4	3.5	3.7	4.2
Kerala	5.8	5.5	5.2	5.5	4.1	5.2	19.4	18.4	16.8	10.8	9.6	14.3
Madhya Pradesh	3.3	11.4	4.1	4.7	11.3	6.6	9.8	8.0	6.9	9.4	9.9	8.8
Maharashtra	3.2	4.8	3.7	3.1	3.9	3.7	8.5	7.8	9.4	8.1	7.7	8.2
Orissa	2.5	4.4	2.8	2.2	4.8	3.2	7.0	5.9	5.5	2.9	2.7	4.9
Punjab	6.5	5.6	4.2	4.8	4.8	4.9	22.8	9.5	25.7	7.6	7.8	13.5
Rajasthan	1.5	2.6	2.1	3.6	3.4	2.7	10.0	9.0	6.2	4.8	3.4	6.6
Tamil Nadu	3.0	4.7	2.3	3.0	5.1	3.6	7.1	6.2	8.4	10.4	6.8	7.7
Uttar Pradesh	7.4	6.2	6.3	6.9	7.0	6.8	11.5	9.3	9.1	7.5	11.1	9.9
Uttaranchal							4.8	1.3	6.9	2.9	4.7	3.9
West Bengal	3.1	4.0	5.6	2.8	3.9	3.8	14.2	11.0	8.7	7.1	11.7	10.3
Religious Group
Hindu							9.0	8.1	7.3	6.8	7.3	7.6
Muslim							11.2	9.4	14.4	6.1	7.0	9.9
Christian							11.1	9.2	6.9	9.0	8.9	9.0
Others							23.0	7.3	9.8	9.1	9.7	11.7
Social Group
ST	1.4	2.9	2.9	3.2	5.5	2.8	5.7	4.3	5.1	5.5	2.1	4.8
SC	4.5	5.4	3.3	3.7	5.6	4.4	9.4	6.5	8.5	5.4	9.0	7.9
OBC							10.6	7.6	7.4	6.8	6.1	7.9
Others	3.8	5.5	4.3	4.2	4.4	4.4	10.0	10.6	9.3	7.2	7.9	8.5
Gender of Household Head
Male	3.9	5.4	4.1	3.8	4.6	4.4	10.1	8.1	8.3	6.7	7.3	8.0
Female	3.2	4.6	3.8	7.6	3.4	4.4	8.5	9.6	7.9	8.4	9.0	8.7
Type of Household
Self-employed	4.2	6.0	4.1	4.3	5.6	4.8	9.6	8.9	7.3	7.8	6.6	8.2
Regular/wage/salary	4.0	3.7	3.7	3.4	4.1	3.8	9.3	6.9	9.6	5.7	6.5	7.2
Casual Labour	3.5	6.6	4.7	5.2	5.3	4.8	11.3	7.8	6.5	4.7	3.8	8.8
Others	2.9	7.8	5.4	7.7	4.1	5.4	7.6	12.5	8.3	10.8	12.1	10.6
All India	3.9	5.3	4.1	4.2	4.5	4.4	9.9	8.3	8.3	6.9	7.5	8.1

Source: Computed from unit record data of NSS 52nd Round and 60th Round.

Notes

References

Alkire

Sabina

(2008). Choosing dimensions: The capability approach and multidimensional poverty (MPRA Paper No. 8862). Retrieved October 10, 2014, from http://mpra.ub.uni-muenchen.de/8862/

Baeza

Packard

T. G.

(2005). Beyond survival: Protecting households from the impoverishing effects of health shocks in Latin America. Washington, DC: World Bank.

Banerjee

Deaton

Duflo

(2004). Health care delivery in rural Rajasthan. Economic and Political Weekly, 39(9), 944–949.

Berki

(1986). A look at catastrophic medical expenses and the poor. Health Affairs, 5(4), 138–145.

Berman

Ahuja

Bhandari

(2010). The impoverishing effect of healthcare payments in India: New methodology and findings. Economic & Political Weekly, 45(16), 65–71.

Bonu

Bhushan

Peters

David. (2007). Incidence, intensity, and correlates of catastrophic out-of-pocket health payments in India. ERD Working Paper No. 102, Asian Development Bank.

Chambers

(1983). Rural development: Putting the last first. London: Longman.

Chowdhury

(2011). Financial burden of transient morbidity: A case study of slums in Delhi. Economic and Political Weekly, 46(33), 59–66.

Commission on Macroeconomics and Health (2001). Macroeconomics and health: Investing in health for economic development. Geneva: World Health Organization.

10.

Dercon

Krishnan

(2000). In sickness and in health: Risk-sharing within households in rural Ethiopia. Journal of Political Economy, 108(4), 688–727.

11.

Dev

Jos

(2002). Social sector expenditures in the 1990s: Analysis of central and state budgets. Economic and Political Weekly, 37(9), 853–866.

12.

Garg

C. C.

Karan

A. K.

(2005). Health and millennium development goal 1: Reducing out-of-pocket expenditures to reduce income poverty – evidence from India. EQUITAP Project: Working Paper # 15.

13.

Gertler

Gruber

(2002). Insuring consumption against illness. American Economic Review, 92(1), 51–76.

14.

Ghosh

(2012). Catastrophic payments and impoverishment due to out-of-pocket health spending. Economic and Political Weekly, 46(47), 63–70.

15.

Government of India (2006). Social, economic and educational status of the Muslim community of India–A report. New Delhi: Prime Minister’s High Level Committee, Cabinet Secretariat, Government of India.

16.

Government of India (2007). Conditions of work and promotion of livelihoods in the unorganised sector, report of the national commission for enterprises in the unorganised sector (NCEUS). New Delhi: Government of India.

17.

Gumber

(2002). Structure of the Indian health care market: Implications for health insurance sector. Regional Health Forum, 4(1&2), 26–34.

18.

Guruswamy

Mohan

Abraham

Ronald J.

(2006). Redefining the poverty line in India. Public Policy Research, 13(3), 135–207.

19.

Kawabata

(2002). Preventing impoverishment through protection against catastrophic expenditure. Bulletin of the World Health Organization, 80(8), 612.

20.

Krishna

(2004). Escaping poverty and becoming poor: Who gains, who loses, and why? World Development, 32(1), 121–136.

21.

Krishnan

T. N.

(2000). Access to health care and burden of treatment in India: An interstate comparison. In Rao

(Ed.), Investing in health (pp. 212–230). New Delhi: SAGE Publications.

22.

Limwattananon

Tangcharoensathien

Prakongsai

(2007). Catastrophic and poverty impacts of health payments: Results from national household surveys in Thailand. Bulletin of the World Health Organization, 85(8), 600–606.

23.

Lindelow

Wagstaff

(2005). Health shocks in China: Are the poor and uninsured less protected? World Bank Policy Research Working Paper No. 3740. Washington, DC: The World Bank.

24.

Meesen

Zang

(2003). Iatrogenic poverty. Tropical Medicine and International Health, 8(7), 581–584.

25.

National Health Accounts India (2004–05). NHA Cell. Ministry of Health and Family Welfare: Government of India.

26.

National Sample Survey Organisation (NSSO) (1992). Morbidity and utilisation of medical services, NSS 42nd round July 1986–June 1987. Sarvekhshana, 15(4), 55–75.

27.

National Sample Survey Organization (NSSO) (1998). Morbidity and treatment of ailments. Report No. 441. New Delhi: Department of Statistics, CSO, GOI.

28.

O’Donnell

van Doorslaer

Wagstaff

Lindelow

(2008). Analyzing health equity using household survey data: A guide to techniques and their implementation. Washington, DC: World Bank.

29.

OECD & WHO (2003). DAC guidelines and reference series – Poverty and health. Paris: OECD Publishing.

30.

Patnaik

Utsa

(2006). Poverty and neo-liberalism in India. Retrieved October 10, 2014, from http://www.ideaswebsite.org/pdfs/neo-liberalism(2).pdf

31.

Patnaik

Utsa

(2007). Neoliberalism and rural poverty in India. Economic and Political Weekly, 42(30), 3132–3150.

32.

Pradhan

Prescott

(2002). Social risk management options for medical care in Indonesia. Health Economics, 11(5), 431–446.

33.

Rajaratnam

Abel

Duraisamy

John

K. R.

(1996). Morbidity pattern, health care utilization and per capita health expenditure in a rural population of Tamil Nadu. National Medical Journal of India, 9(6), 259–262.

34.

Ranson

M. K.

(2002). Reduction of catastrophic health care expenditures by a community based health insurance scheme in Gujarat, India: Current experiences and challenges. Bulletin of the World Health Organization, 80(8), 613–621.

35.

Russell

(2004). The economic burden of illness for households in developing countries: A review of studies focusing on malaria, tuberculosis, and human immunodeficiency virus/acquired immunodeficiency syndrome. American Journal of Tropical Medicine and Hygiene, 71(Supp. 2), 147–155.

36.

Sen

Pronab

(2005). Of calories and things. Economic and Political Weekly, XL(43), 4611–4618.

37.

Shahrawat

Renu

Rao

Krishna D.

(2012). Insured yet vulnerable: Out-of-pocket payments and India’s poor. Health Policy and Planning, 27(3), 213–221.

38.

van Doorslaer

O’Donnell

Rannan-Eliya

R. P.

Somanathan

Adhikari

S. R.

Akkazieva

, (2005). Paying out-of-pocket for health care in Asia: Catastrophic and poverty impact. EQUITAP Working Paper 2, May 2005. Erasmus University, Rotterdam. Retrieved from http://www.equitap.org

39.

van Doorslaer

O’Donnell

Rannan-Eliya

(2007). Catastrophic expenditures on health care in Asia. Health Economics, 16(11), 1159–1184.

40.

Visaria

Gumber

(1994). Utilisation of and expenditure on health care in India. Ahmedabad: Gujarat Institute for Development Research.

41.

Wagstaff

(2005). The Economic Consequences of Health Shocks. World Bank Policy Research Working Paper WPS3644. Washington, DC: World Bank.

42.

Wagstaff

van Doorslaer

(2001). Paying for health care: Quantifying fairness, catastrophe and impoverishment with applications to Vietnam 1993–98. World Bank Policy Research Working Paper No. 2715. Washington DC, World Bank.

43.

Wagstaff

van Doorslaer

(2003). Catastrophe and impoverishment in paying for health care: With application to Vietnam 1993–98. Health Economics, 12(11), 921–933.

44.

Whitehead

Dahlgreen

Evans

(2001). Equity and health sector reforms: Can low income countries escape the medical poverty trap? Lancet, 358(9284), 833–836.

45.

World Bank (2008). Quantitative techniques for health equity analysis – Technical note # 19.

46.

Wyszewianski

(1986). Financially catastrophic and high cost cases: Definitions, distinctions and their implications for policy formulation. Inquiry, 23(4), 382–394.

47.

Evans

D. B.

Carrin

Aguilar-Rivera

A. M.

Musgrove

Evans

(2007). Protecting households from catastrophic health spending. Health Affairs, 26(4), 972–983.

48.

Evans

D. B.

Kawabata

Zeramdini

Klavus

Murray

C. J. L.

(2003). Household catastrophic health expenditure: A multicountry analysis. The Lancet, 362, July 12.