ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846
-A A +A

Measuring Catastrophic Healthcare Expenditure

Sunil Rajpal ( is a doctoral scholar at the Indian Council of Social Science Research, Delhi. William Joe ( teaches at the Institute of Economic Growth, Delhi.

Catastrophic household healthcare expenditure is a prominent policy concern. The National Health Policy 2017 takes explicit cognisance of this issue and presents an empirical formulation to examine its incidence and patterns. However, the policy needs to account for household size variations to counter an implicit bias that tilts the estimates to reflect a higher concentration of catastrophic expenditure among the rich. This concern is illustrated using health data from the 71st round of the National Sample Survey. Further, a minor modification to unravel the socio-economic gradient in catastrophic healthcare expenditure has also been discussed.

The authors are thankful to an anonymous reviewer for helpful comments and suggestions. 

Affordability is a key principle under the National Health Policy 2017 (NHP) of India. The NHP states that,

As costs of care increases, affordability, as distinct from equity, requires emphasis. Catastrophic household healthcare expenditures defined as health expenditure exceeding 10% of its total monthly consumption expenditure or 40% of its monthly non-food consumption expenditure, are unacceptable. (GoI 2017: 2)

The NHP deserves appreciation for upholding the concern of financial protection in health, but the approach towards measurement of catastrophic household healthcare expenditure delimits the normative rigour of this principle. In general, health expenditure is considered as catastrophic when it endangers the family’s ability to maintain its customary standard of living (Berki 1986). The NHP definition, however, is related to two alternative empirical approaches presented in Xu et al (2003) and Wagstaff and Doorslaer (2003). These methods were developed to facilitate measurement of catastrophic health expenditure across countries with significant data gaps on household income and expenditure. But the prescribed approach lacks coherence due to its total neglect of household size or composition in the estimation process. The problem is akin to the use of total gross domestic product (GDP) versus per capita GDP in ranking of countries or economies, which can provide contrasting conclusions. For instance, currently, India ranks third in terms of GDP (in purchasing power parity [PPP]) across countries, whereas it ranks 123rd in terms of per capita GDP (in PPP). This elementary concern of neglecting the per capita dimension assumes salience for policymaking on catastrophic healthcare payments in India. In particular, such neglect can yield misleading conclusions regarding the socio-economic gradient in the incidence of catastrophic payments.

These apprehensions are elaborated using a hypothetical illustration in Table 1. Household A has a lower total consumption expenditure (₹8,000) than household B (₹10,000), but, due to smaller household size, the per capita consumption levels are higher for the former household. Suppose both households spend ₹1,000 on healthcare services. In relative terms, this implies that household A spends 12.5% of its total consumption expenditure on healthcare, whereas household B spends only 10%. Thus, based on the NHP approach, household A is experiencing catastrophic health expenditure because it exceeds the 10% threshold of total consumption expenditure. Similarly, the healthcare expenditure of household A accounts for 50% of the total non-food expenditure (₹2,000). Using the NHP definition, household A is again categorised as experiencing catastrophic health expenditure because its health expenditure accounts for more than a 40% share in non-food expenditure. Importantly, despite allocating a greater share of per capita consumption and per capita non-food expenditure towards healthcare, either approach fails to consider household B to be incurring catastrophic payments.

Clearly, the NHP definition has an implicit pro-rich bias whereby similar-level health expenditure among households with higher per capita expenditure can appear as catastrophic, whereas it gets neglected for a household with lower per capita expenditure. Determination of catastrophic health expenditure without considering the dimension of household size can, therefore, offer counter-intuitive inferences.

Modified Approach

Overlooking such elementary concerns can have non-trivial implications for the narrative of magnitude and socio-economic patterns in the distribution of catastrophic healthcare expenditure in India. Therefore, it is necessary to account for variations in household size while estimating the incidence of catastrophic healthcare expenditure. In particular, we present a modified approach that retains the basic NHP formulation, but replaces the denominator with per capita annual household consumption expenditure.

Formally, let THE denote total annual household health expenditure. Let HCE denote total annual household consumption expenditure, and PHCE denote per capita household annual consumption expenditure. Also, let Ca denote the threshold used for identifying catastrophic expenditure where threshold values can be set at any consensual level such as 10% or 20%. Under the NHP approach, a household can be classified as incurring catastrophic expenditure if THE/HCE > Ca(where Ca= 0.1 or 0.2). However, we recommend a minor but important modification whereby health expenditure can be considered catastrophic if THE/PHCE > Ca(where Ca= 1 or 2). Consequently, the interpretation is also modified and becomes more intuitive. The modified interpretation will estimate the percentage of households whose healthcare expenditure is more than the annual consumption expenditure of an average household member. It may be noted that varying levels of thresholds can be used to understand the patterns and intensity of such payments. For instance, using Ca= 0.5 will provide the proportion of households whose healthcare spending is more than the half-yearly consumption expenditure of one average member, whereas a threshold of Ca= 2 will reveal expenditure greater than the annual consumption expenditure for two years. Moreover, it is a straightforward process to extend this approach in the context of food-based expenditure.

Inpatient Care Expenditure

For an empirical illustration of the NHP approach and the modified approach, we use the data from the 71st round of the National Sample Survey (NSS 2015) on social consumption: health, and estimate the incidence of catastrophic healthcare expenditure in India while accessing inpatient care services during the survey reference period of the past one year. This survey was conducted from January to June 2014 with a total sample of 65,932 households (36,480 rural, 29,452 urban). For comparison, we present the estimates of catastrophic expenditure by using two different thresholds. Concentration index is computed to discern the association of catastrophic expenditure with socio-economic status (Erreygers 2009). A negative concentration index value would imply greater concentration of catastrophic payments among the poorer households, while a positive concentration index value implies the same for richer households.

Table 2 shows that the average annual household inpatient care expenditure among households with hospitalisation cases is ₹24,300. The average inpatient care expenditure among poorest households is ₹15,100 whereas it is three times higher (₹44,500) amongst the richest households. A significant differential in average inpatient care expenditure is observed across social groups and occupational groups. The hospitalisation expenditure among backward social groups—particularly, Scheduled Castes (SCs) and Scheduled Tribes (STs)—is much lower than others. Similarly, hospitalisation expenditure among casual labour households is much lower than other occupations. Significant rural–urban differentials are also apparent. Further, among poorest households with hospitalisation cases, the average annual household consumption expenditure is ₹53,500 whereas the same is about 3.5 times higher among the richest households. But, the average annual per capita expenditure among the richest households is 5.6 times higher than the poorest ones. The stark contrast and relative gap thus becomes apparent when adjusted for household size.

Table 3 (p 24) presents the incidence of catastrophic inpatient care expenditure. Based on the 10% threshold specified under the NHP approach, it is estimated that every second household seeking hospitalisation care incurs catastrophic health expenditure. The NHP approach, however, overlooks the income-related gradient in the incidence of catastrophic expenditure. The concentration index value [-0.01] for the NHP approach confirms the absence of a socio-economic gradient. Clearly, it is puzzling that there is hardly any variation in the incidence of catastrophic expenditure across socioeconomic categories. Although, Table 2 clearly indicates that households with higher average consumption expenditure tend to have higher hospitalisation expenditure, the absence of an income-related gradient emerges as a key methodological concern. Similarly, the incidence of catastrophic expenditure lacks systematic variations across occupation groups. Incidence is higher in rural areas (51.7%) than urban areas (46.1%), but the concentration index does not reveal any significant socio-economic gradient in either of the areas.

The modified approach, however, overcomes these limitations to present estimates that are consistent with the conceptual underpinning of catastrophic health expenditure. Unlike the NHP approach, the modified approach uses more intuitive catastrophic thresholds to estimate the percentage of households whose healthcare expenditure is more than the annual consumption of expenditure of one member or two members. Table 3 presents the incidence of catastrophic healthcare expenditure based on the modified approach.

Results based on the one-member threshold reveals that while seeking hospitalisation care, 29% households had to incur expenditure greater than the yearly consumption expenditure of a household member. Similarly, 14% households had to incur expenditure greater than the combined annual consumption expenditure of two members of the household. The modified approach reveals the systematic association between household well-being and the incidence of catastrophic expenditure. Compared to richer households, the poorer ones are more likely to experience catastrophic consequences. This is further confirmed by the significantly negative concentration index values. With the increase in catastrophic thresholds, the concentration of incidence of catastrophic payments also increases among the poor. Further, the method performs well to reveal greater incidence of catastrophic payments among casual labour households than other occupations.

Concluding Remarks

Early conceptualisations of catastrophic expenditure adopted an absolute approach such that health expenditure greater than a particular fixed amount was regarded as catastrophic (Wyszewianski 1986). A particular level of health expenditure can, however, account for a higher share of income among poorer households and a lower share among richer households. A relative scale based on percentage share in the total income was, therefore, preferred to estimate the incidence of catastrophic expenditure. The relative approach acknowledges that healthcare payments should be based on the ability-to-pay principle. The common approaches to measure catastrophic payments aim to capture this fundamental concern (Wagstaff and Doorslaer 2003; Xu et al 2003). However, while applying these methods on consumption expenditure data, we need to account for household size variations, else it can provide counter-intuitive inferences. In fact, previous studies based on an NHP-type approach are puzzled to observe the lack of a socio-economic gradient in catastrophic health payments (Ghosh 2011; Van Doorslaer et al 2007; Khan et al 2017). It is no surprise that household size is higher among low-income households and it shrinks the per capita availability of household resources. After accounting for such variations, the modified approach confirms a greater incidence of catastrophic health payments among the poorer households.

In conclusion, it may be reiterated that the NHP definition for measurement of catastrophic expenditure should explicitly account for household size variations to reveal the socio-economic gradient in the incidence of catastrophic expenditure. This definitional anomaly also presents a unique opportunity to develop an alternative policy definition for the measurement of catastrophic household healthcare expenditure in India. In particular, policies on financial protection in health can emphasise the understanding of the various sources of financing household healthcare expenditure in India. For instance, it is disconcerting that in 2004, about 60% of rural households and 40% of urban households had to rely on borrowings and sale of physical assets to finance inpatient care services (Joe 2015). Such information on the magnitude of distressed financing can be used to complement assessments of catastrophic healthcare expenditure in India and other similar data constrained settings. Similarly, as indicated by NSS data, several households with poor ability to pay are likely to delay or forego treatment. These households should be identified for policy analysis and their proportions should be viewed in conjunction with the incidence of catastrophic expenditure and incidence of distress financing, in order to outline these as central concerns in the political economy of healthcare in India.


Berki, S E (1986): “A Look at Catastrophic Medical Expenses and the Poor,” Health Affairs, Vol 5, No 4, pp 138–45.

Erreygers Guido (2009): “Correcting the Concentration Index,” Journal of Health Economics, Vol 28, No 2, pp 504–15.

Ghosh, Soumitra (2011): “Catastrophic Payments and Impoverishment due to Out-of-pocket Health Spending,” Economic & Political Weekly, Vol 46, No 47, pp 63–70.

GoI (2017): “National Health Policy, 2017,” Ministry of Health and Family Welfare, Government of India.

Joe, William (2015): “Distressed Financing of Household Out-of-pocket Health Care Payments in India: Incidence and Correlates,” Health Policy and Planning, Vol 30, No 6, pp 728–41.

Khan, Jahangir A, Sayem Ahmed and Timothy G Evans (2017): “Catastrophic Healthcare Expenditure and Poverty Related to Out-of-pocket Payments for Healthcare in Bangladesh: An Estimation of Financial Risk Protection of Universal Health Coverage,” Health Policy and Planning, Vol 32, No 8, pp 1102–10.

NSS (2015): “Key Indicators of Social Consumption in India: Health,” National Sample Survey 71st Round, Ministry of Statistics and Programme Implementation, National Sample Survey Office.

Van Doorslaer, Eddy et al (2007): “Catastrophic Payments for Health Care in Asia,” Health Economics, Vol 16, No 11, pp 1159–84.

Wagstaff, Adam and Eddy van Doorslaer (2003): “Catastrophe and Impoverishment in Paying for Health Care: With Applications to Vietnam 1993–98,” Health Economics, Vol 12, No 11, pp 921–33.

Wyszewianski, Leon (1986): “Financially Catastrophic and High-cost Cases: Definitions, Distinctions, and Their Implications for Policy Formulation, Inquiry, Vol 23, No 4, pp 382–94.

Xu, Ke, Jan Klavus, Kei Kawabata, David B Evans, Piya Hanvoravongchai, Juan Pablo Ortiz, Riadh Zeramdini and Christopher J L Murray (2003): “Household Health System Contributions and Capacity to Pay: Definitional, Empirical and Technical Challenges,” Health Systems Performance Assessment: Debates, Methods and Empiricism, Christopher J L Murray and David B Evans (eds), World Health Organization, Geneva.

Updated On : 14th Aug, 2018


(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top