Accounting for horizontal inequity in the delivery of health care: A framework for measurement and decomposition

https://doi.org/10.1016/j.iref.2019.10.010Get rights and content

Abstract

The pursuit of equity is an objective of many healthcare systems. Horizontal equity, interpreted as “equal treatment for equal need”, has received much attention in both the policy and academia arenas. By combining the indirect standardization method with regression-based Shapley value decomposition, the paper aims to propose a framework for measuring and decomposing horizontal inequity and to investigate the contributors to horizontal inequity in health care delivery in China using the China Health and Nutrition Survey (CHNS) dataset. The horizontal inequity indicated by the Gini coefficient of indirectly standardized healthcare expenditure accounts for approximately 68.63 percent of the overall inequality, and the non-need factors, such as household registration, region, work status, education, income, insurance, and marital status, explain between 50 and 70 percent of the inequity, with household registration and region being the two largest contributors.

Introduction

Achieving equity and equality in health care delivery is a widely pursued but seldom accomplished policy objective in many countries. There are many theoretical and empirical studies on the inequity in health and health care (Wagstaff & van Doorslaer, 2000a; Fleurbaey & Schokkaert, 2009; van de Poel, van Doorslaer, & O’Donnell, 2012; Terraneo, 2015). Among the various dimensions of healthcare inequity, horizontal equity receives much attention in the literature. Horizontal equity means “equal treatment for equal need (hereafter ETEN)” and is referred to as “unfair inequality” in Fleurbaey and Schokkaert (2009). This indicates that individuals with the same healthcare need should receive the same amount of resources, irrespective of other socioeconomic factors, such as education, household registration and area of residence (Wagstaff, van Doorslaer, & Paci, 1991). While health inequalities attributable to biological variations or free choice are unavoidable, others due to the uneven distribution of social determinants of health are avoidable. Thus, investigating horizontal inequality in health care has significant policy implications.

China, with a population of 1.4 billion, is one of the largest developing countries, and its healthcare system determines the health welfare of approximately one-fifth of the world’s population. Its healthcare system is undergoing a major reform, one of the most complex and far-reaching efforts ever undertaken by any public health system in the world. For example, China decentralized the fiscal system in the mid-1980s to rectify the inefficiencies of the centralized command system. The decision making of health care spending was also decentralized to provinces and local governments. A decentralized health care system might increase efficiency in terms of expenditure and investment. However, the disarray in decentralization diminishes the government’s role in managing public health programs and aggravates inequality in the accessibility and delivery of healthcare provision. Under this decentralized allocation of decision-making power, how to maintain equity in the health care delivery is a paramount issue for the 1.4 billion population and thus deserves a thorough scrutiny.

Given the dominant importance of horizontal inequity in the literature, coupled with the unprecedented healthcare system reform in China, this paper attempts to achieve two objectives: first, proposing a framework for measuring and decomposing horizontal inequity and second, exploring the possible sources of horizontal inequity in healthcare delivery in China.

The contributions of the paper are four folds: 1) The method we propose to measure and decompose horizontal inequity is closely related to the concept of egalitarian-equivalence in the literature on fair allocation because they are both inspired by the ideal situation in which all individuals have the same circumstances (Fleurbaey & Schookaert, 2012; Fleurbaey & Schokkaert, 2009). 2) The method satisfies path independence (Fortin, Lemieux, & Firpo, 2011) and would not be contingent on the model specification. 3) We identify the difference between horizontal inequity and overall inequality with raw data. As shown in Table 6, the horizontal inequity is much smaller than the overall inequality and has an obviously different theoretical foundation. 4) While most papers study the horizontal inequity for developed countries, there is a scarcity of studies on the topic for China. This paper fills the void by contributing some evidence for the largest developing country.

We obtain the following findings. First, contrary to the common belief, we find that the horizontal inequity of indirectly standardized healthcare expenditure (ISHE inequality) from 1991 to 2011 mainly results from the non-need factors, such as household registration, region, work status, education, insurance, and marital status, rather than from household income per capita. To be more specific, household registration is the most significant contributor, which accounts for approximately 20 percent of the total horizontal inequity. Region is the second largest factor, contributing 11.24 percent to 18.32 percent. Table 4 lists the contributions of all the variables. It is notable that household income per capita only ranks the sixth largest contributor in most years, ranging between 3 and 5 percent. Second, we find that the horizontal inequity indicated by the Gini coefficient of indirect standardized healthcare expenditure accounts for approximately 68.63 percent of the overall inequality.1 This indicates that the overall inequality with raw data in health care delivery does not fully reflect the inequity in reality.

The remainder of this paper is structured as follows. Section 2 introduces the theory of horizontal equity. Section 3 reviews the previous literature. Section 4 elaborates the new framework for measuring and decomposing horizontal inequity. Section 5 investigates the sources of inequity in China’s healthcare system. Section 6 carries out the empirical analyses of health care delivery in China. Section 7 concludes.

Section snippets

Horizontal (in)equity and ETEN

There are many existing theories and specifications of horizontal equity (Culyer, Van Doorslaer, & Wagstaff, 1992; Le Grand, 1991; Mooney, Hall, Donaldson, & Gerard, 1991; Wagstaff et al., 1991). Mooney et al. (1991) define horizontal equity according to the idea that individuals with equal need should enjoy the same access to health care. However, even with the same access to the health care service, individuals usually end up consuming different amounts due to different demand curves. If

Previous empirical research

For decades, the existing literature has focused mainly on developed countries, such as the UK (Morris, Sutton, & Gravelle, 2005), the US (van Doorslaer et al., 2000) and the Netherlands (Wagstaff & van Doorslaer, 2000a) and international comparisons among developed countries (Lu et al., 2007; Terraneo, 2015; Wagstaff et al., 1991). This may be because many healthcare systems in developed countries are based on the principle of horizontal equity (Kelley & Hurst, 2006; Terraneo, 2015). There is

A new framework of measuring and decomposing horizontal inequity

As described in Section 2, horizontal equity is interpreted as the principle of equal treatment for equal need irrespective of social economic status. Horizontal inequity is the “unequal treatment of the equal need” (Jenkins, 1988). Obviously, it is essential to distinguish between “need” factors and “non-need” factors (Culyer & Wagstaff, 1993; Culyer, 1995; O’Donnell, van Doorslaer, Wagstaff, & Lindelow, 2008). Morris et al. (2005) state that need factors are those that ought to affect the

Inequity in China’s healthcare system

Since market-oriented reforms began in 1978, China has implemented a strategy of promoting unbalanced development in economic and social sectors, which has led to substantial inequalities across regions, between urban and rural areas and between coastal and inland districts (Qin & Hsieh, 2014; Zhang, Du, Zhuge, & Tong, 2019). The market-oriented reforms have also changed the traditional healthcare system established for a low-level economy based on an equalitarian health policy.

Notably,

Data and variables

To investigate the determinants of healthcare expenditure and its inequity in China, we employ repeated cross sections from the China Health and Nutrition Survey5 (CHNS), which is a longitudinal dataset conducted in nine waves (1991, 1993, 1997, 2000, 2004, 2006, 2009 and 2011). The CHNS includes information such as demographics, socioeconomic status, health and nutrition, so we can examine the effects of Chinese economic and social transformation

Conclusion

In this paper, a new framework is proposed combining indirect standardization mothed with regression-based Shapley value decomposition to investigate horizontal inequity in health care delivery in China using the China Health and Nutrition Survey (CHNS) dataset. On the one hand, the indirect standardization could control for the differences of need factors, such as severity of illness, age, and sex, and identify the difference between horizontal inequity and overall inequality with raw data. On

Authors’ contributions

Zhao set the framework and did the main empirical work; Cao did parts of the empirical work and wrote the main parts of paper; Ma cleaned the data and wrote parts of the paper. All authors read and approved the final manuscript.

Funding

This work was supported by the Philosophy and Social Science Foundation of Jiangsu Colleges and Universities (grant number 2019SJA0262); the Philosophy and Social Science Foundation of China (grant number 16AGL014); and the National Natural Science Foundation of China (grant number 71603046).

Declaration of competing interest

None.

Acknowledgements

I am grateful to the editor and anonymous referees for useful comments. The remaining errors are mine.

References (39)

Cited by (10)

View all citing articles on Scopus
View full text