Understanding equity of institutional delivery in public health centre in India: An assessment using Benet Incidence Analysis

Background: The National Health Mission (NHM), the largest ever publicly funded health programme worldwide used over half of the national health budget in India and primarily targeted to improve maternal and child health in the country. Though increased public health investment in India has improved the health care utilization and health outcomes across geographies and socio-economic groups, little is known on equity aspect of such large scale investment. In this context, this paper examines the utilisation pattern of delivery care and equity aspect of public investment in India. Methods: Data from the recent round of National Family Health Survey (NFHS 4) conducted during 2015-16 used in the analyses. A total of 148,645 last birth delivered in an institution in a reference period of ve years has been used in the analyses. The out-of-pocket (OOP) payment on delivery care is the dependent variable. Benets Incidence Analysis (BIA), descriptive statistics, concentration index (CI), and concentration curve (CC) used in the analysis. Results: Utilization of natal care from public health centres in India is pro-poor and have a strong economic gradient. However, about 28% mothers in richest wealth quintile did not pay for delivery in public health centres compared to 16% among the poorest wealth quintile. The median OOP payments for institutional delivery from public health centre was ₹1200 for mothers belonging to the poorest wealth quintile compared to ₹2000 for mothers belonging to richest wealth quintile. Benet incidence analyses suggests that about 17.4 % of subsidies were used among poorest, 19.3% among richest quintile and 21.9% in the middle quintile. The concentration index of institutional delivery from public health centres was -0.161 [95% Condence Interval: -0.158, -0.165] compared to 0.296 [95% Condence Interval: 0.289, 0.303] from private health centres. Conclusion: Utilisation of natal care services from public health centres in India is pro-poor. However, benet-incidence analyses suggest that the use of subsidy is relatively higher among wealthier and better off mothers in India. from ₹7,000 to ₹14,600. The pattern is almost similar for the four selected states with some exception. Uttar Pradesh and Karnataka show a similar pattern of increasing OOP for institutional delivery with the increase in economic status. However, in case of Maharashtra and Rajasthan, some interesting patterns have been found. Overall, poorest women of Maharashtra spent ₹7,937 regardless their place of delivery. Furthermore, women belonging to poorest section and delivered at public health centre spent almost double than that of richest section. Similarly, for Rajasthan, richer section spent the lowest of ₹1,685 than the women belonging to other quintile in case of public health centre delivery. high performing states. Institutional delivery in Low Performing States (LPS) varied from 12.9% among women from richest wealth quintile to 23.5% among women in poorest quintile whereas it varied from 12.5% among women from richest quintile to 25.7% among woman from poorest quintile in High Performing States (HPS). Considering the median OOP of service, the share of public subsidy was maximum for the women from middle quintile in both LPS and HPS. However, the share was minimum for the poorest quintile (18.7%) in LPS and richest quintile in HPS (17.7%). When mean OOP of delivery care was taken, the public subsidy was maximum for the women from the poorer wealth quintile (24.3%) in LPS and middle quintile in HPS (22.7%) (Appendix 2). Also, the benet incidence of the public subsidy was computed for women with education less than ve years and more than ve years. The utilization rate of public health centre among women with less than ve years of education varied from 21.9% among women from richest quintile to 16.7% among women from richest quintile while it varied to 25.9% among women from poorest quintile 12.1% among women from richest quintile for women with more than ve years of education. Considering the median OOP of delivery care service, the benet incidence of public subsidy was highest among women’s from the richer quintile (22.1%) having less than ve years of education In comparison the share was highest among women from middle wealth quintile (21.4%) having more than ve years of education.


Introduction
Increasing health spending and rising health inequality is concomitant across geographies and socio-economic groups [1][2][3][4]. Rising health spending is associated with increasing public investment on health and declining out-of-pocket (OOP) payments on health spending [5][6]. Despite these, the catastrophic health spending (CHS) and impoverishment, resulting from OOP payment have been increasing in many developing countries [7][8][9]. The CHS, impoverishment and the poverty impact of health spending vary across countries and largely depend on income level, public policies, coverage of health insurance schemes, type of provider, payment method, disease burden and demographics [10][11][12]. Globally about 1.3 billion people do not have access to effective and affordable health care. Of those who have access about 170 million are forced to spend more than 40% of their household income on medical treatment. Over 100 million people annually are pushed into extreme poverty due to health spending [13].
Equity and e ciency are the two pillars of public health investment worldwide. Goal 4 and 5 of Millennium Development Goals (MDGs) and goal 3 and 10 of Sustainable Development Goals (SDGs) has outline speci c goals to reduce inequality in access and utilisation to quality health services [14][15]. Most of the welfare government has made large-scale investments to increase the access and utilisation of health care services. The periodic evaluation suggests mixed impact of public health investment on health outcome and health care utilisation [16][17][18][19]. Studies used varying approaches in understanding the impact of public health investment (bene t-incidence analysis, individual preference, concentration curve, concentration index). Among these, bene t incidence analyses (BIA) has been increasingly used in health economics literature [20-24, 44-45. 48]. Bene t incidence analysis is a tool which shows whether the subsidies are directed to the poorer section or it helps the better-off section of the society.
It also involves the estimates of monetary value of the services and its distribution along the population [25]. Effectiveness of the distribution of limited resources to meet the needs of the poor is captured along with extent of e cient resource allocation [21].
Studies from African and Asian countries suggest that the impact of public subsidy on health care is pro-poor for some countries and pro-rich for others. A systematic review suggests that in most of the lower-middle income countries (LMICs), a pro-rich distribution of public subsidy was observed for health care service except for Nigeria [20]. In Nigeria, individuals belonging to the rural areas and poorest wealth quintile had the highest share of bene ts received for utilisation of public health care services [26]. In Kenya, the distribution of public subsidy at primary health centres was pro-poor while that at hospital level were pro-rich [27]. The distribution of health services was pro-poor in Hong Kong and Malaysia and it was pro rich in Nepal, Bangladesh and India [24]. In India, the share of public subsidy for in-patient and out-patient services was 40% in the richest quintile compared to 9% in the poorest quintile [28].
Over a decade ago, the state of maternal and child health was poor in the country. In 2002-03, the maternal mortality ratio was 286 per 100,000 live births and the under-ve mortality was 74 per 1,000 live births [29][30]. Over half of the mothers did not delivered at a health centre. The institutional delivery among women belonging to the poorest wealth quintile was 12.8% compared to 83.6% in richest quintile in 2005 [30]. The inequality in health care utilisation was large in many other health services [4,[31][32][33]. As a policy response, the Government of India in 2005 revamped the health programme in the country and launched the National Health Mission (NHM), the largest ever health program worldwide. The main objective of NHM was improvement of maternal and child health care in poorer regions of the country and among the poor and vulnerable section of the population. The NHM had an estimated annual budget of over ₹26,691 cr. in 2017-18 accounting more than half of health budget of the union government [34]. The large scale public health investments have reduced the maternal and child mortality in the country. Delivery from public health centres has increased from 18% in 2005-06 to 52.1% by 2015-16 [30,35]. Despite the increase in coverage of services, the out-of-pocket (OOP) expenditure and catastrophic health spending (CHS) on delivery care remained high [36][37][38]. Studies suggest that inequality in health care services has widened across state, rural and urban area and among wealth quintile [39][40].
A number of studies in India have used BIA approach to examine the bene ts of public subsidy on inpatient care, out-patient care and delivery care. The distribution of public subsidies in Karnataka was six times higher for richest 20% of the population compared to poorest 20% [41]. In Northeast India, the bene ts for inpatient care in urban area were pro-poor whereas for rural areas, the bene ts were pro-rich in nature [42]. A recent study found a pro-rich distribution of public subsidy for inpatient care of non-communicable diseases (NCDs) among elderly [43]. In West Bengal, the bene t of public subsidy was highest for individuals belonging to lower-middle group in rural areas and upper-middle income group in urban areas [44]. In 2004, for inpatient care, the share of public subsidy was maximum for poorest section (41.5%) in Tamil Nadu, lower middle in West Bengal (32.3%) and richest section (38.9%) in Rajasthan. In 2014, the share of public subsidy was highest for poorest section of the society across all three states. Across rural areas, the share of public subsidy in Tamil Nadu and Rajasthan shifted from lower middle section to poorest section of population during 2004-14 whereas in West Bengal, the richer section continued to avail the subsidy during same period [45]. In India, though the utilisation of delivery care services was higher among the poor, the bene ts of public subsidy for delivery care services were pro-rich in nature [21].
In developing countries, public investment in health remained low over time and effectiveness of public spending on health care services is an elusive empirical issue. Increasing public health expenditure on health care services does not automatically bene t to all groups of population if the distribution of resources is not equitable [46]. While the average utilisation of services may increase, it may not necessarily bene t the poor and marginalised [47]. Therefore, it is important to assess empirically whether public spending in India truly bene ts the poor section of the population. The national average in utilisation of delivery care services from public health centres conceals large variations across states and economic groups. Though there has been increase in utilisation of maternal services from public health centres, little is known to whom it is bene ting. It is unclear whether the bene ts are largely pro-poor or pro rich. In this context, this paper examines the equity in the distribution of public subsidy among the mothers using public health centres for institutional delivery using the BIA approach.

Data And Methods
Unit data from the recent round of National Family Health Survey (NFHS-4) conducted during 2015-16 has been used for the analysis. NFHS 4 is the fourth in the series of Demographic Health Survey (DHS) in India that aimed to provide reliable estimates of utilization of maternal and child health services, contraception, nutrition etc. along with the socio-economic and demographic condition of the households. A total of 601,509 households, 699,686 ever-married women in the age group 15-49 and 112,122 men in the age group 15-54 were successfully interviewed across all states and union territories of India. The NFHS-4 for the rst time included a set of policy relevant question on OOP payment on delivery care (de ned as the expenditure net of reimbursement) for the last birth delivered at a health centre and reimbursement under Janani Suraksha Yojana (JSY). Findings from the survey along with sampling design, methodology, and results are available in the national report [35].
Unit data from the kids le has been used, which provides details of births to mothers during ve years preceding the survey. A total of 259,627 births were reported of which 190,898 were last births and 148,645 were conducted in the health centres (institutional delivery). The unit data was cleaned for factual errors on OOP payment before the analysis. The details and procedure for data cleaning are available elsewhere [36].

Statistical Analysis
Descriptive statistics, Bene t Incidence Analysis (BIA), and Concentration Index (CI), and Concentration Curve (CC) are used in the analysis.

Bene t Incidence Analysis
To determine the distribution of bene ts received by various socio-economic groups using public health services for delivery care, Bene t Incidence Analysis has been used. One of the di culties in bene t incidence analyses is obtaining the cost of services. In the absence of cost of services, the modal value of OOP payment for delivery has been used in literature [43]. However, we have used the median value in our study since a signi cant proportion of women reported zero OOP (varying from 7-10% across the wealth quintile) of delivery at the accredited private health centres (JSY under NHM programme has such provision) and so the modal value for OOP payments became zero. Further, to examine the robustness of the result obtained from median value, we have also estimated bene t incidence using mean value. We have estimated the bene t incidence of a particular group j utilizing service i (institutional delivery) from public health centres. The OOP payment in private health centres has been taken as synonymy to cost of services. In case of maternal care, most of the health insurances in India do not provide any coverage/reimbursement and so OOP is equivalent to household expenditure. In case, no charge was levied, the OOP payment was considered as zero.
Mathematically, the formula for calculating Bene t Incidence is as follows: The following steps have been used in the study.
i. Computing wealth quintile (population ranked by wealth) as a measure of socio-economic status.
ii. Estimating the utilization rate for delivery care in public health centres for each quintile.
iii. Estimating net subsidy at public health centres for each quintile (obtained by deducting the median OOP payment on delivery care in public health centres from median OOP payment in private health centres) iv. Estimating individual subsidy for each quintile by multiplying the net subsidy with the utilization rate.
v. Calculating Bene t Incidence for each quintile by taking percentage share of individual subsidy.

OOP payment and Cost of service on institutional delivery
We have computed the OOP payment by quintile for mothers delivering at public health centres. Further, NFHS 4 does not provide any information on the actual cost of delivery care at the public health centre. Hence in line with previous literature, we have used OOP payment on delivery care in private health centres as the proxy for the actual cost of delivery care in public health centres [44][45].

Concentration Index (CI) and Concentration Curve (CC)
To examine the economic inequality in the utilization of delivery care services from public/private health centre, CC and CI were used. CC and CI are the commonly used measures by researchers across the globe to measure health inequality [24,[49][50]. The inequality is graphically represented through CC and plots the cumulative proportion of the population (ranked by wealth) against the cumulative proportions of the population utilizing delivery care services from public/private health centres. If CC overlaps with the line of equality, then the extent of utilization of services from public/private health centres is evenly distributed across the wealth group. However, if CC lies above the line of equality, it implies a pro-poor concentration of utilization of delivery care services while if CC lies below the line of equality, it implies a pro-rich concentration in utilization of delivery care services. CI is de ned as twice the area between the CC and the line of equality. The value of CI ranges from -1 to +1, with a zero value of CI suggesting equal distribution of utilization of services across the wealth group. The negative value of CI signi es a pro-poor distribution of utilization of delivery care services while a positive value of CI signi es a pro-rich distribution. Figure 1 presents the distribution of institutional delivery by wealth quintile and type of health centres in India. Among all institutional deliveries in poorest wealth quintile, 86% were in public health centre compared to 14% in private health centre. The utilization of institutional delivery from public health centres declines with an increase in the economic well-being of the households. On the other hand, economic gradient in utilization of institutional delivery in private health centres is strong and positive. In richest wealth quintile, about two-third utilized private health centre for delivery care. However, majority of the women from the poorest, poorer and middle quintile utilized public health centre for delivery. Table 1 presents the % of women who paid or availed services without payment at private and public health centre by wealth quintile in India. Among those availed services from public health centres, the proportion of women who did not pay for delivery care increases by wealth quintile. For example, about 17% of women in poorest wealth quintile did not pay for delivery care compared to 28% in richest wealth quintile. In case of private health centres the proportion is similar across wealth quintile.   India  Poorest 3760  15947  1100  27741  2484  15475  1000  24518  11628  16551  7000  3223   Poorer  5308  19850  1500  31802  3440  20240  1100  26635  12900  16082  8200  5167   Middle  7160  19834  2100  31891  3679  19045  1200  24053  14997  19335  10000  7838   Richer  9326  16824  3900  29825  3392  12909  1200  18676  16593  18158  12000  11149   Richest  13940  20090  8000  27386  4096  17151  1000  11733  19305  19536  14600  15653   Total  7984  18957  2200  148645  3331  17395  1050  105615  16513  18736  11000 Middle  5439  25056  1000  2166  3564  26977  700  1733  12705  13215  9000  433   Richer  5119  11386  1300  1933  1685  7878  500  1372  12736  13987  8200  561   Richest  9308  28185  3000  2050  4364  35909  500  1076  14618  14459  10000  974   Total  5574  23069  1000  10272  2972  25032  500  7745  13046  13635  8600  2527   Maharashtra  Poorest 7937  39790  1000  548  7367  44088  700  452  10316  8604  9000  96   Poorer  5494  17842  1500  1192  3111  18548  1000  904  12770  13046  8590  288   Middle  8203  18315  3100  1672  4201  17646  1200  1089  14862  17462  9200  583   Richer  9381  14385  5000  1689  3070  8245  1402  832  15887  16339  10500  857   Richest  15426  21183  10040  1349  4642  17193  1900  334  18811  21192  13200  1015   Total  9840  20468  4500  6450  4019  20491  1100  3611  16414  18344  11000  2839   Karnataka  Poorest 4933  10304  1500  370  4032  8656  1200  333  12846  17848  7100  37   Poorer  5584  13400  1800  1213  2966  7356  1200  1011  18662  24560  13000  202   Middle  7499  17416  2500  1708  4291  15903  1500  1264  16823  18268  11000  444   Richer  10376  17291  4050  1537  3940  6531  2000  875  19217  22726  14030  662   Richest  20130  23929  12000  711  8833  15083  4000  187  24843  25332  16000  524   Total  10220  18567  3300  5539  4249  11605  1700  3670  20650  23256  15000  1869   Table 3 presents the bene t incidence of the public subsidy on delivery care by wealth quintile in India. The utilization rate at public health centre varied from a maximum of 24.8% among poorest quintile to 12.3% among richest quintile. During 2015-16, public subsidy was the maximum for the middle wealth quintile (21.9%) followed by richer (21.8%) wealth quintile while it was least for the poorest quintile (17.4%). Considering the mean cost of service in private health centre, the bene t of public subsidy was maximum for the women from the middle quintile (21.9%) followed by richer quintile (20.8%) (Appendix 1).  Table 4 presents results of bene t incidence on institutional delivery by place of residence, low/high performing states and educational attainment of women in India. In the urban area, the utilization rate from public health centre varied from 28.2% among women belonging to the poorest quintile to 11.6% among women from richest quintile whereas in rural areas, the utilization rate varies from 14.5% among women from richest quintile to 23.2%

Result
among women belonging to the poorest quintile. Considering median OOP of delivery care, in urban area, the share of bene t received was highest for women belonging to poorer quintile (23.1%) followed by women from middle quintile (22.5%) and poorest quintile (20.5%). In case of rural areas, the share of public subsidy was highest among women belonging to richer quintile (21.7%) followed by middle (20.9%) while it was least among women from poorest quintile (17.5). Further, bene t incidence was computed for low and high performing states. Institutional delivery in Low Performing States (LPS) varied from 12.9% among women from richest wealth quintile to 23.5% among women in poorest quintile whereas it varied from 12.5% among women from richest quintile to 25.7% among woman from poorest quintile in High Performing States (HPS). Considering the median OOP of service, the share of public subsidy was maximum for the women from middle quintile in both LPS and HPS. However, the share was minimum for the poorest quintile (18.7%) in LPS and richest quintile in HPS (17.7%). When mean OOP of delivery care was taken, the public subsidy was maximum for the women from the poorer wealth quintile (24.3%) in LPS and middle quintile in HPS (22.7%) (Appendix 2). Also, the bene t incidence of the public subsidy was computed for women with education less than ve years and more than ve years. The utilization rate of public health centre among women with less than ve years of education varied from 21.9% among women from richest quintile to 16.7% among women from richest quintile while it varied to 25.9% among women from poorest quintile 12.1% among women from richest quintile for women with more than ve years of education. Considering the median OOP of delivery care service, the bene t incidence of public subsidy was highest among women's from the richer quintile (22.1%) having less than ve years of education In comparison the share was highest among women from middle wealth quintile (21.4%) having more than ve years of education.  Figure 2 presents the concentration curve (CC) for women utilizing institutional delivery at public and private health centre respectively. The CC for women utilizing public health centre was above the line of equality indicating a pro-poor concentration of use of public health centre for delivery care services whereas CC was below the line of equality for women utilizing private health centre suggesting a pro-rich concentration of use of private health centre for delivery care services. Table 5 presents the public-private differential of concentration index for institutional delivery by place of residence, low/high performing states and educational attainment of women in India. The CI values was higher for women residing in urban area using public health centre (CI: -0.209) for delivery care compared to women delivering in private health centre (CI: -0.112). Similarly, the CI for mother using private health centre for delivery services was higher for mothers from rural area (CI: 0.281) compared to urban area (CI: 0.217). The CI value of women residing in HPS (-0.177) using public health service was higher compared to their counterparts residing in LPS (-0.113). On the contrary, the CI value was higher for women residing in LPS (0.318) utilizing private health centre for delivery care services compared to those residing in HPS (0.226).

Discussion
Resource constraints in the public healthcare system poses several challenges to meet the healthcare need of the poor and marginalized population and likely to increase the socioeconomic inequality in health outcome. Resources used for public health services has opportunity cost and in this context equity in health care is assumed to be signi cant. The NHM in India, the largest ever public health programmes worldwide that has been operational for over 15 years. While over half of the national health resources were invested in NHM, large-scale investment aimed to achieve multiple objectives including increasing service coverage, reducing inequality in health care and health outcomes and reducing OOP payment and CHS speci cally among the poor and disadvantages. Although the priorities of these schemes are usually aimed at bene ting the economically weaker section of the population, no attempt has been made to understand the bene t-incidence of these programmes in India. This is the rst ever study to investigate the distribution of public subsidy among mothers using public health centres considering institutional delivery as the case from NFHS-4, a large scale population based survey. The salient ndings of the paper are as follows: First, the utilisation of delivery care services has a strong economic gradient. Mothers belonging to the poorest and poorer wealth quintile use more of public health centre for delivery care while mothers from richer and richest wealth quintile use private health centre for delivery care services. Second, among those using public health services, a higher proportion of mothers from richest and richer wealth quintile did not pay for the services compared to poorest and poorer wealth quintile. Third, the median OOP on delivery care was about seven times higher for mothers belonging to richest wealth quintile compared to mothers from poorest quintile. The pattern was similar for deliveries in public and private health centres. However, large variation was observed in case of private health centres. Fourth, the public subsidy was relatively used more from the mothers in the middle wealth quintile suggesting that the public subsidy was not pro-poor in nature. The public subsidy in urban area was used among the poor section of the society whereas in rural areas, it was largely used among the richer section of the society when both mean and median OOP was considered. Further, the public subsidy was higher among mothers from poorer section of the society in LPS whereas the subsidy was higher among mother from middle section of the society in HPS. Fifth, the CC for mother utilizing public health centres for delivery care was above the line of equality suggesting a pro-poor concentration of utilization of public health service whereas the CC was below the line of equality suggesting a pro-rich concentration of utilization of private health services. The CI value of -0.161 for public health centre and 0.296 for private health centre further con rms the concentration of use public health centres among the poor and private health centre among the rich. Further, the state variation in concentration index was large for both public and private health service.
We provide some plausible explanation to our ndings. The utilization of delivery care from public health centres is higher among the poorest and poorer section of the population as these services at public health centres are provided at free or very nominal cost and poor people has limited ability to pay for services. On the other hand, services from private health centres are expensive in nature and are mostly used by richer wealth quintile. The higher dependency and utilization from the private health sector led to rising inequality in the health care services. The OOP expenditure incurred during delivery may be due to the complications in delivery care, caesarean delivery, cost of medicine, transportation and accompanying person etc. This OOP payment has strong and positive economic gradient. The reason is, mothers from higher economic strata have higher ability to pay for services and so they seek for better quality of care [51]. However, among those using the public health centres for delivery care, the extent of bene t was relatively higher among richer and richest quintile. The potential cause of this trend is that, the poorer women are paying more for similar services at public health centres. Although, the marginalized women should have received various reimbursement and incentives from NHM and other maternal programmes, studies suggest that, these incentives are either insu cient or there are some other factors accounting for utilization inequality, such as, low education attainment, low quality public health facility in poorer areas [20,52]. The cash assistance programmes are not enabling factors for institutional delivery as there are some other informal payments associated with it, which in turn, opts out the marginalized section [53]. On the other hand, women from richer quintile are more aware about the ongoing programmes and incentives; so they enjoyed more bene ts compared to poor. Regional variation in subsidy utilization could be another possible reason of unequal distribution of public subsidies. For instance, poor mothers from LPS enjoy higher bene t of subsidy which can be explained by the introduction of various maternal and child health programmes under NHM. The JSY under NHM provides conditional cash assistance to the eligible women from weaker section of the population. Although inequality still exists, the level of inequality has reduced and institutional delivery has increased signi cantly across all groups in LPS [52,54]. Besides increasing facility based delivery, JSY has signi cantly increased contraceptive use, breastfeeding practice and post-natal check-up which are closely associated with accessing public health facilities [55].
We outline the following limitations of the study. First, we have used NFHS data to estimate utilization pattern, OOP payments, bene t incidence based on self-reported, there could be some recall bias. Second, we have used median cost of private health centres as proxy to cost of services in public health centres. An appropriate study on costing could be more robust to provide exact scenario. Third, our results could not cover the impact of recent initiatives such as effect of Ayushmann Bharat, Pradhan Mantri Matru Vandana Yojana etc.

Conclusion
Public health spending should bene t the poor and marginalized section of the society to achieve equity in health outcome. Focused policies are needed in order to maintain equity across all the wealth quintile. At the national level, policies such as, Rashtriya Swasthya Bima Yojana (RSBY), Ayushman Bharat, Pradhan Mantri Matru Vandana Yojana (PMMVY) has been providing the protection against nancial risks to the economically weaker section of the population. These policies are signi cant to change the very outline of health care access, utilization and OOP expenditure. It is very much recommended to continue such policies with greater monitoring surveillance to make it more pro-poor, so that disadvantageous can get the substantial support.

Declarations
Ethics approval and consent to participate: As the analysis is based on secondary data available in the public domain, it needs no prior approval.
Consent for publication: This manuscript is an original work and has been done by the authors, SKM, RM, SM and SS who all are aware of its content and approve its submission. This manuscript has not been published elsewhere in part or in entirety, and is not under consideration by another journal. All authors gave their consent for publication in International Journal for Equity in Health.
Availability of data and material: The dataset used and analysed for the current study is available in DHS repository, [https://dhsprogram.com/data/dataset/India_Standard-DHS_2015.cfm? ag=0] Competing interests: The authors declare that they do not have any competing interest.  Concentration curve for mothers using delivery services at public and private health facility in India, 2015-16.

Figure 3
Concentration Index of institutional delivery by public and private facility in selected states of India, 2015-16. Supplementary Files