Distribution of government health financing benefits among women who delivered in public institutions in Bangladesh: a nationally representative cross-sectional study

Summary

Background

Equitable access to institutional delivery care is crucial for reducing maternal mortality. Although Bangladesh has made progress in this regard, significant challenges persist in achieving equitable access to institutional delivery care, particularly for economically disadvantaged populations. The objective of this study was to investigate the distribution of public health financing benefits among women who delivered in public facilities in Bangladesh.

Methods

This study was conducted based on the data from the Bangladesh Demographic and Health Survey (BDHS) 2022, which includes a sample of 3360 women (age 15–49 years) who had a history of institutional delivery during two years preceding the survey. Descriptive and econometric analyses, including Benefit Incidence Analysis (BIA), Concentration Curves (CC), and Concentration Indices (CIX), were employed to assess the distribution of public subsidies for delivery care among different socio-economic groups. Socioeconomic inequality in utilisation of delivery care services was evaluated using concentration curves, while BIA estimated the distribution of public healthcare benefits across wealth quintiles. Logistic regression was used to determine the factors associated with distress financing—whether the household of the women had to sell their assets or had to resort of borrowing to avail services.

Findings

Among the poorest quintile, 38% utilised public facilities, compared to 17% of the women with highest income. The concentration curve for public facility use indicated a pro-poor distribution (CIX −0.031). BIA revealed that the poorest quintile received 24.5% of public subsidies, whereas the wealthiest quintile received 13.7%. However, in tertiary care facilities and for caesarean delivery, the wealthiest group benefitted the most, receiving 23.5% and 26% of the subsidies, respectively. Odds of distress financing was significantly higher among women from poorer or poorest households compared to the richest group (aOR 4.35, 95% CI 3.16–6.03 for poorest and aOR 2.74, 95% CI 2.00–3.77 for poorer).

Interpretation

Public health subsidies in Bangladesh equitably benefit women with lower income, though inequities remain, particularly in tertiary care facilities and for caesarean deliveries. Despite this, women with lower income are more vulnerable to distress financing for delivery care.

Funding

None.

Research in context.

Evidence before this study

Before undertaking this study, we conducted a comprehensive review of existing literature on the distribution of public health financing benefits for maternal care in Bangladesh. We searched databases including PubMed, Scopus, and Google Scholar, as well as reports from the Bangladesh Demographic and Health Survey (BDHS) (2011–2022) and government health expenditure reports. The search included studies published up to 2024 in English, using terms such as “benefit incidence analysis,” “maternal health equity,” “public health subsidies,” and “distress financing in Bangladesh.” Prior research has established that public health services in Bangladesh are more frequently utilized by lower-income populations. However, studies examining the equitable distribution of government health subsidies in institutional delivery care are limited. Existing evidence indicates persistent socioeconomic disparities in maternal healthcare access. However, detailed benefit incidence analysis (BIA) focusing on institutional deliveries in public health facilities remained scarce.

Added value of this study

Our study provides an in-depth assessment of the distribution of public health subsidies for institutional delivery care in Bangladesh, specifically through the lens of benefit incidence analysis. We find that while public healthcare services predominantly benefit women with low-income, significant disparities exist at the tertiary care level, where women with high-income receive a disproportionately high share of public subsidies. Additionally, women with lower-income remain highly vulnerable to distress financing despite receiving public subsidies, highlighting financial barriers in access to maternal care. This study bridges a crucial knowledge gap by illustrating how financial protection mechanisms remain insufficient for disadvantaged populations.

Implications of all the available evidence

The findings of this study suggest that while public delivery care facilities in Bangladesh support equitable access at lower-tier institutions, targeted policy interventions are necessary to improve financial protection for the most vulnerable populations. Strengthening referral mechanisms and implementing community-based health insurance schemes could mitigate access and economic barriers. Additionally, policies should address the disproportionate allocation of subsidies at tertiary care facilities to ensure fairer distribution of healthcare resources.

Introduction

The Sustainable Development Goals (SDGs) include a critical target aimed at reducing the global maternal mortality ratio (MMR) to less than 70 per 100,000 live births by 2030.¹ Bangladesh, as a committed signatory to the SDGs, has made significant progress in this area, with the current MMR standing at approximately 136 per 100,000 live births—a notable reduction from 322 in 2001.² However, despite these advancements, considerable challenges remain, particularly in achieving universal access to institutional deliveries, which are essential for improving maternal health outcomes. Approximately one-third of women in Bangladesh continue to give birth at home without skilled assistance, predominantly among poorer households and rural communities.³ Multiple factors contribute to the persistence of home births, including cultural traditions, geographical barriers, and, most critically, financial constraints.3, 4, 5, 6 These challenges emphasise the entrenched disparities in access to maternal healthcare, particularly for the most vulnerable segments of the population.⁷

The healthcare system in Bangladesh operates as a mixed model, comprising both public and private sectors. Public healthcare facilities, spanning from primary health centres to tertiary hospitals, provide essential maternal and child health services, with a specific focus on ensuring the access to the services by the low-income populations. Government initiatives have expanded the availability of services such as antenatal care and emergency obstetric care, aiming to reduce maternal mortality and morbidity.⁸ While these initiatives demonstrated some success in increasing access to maternal health services, the coverage remains insufficient to meet the needs of all eligible women, leaving significant gaps in service delivery.9, 10, 11

Despite these efforts, many public healthcare facilities face constraints in terms of resources and infrastructure, limiting their capacity to deliver high-quality care consistently.^12,13 Although government policy provides maternal and newborn care under the Essential Health Service Package,¹⁴ the availability of essential supplies is often inadequate, particularly in primary care facilities. Consequently, patients frequently incur out-of-pocket (OOP) expenses to obtain necessary treatments that should be available at no cost. Besides, the allocation of the government health budget of the country across various levels of healthcare facilities is not equally distributed. For instance, in the 2020–2021 budget of the Government of Bangladesh, 25% of the total health budget was allocated to primary care facilities, 39% to secondary care facilities, and 36% to tertiary care facilities. As a result, low-income families, who are often reliant on public healthcare, may experience limitations in service provision. In response, some turn to private healthcare providers, where the costs of care are significantly higher.¹⁵ Even within public facilities, wealthier households often incur higher OOP expenses, as they frequently opt for additional services such as private cabins, better-quality medications sourced from outside the hospital, and improved food options. Additionally, informal payments or provider-initiated expenses, such as investigations outside the public facilities, may disproportionately affect wealthier patients, as healthcare staff may perceive them as being able to afford such costs. Besides, the intended affordability of public healthcare facilities is often undermined by out-of-pocket expenses, which can include the cost of transportation, medications, and supplies that are often not available in public facilities.¹⁶ These financial burdens might exacerbate inequities in access to public facilities, limiting the ability of economically disadvantaged populations to fully benefit from government-funded healthcare services. However, limited studies focused on the estimating those persistent inequities.

In this context, Benefit Incidence Analysis (BIA) provides an efficient framework for evaluating the distribution of public healthcare benefits among different socio-economic classes. BIA assesses the extent to which different segments of the population benefit from public spending, helping to determine whether government-funded services are effectively reaching the most vulnerable populations.¹⁷ Despite established inequities in utilisation of public-funded maternal care in Bangladesh,¹⁸ there are limited attempts of in-depth investigation of these inequities regarding child delivery care. Hence, the objective of the present study was to investigate the distribution of government health financing benefits among women who delivered in public institutions in Bangladesh.

Methods

Data source

We used data from the 2022 Bangladesh Demographic and Health Survey (BDHS), a nationally representative survey that captures comprehensive information on household socio-demographic and economic characteristics, maternal and child health, family planning, sexual and reproductive health, non-communicable diseases, healthcare utilisation, and more.¹⁹ The BDHS employed a multistage stratified sampling design, ensuring coverage across urban and rural areas of Bangladesh. A total of 19,987 women (age 15–49 years) were interviewed: 7007 from urban and 12,980 from rural areas. For the present analysis, we focused on a subset of the BDHS data specifically related to institutional deliveries that occurred within the two years preceding the survey. Our sample included women with complete data on out-of-pocket expenditure (OOPE) for child delivery care, resulting in a final sample size of 3360 women (age 15–49 years) (Fig. 1). The protocol for the 2022 BDHS received ethical clearance from both the ICF Institutional Review Board ethics committee and the Bangladesh Medical Research Council (BMRC). Informed written consent was obtained from each participant before commencement of the interview.

Fig. 1.

Participant selection procedure for the study. OOPE: out-of-pocket expenditure.

Variables

Outcome variables

We considered two outcome variables in our analysis: OOPE for child delivery care and distress financing for these expenditures. We defined OOPE as the total amount a woman or her household paid for child delivery care in a healthcare facility, without receiving any reimbursement. In the BDHS 2022, the survey captured this information through a direct question asked to the mother: “How much did you pay in total for your last delivery?”

In addition to OOPE, we also examined distress financing, which was defined as the households resorted to borrowing money or selling assets to cover the costs of child delivery care. Distress financing was identified based on the mother’s response to the question: “What was the source of the money required for this delivery?” If she indicated borrowing, selling or mortgaging assets as a source of the fund, either partially or completely, we classified it as distress financing.

Independent variables

We included several independent variables including residence (urban or rural), maternal age (categorised as <25 years or ≥25 years), educational attainment (<10 years or ≥10 years of schooling), wealth quintile, number of antenatal visits (fewer than four visits or four or more visits), place of delivery (public or private facility), and mode of delivery (vaginal delivery or caesarean section). The wealth quintile variable was constructed using an asset-based wealth index that incorporated various household characteristics, including land ownership, drinking water source, access to electricity, type of housing, sanitation facilities, household durables, and the number of rooms per person. Principal component analysis (PCA) was used to generate a wealth score, which was then used to categorise households into five wealth quintiles: poorest, poorer, middle, richer, and richest.¹⁹

In the context of health system of Bangladesh, public facilities included community clinics, union health and family welfare centres, upazila (sub-district) health complexes, district hospitals, maternal and child welfare centres, medical college hospitals, and specialised government hospitals. We further categorised these into primary care facilities (community clinics, union health and family welfare centres, and upazila health complexes), secondary care facilities (district hospitals, and maternal and child welfare centres), and tertiary care facilities (medical college hospitals and specialised government hospitals). Private delivery care facilities encompassed private clinics, private hospitals, and private medical college hospitals.

Statistical analysis

We conducted both descriptive statistics and bivariate analysis, as well as econometric analysis, including Benefit Incidence Analysis (BIA), Concentration Curves (CC), and Concentration Indices (CIX). Categorical variables were summarised using frequencies and percentages, while continuous variables were presented as means and standard deviations (SD). We excluded any missing data when calculating proportions. A p value less than 0.05 was considered as statistically significant. All the analyses were conducted using R (version 4.2.3).

Concentration curve and concentration index

To assess socio-economic inequality in the utilisation of delivery care services, we constructed CCs and CIXs. The CC visually represents the cumulative proportion of women, ranked by wealth index, against the cumulative proportion of women who utilised either public or private healthcare facilities for childbirth. A CC below the line of equality indicates that women from wealthier households were more likely to use these services, whereas a curve above the equality line suggests that women from poorer households had higher utilisation rates. As our variable of interest (type of delivery care facility) was binary, we applied Erreygers’ correction to the CIX to ensure accurate measurement of socioeconomic inequality. This correction is essential for binary outcomes and provides a normalized index, where negative values suggest that utilisation is concentrated among women with lower income and positive values suggest concentration among women with higher income.

Benefit incidence analysis

We employed BIA to explore the extent to which public subsidies for institutional child delivery care were distributed across different socio-economic groups. BIA is an established method for evaluating whether public healthcare spending benefits poorer segments of the population. Since the BDHS 2022 did not collect direct data on the actual cost of delivery care in public healthcare facilities, we used the median OOPE at private facilities as a proxy for the actual cost of delivery care at public facilities. Additionally, there is a paucity of data regarding the actual cost of delivery care in Bangladesh. Although some studies have estimated the government expenditure for essential maternal and childcare packages in Bangladesh,¹⁴ these estimates do not reflect the total cost, which is necessary for conducting a BIA. Instead, these estimates primarily represent the government’s expenditure on service provision. While it is acknowledged that the cost of public sector services may exceed the actual delivery care costs, due to the profit-maximisation objectives of private providers and the insufficient regulation of the private sector in Bangladesh, the lack of accurate service cost data has led several studies, including one from India, to use OOPE at private facilities as a proxy for estimating treatment costs.²⁰

To estimate benefit incidence, we first ranked households by their wealth index and grouped them into quintiles. Then, we calculated the utilisation rate of public healthcare facilities for child delivery in each wealth quintile. The net public subsidy was estimated by subtracting the median OOPE in public facilities from that in private facilities, assuming that the difference represents the cost subsidised by the government. The individual subsidy was calculated by multiplying the utilisation rate for each quintile by the net public subsidy, and finally, the overall benefit incidence was calculated as the proportion of the total subsidy received by each wealth quintile. For subgroup analysis of the individual subsidy, we considered the group-specific median OOPE in the private facilities of the respective subgroup. For instance, subsidy in the primary, secondary and tertiary care facilities were calculated based on the median OOPE in the private facilities at the similar levels. Similarly, for subsidies in the urban and rural areas were calculated based on the median OOPE of the private facilities of that geographic location.

Distress financing

To assess the prevalence of distress financing across wealth quintile, we calculated the proportion of women who relied on borrowing, asset sales or mortgages to fund their childbirth expenses. To further explore the determinants of distress financing, we used logistic regression models, adjusting for potential covariates such as residence, maternal age, education, and mode of delivery. Separate models were fitted for overall delivery care, as well as for public and private facilities, to explore whether the risk for distress financing among different socio-economic groups differed between these settings.

Role of the funding source

There was no funding source for this study.

Results

Sociodemographic characteristics

Approximately, 38% of the women in the sample resided in rural areas, with an average age of 25.6 years (SD 5.52 years). Nearly 59% of the women had completed fewer than ten years of schooling. Socioeconomically, the distribution was fairly even, with 16% falling into the poorest quintile, 17% in the poorer, 21% in the middle, 22% in the richer, and 24% in the richest quintile. Just over half of the women attended more than four antenatal care (ANC) visits. The majority (around 73%) gave birth at private facilities, and a notable proportion of these institutional deliveries were by caesarean sections (Table 1).

Table 1.

Socio-demographic characteristics of the participants (n = 3360).

Characteristic	n	%
Residence
Urban	1288	38.33
Rural	2072	61.67
Maternal age (years) (mean 25.6, SD 5.52)
<25	1576	46.9
≥25	1784	53.1
Educational attainment (years)
<10	1994	59.35
≥10	1366	40.65
Wealth quintile
Poorest	530	15.77
Poorer	579	17.23
Middle	699	20.80
Richer	743	22.11
Richest	809	24.08
Number of antenatal visits
<4	1565	48.72
≥4	1647	51.28
Place of delivery
Public facility	896	26.67
Private facility	2464	73.33
Mode of delivery
Vaginal delivery	1002	29.82
Caesarean section	2358	70.18

Utilisation of public and private delivery care facilities

The use of public delivery care facilities showed a clear decline as socio-economic status increased. Among women in the poorest quintile, around 38% delivered in public facilities, whereas only 17% of the wealthiest women opted for public care. In contrast, private delivery facilities were used by 62% of women in the poorest quintile, compared to nearly 83% in the richest (Fig. 2).

Fig. 2.

Proportion of the women delivered at public and private facilities (n = 3360).

The concentration curve for public facility utilisation (Fig. 3a) appeared above the equality line, with a concentration index (CIX) of −0.031 (SE 0.002), indicating a strong concentration of public service use among lower socio-economic groups. Conversely, the curve for private facility use fell below the line of equality, yielding a CIX of 0.011 (SE 0.001), highlighting a concentration of private care use among wealthier groups. This pattern held consistent across both urban (CIX −0.026, SE 0.002 for public facilities and CIX 0.017, SE 0.001 for private facilities) (Fig. 3b) and rural settings (CIX −0.019, SE 0.002 for public facilities and CIX 0.011, SE 0.001 for private facilities) (Fig. 3c), as well as for vaginal deliveries (CIX −0.009, SE 0.001 for public facilities and CIX 0.013, SE 0.002 for private facilities) (Fig. 3d). For caesarean deliveries, however, while public facility use remained concentrated among poorer women (CIX −0.022, SE 0.003), private facility use showed a more even distribution across socio-economic groups, with a minimal CIX of 0.003 (SE 0.001), indicating no strong economic bias in accessing private care for caesareans (Fig. 3e).

Fig. 3.

Concentration curve and concentration Index of institutional delivery by public and private facilities (a. Overall, b. Urban facilities, c. Rural facilities, d. Vaginal delivery, e. Caesarean section).

Benefit incidence analysis

Overall, the poorest socio-economic group received the largest share of the public subsidy at 24.5%, while the wealthiest quintile received the lowest allocation at 13.7%. This distribution pattern persisted for both primary and secondary care facilities, where the poorest groups were the primary beneficiaries. However, a distinct pattern emerged at tertiary care facilities, where the wealthiest quintile received the largest share of the subsidy (23.5%), surpassing all other groups, including the poorest (20.3%) and middle-income groups (21.7%). In urban areas, women from the poorest, poorer, and middle quintiles received 32%, 28%, and 21% of the benefits, respectively. This contrasts with rural areas, where benefit allocation was more evenly distributed across socio-economic strata, with the middle and richer quintiles consuming a substantial 44% of the benefits jointly.

Regarding vaginal deliveries, allocation of benefits was high in the poorest quintile, which received 36% of the subsidy, followed by the poorer group with 23.5%. In contrast, for caesarean sections, the wealthiest quintile received the highest portion of the public subsidy (26%), followed by the middle-income (24.5%) and poorest groups (18.5%). Among women with lower educational attainment (possibly due to an opportunity gap), the poorest quintile received the highest proportion of subsidised care at 29.7%, while the richest quintile received only 10%. However, a reverse pattern was observed for women with higher educational attainment. In this group, the richest and richer quintiles each received 23% of the public subsidy, followed by the middle (22.5%) and poorer (18%) groups. The poorest quintile, in this case, received the smallest share of the benefits at 12% (Table 2).

Table 2.

Utilisation rate, out-of-pocket payment (OOP, in Bangladeshi Taka [BDT]), and benefit incidence of delivery care at public facilities by wealth quintiles.

Characteristics	Wealth quintile	Number women utilising public facilities (1)	Utilisation rate by wealth quintile (2)	Median OOP expenditure in public facilities reported by participants (3)	Median cost of service in private facilities reported by participants (4)	Net subsidy at public facilities (5 = 4−3)	Individual subsidy (6 = 5 × 2)	Benefit incidence (%) (7 = 6/total individual subsidy x100)
Overall	Poorest	200	0.22	5000	20,000	15,000	3348	24.49
	Poorer	171	0.19	5000	20,000	15,000	2863	20.94
	Middle	192	0.21	6000	20,000	14,000	3000	21.94
	Richer	193	0.22	8000	20,000	12,000	2585	18.91
	Richest	140	0.16	8000	20,000	12,000	1875	13.72
Level of care
Primary	Poorest	99	0.27	3000	20,000	17,000	4624	28.85
	Poorer	73	0.20	3500	20,000	16,500	3309	20.65
	Middle	66	0.18	4000	20,000	16,000	2901	18.10
	Richer	77	0.21	5000	20,000	15,000	3173	19.80
	Richest	49	0.13	5000	20,000	15,000	2019	12.60
Secondary	Poorest	77	0.22	7000	20,000	13,000	2852	23.93
	Poorer	66	0.19	7000	20,000	13,000	2444	20.51
	Middle	90	0.26	7000	20,000	13,000	3333	27.97
	Richer	66	0.19	10,000	20,000	10,000	1880	15.78
	Richest	52	0.15	10,500	20,000	9500	1407	11.81
Tertiary	Poorest	24	0.13	6000	20,000	14,000	1856	20.29
	Poorer	32	0.18	10,000	20,000	10,000	1768	19.32
	Middle	36	0.20	10,000	20,000	10,000	1989	21.74
	Richer	50	0.28	15,000	20,000	5000	1381	15.10
	Richest	39	0.22	10,000	20,000	10,000	2155	23.55
Residence
Urban	Poorest	84	0.25	5000	20,000	15,000	3772	31.66
	Poorer	74	0.22	5000	20,000	15,000	3323	27.89
	Middle	65	0.19	7000	20,000	13,000	2530	21.23
	Richer	69	0.21	15,000	20,000	5000	1033	8.67
	Richest	42	0.13	10,000	20,000	10,000	1257	10.55
Rural	Poorest	116	0.21	4500	20,000	15,500	3199	21.37
	Poorer	97	0.17	5000	20,000	15,000	2589	17.29
	Middle	127	0.23	5000	20,000	15,000	3390	22.64
	Richer	124	0.22	5000	20,000	15,000	3310	22.10
	Richest	98	0.17	5750	20,000	14,250	2485	16.60
Mode of delivery
Vaginal delivery	Poorest	151	0.26	3000	7000	4000	1050	35.85
	Poorer	113	0.20	3500	7000	3500	688	23.47
	Middle	127	0.22	4500	7000	2500	552	18.84
	Richer	113	0.20	5000	7000	2000	393	13.41
	Richest	71	0.12	5000	7000	2000	247	8.43
Caesarean section	Poorest	49	0.15	15,000	22,000	7000	1069	18.50
	Poorer	58	0.18	16,250	22,000	5750	1039	17.98
	Middle	65	0.20	15,000	22,000	7000	1417	24.53
	Richer	80	0.25	19,000	22,000	3000	748	12.94
	Richest	69	0.21	15,000	22,000	7000	1505	26.04
Educational attainment
<10 years	Poorest	173	0.28	5000	20,000	15,000	4185	29.66
	Poorer	131	0.21	5000	20,000	15,000	3169	22.46
	Middle	135	0.22	6000	20,000	14,000	3048	21.60
	Richer	117	0.19	8000	20,000	12,000	2265	16.05
	Richest	64	0.10	6000	20,000	14,000	1445	10.24
≥10 years	Poorest	27	0.10	5000	21,000	16,000	1565	12.20
	Poorer	40	0.14	5000	21,000	16,000	2319	18.07
	Middle	57	0.21	7000	21,000	14,000	2891	22.53
	Richer	76	0.28	10,000	21,000	11,000	3029	23.60
	Richest	76	0.28	10,000	21,000	11,000	3029	23.60

Note: Opportunity gap may impact educational attainment. US$ 1 = 121 BDT (as of April 10, 2025).

Distress financing

Distress financing was most prevalent in the poorest socio-economic quintile. Specifically, 34% of women in this group relied on distress financing, followed by 25.5% in the poorer quintile, 22% in the middle quintile, 21% in the richer quintile, and only 10% in the richest quintile. This trend was consistently observed across both public and private delivery care facilities, where the poorest segments of society exhibited the highest rates of distress financing compared to wealthier groups (Table 3).

Table 3.

Prevalence and risk of distress financing for delivery care according to wealth quintiles (Logistic regression model).

Wealth quintile	Overall			Public			Private
	n (%)	aORa (95% CI)	p-value	n (%)	aORa (95% CI)	p-value	n (%)	aORa (95% CI)	p-value
Poorest	184 (34.72)	4.35 (3.16–6.03)	<0.001	56 (28.00)	6.12 (2.90–14.10)	<0.001	128 (38.79)	4.04 (2.82–5.84)	<0.001
Poorer	148 (25.56)	2.74 (2.00–3.77)	<0.001	27 (15.79)	2.58 (1.18–6.03)	0.022	121 (29.66)	2.85 (2.01–4.05)	<0.001
Middle	154 (22.03)	2.34 (1.73–3.18)	<0.001	30 (15.63)	2.82 (1.32–6.48)	0.011	124 (24.46)	2.30 (1.65–3.21)	<0.001
Richer	156 (21.00)	2.18 (1.62–2.94)	<0.001	26 (13.47)	2.15 (0.99–4.96)	0.058	130 (23.64)	2.23 (1.62–3.09)	<0.001
Richest	83 (10.26)	1		10 (7.14)	1		73 (10.91)	1

^aAdjusted for residence, maternal age, educational attainment and mode of delivery.

In the logistic regression analysis, the poorest socio-economic group was almost four times more likely to experience distress financing (aOR 4.35, 95% CI 3.16–6.03, p < 0.001) compared to the richest group. Similarly, women from the poorer quintile faced 2.7-fold increased odds (aOR 2.74, 95% CI 2.00–3.77, p < 0.001). The association between socio-economic status and distress financing persisted when stratifying by the type of facility used. Among women who gave birth in public facilities, those in the poorest quintile had nearly six times higher odds (aOR 6.12, 95% CI 2.90–14.10, p < 0.001) compared to those in the richest quintile. A similar but slightly reduced disparity was observed in private facilities, where women in the poorest group had approximately four times the odds (aOR 4.04, 95% CI 2.82–5.84, p < 0.001) of distress financing compared to the wealthiest group (Table 3).

Discussion

Overall, we found that public delivery care services in Bangladesh were more equitably utilised by poorer households in both urban and rural settings, particularly for vaginal deliveries. An alarming trend has emerged in tertiary care facilities, which are expected to provide higher-quality services, as well as for caesarean sections, showing that women with higher income disproportionately benefited from public subsidies. Additionally, distress financing remained a significant burden for women with lower income, further exacerbating their economic vulnerabilities in healthcare.

One of our major findings was that public delivery care facilities were primarily used by lower socio-economic groups, suggesting that the public sector’s effort to provide subsidised healthcare services for the poor was meeting its objectives. However, it could also indicate that due to affordability issues, women with low income had limited options other than going to public facilities. BIA showed that the poorest quintile of the society received a significant portion of the public healthcare subsidy, particularly at primary and secondary care facilities, where they benefited most from this support. Despite these positive findings at lower-tier facilities, we observed that a disproportionate share of public subsidies at tertiary care facilities went to women with higher income. Similar phenomenon was also observed in a previous study regarding maternal care utilisation in Bangladesh which reported that benefits utilisation from the primary level of health facilities was concentrated among the pro-poor group, while a pro-rich concentration was observed at higher level facilities.¹⁸ These urban-based facilities provide specialised maternal healthcare, such as complex deliveries and caesarean sections, and draw patients from a broad geographic and socioeconomic range. Women with higher income, who can afford the costs associated with travel and care, often dominate their utilisation. Households with higher incomes also tend to have better access to information about available services and greater mobility, allowing them to benefit from higher-tier healthcare services. These issues could also explain the disproportionate allocation of benefit among pro-rich group in case of caesarean deliveries. This unequal access means women with lower income are more likely to rely on lower-tier public facilities or forgo necessary care, further exacerbating healthcare inequities. To mitigate this imbalance, strengthening referral systems between primary, secondary, and tertiary healthcare facilities would ensure that women get access to the tertiary care facilities based on medical need rather than socioeconomic status. By streamlining referrals and prioritising those requiring advanced care, healthcare resources can be more efficiently allocated, reducing overcrowding at tertiary facilities and ensuring that women with lower income have equal access to specialised services.²¹

A similar trend was also observed among women with higher educational attainment where women with higher income disproportionately received public resources. This pattern reflects the influence of education on healthcare access–wealthier women, typically with higher education, are more informed about available services and better equipped to navigate the healthcare system. This phenomenon might especially pronounce in tertiary care facilities, where women with higher educational levels gain access, yet the benefits disproportionately favour the wealthiest. While educated women may frequent these facilities more often, the real advantages accrue to wealthier individuals rather than the poorest. There is limited evidence regarding the influence of educational attainment on benefit utilisation of public facilities in context of Bangladesh, however, similar finding was reported in India, where richest group of educated women received the highest benefit of public delivery care facilities.²⁰

Another striking observation from the study was the equitable public service utilisation coexisting with higher rates of distress financing among poorer populations. Although the poorest quintile received a larger share of public subsidies for delivery care, they were still more vulnerable to distress financing—borrowing money or selling assets to cover out-of-pocket expenses. Nearly one-third of the women with lowest income resorted to distress financing for their delivery expenses, compared to only ten percent of the women with highest income. Though the poor portion of the society is getting subsidy from the public healthcare facilities, these subsidies are not adequate to alleviate their poverty. Besides, while public healthcare facilities in Bangladesh offer subsidised services, they are not entirely free. Incidental costs like purchasing drugs, travel cost, or making informal payments to healthcare providers can strain poorer households.²² However, this phenomenon might also attributable to the methodology we used to calculate benefit incidence. Since, we relied on the median OOPE at private facilities to estimate benefit incidence, this approach might lead to an overestimation of the subsidies received by the poorest quintile. Importantly, this overestimation might not necessarily reflect their actual vulnerability to distress financing, as it did not account for the financial hardships, they experienced despite receiving public subsidies. Additionally, the high rate of caesarean sections in Bangladesh, not only in private facilities, but also in public ones,²³ likely contributes to higher OOPE and the need for distress financing. The healthcare financing system in Bangladesh, which remains heavily reliant on out-of-pocket payments, exacerbates this issue. While the public sector subsidises many services, households still bear a significant portion of healthcare costs.²⁴ Poorer households, with limited savings and access to formal credit, may be forced to borrow or sell assets to manage these expenses.²⁵

Our study has several policy implications. First, the financial risk protection mechanisms for poorer households should be strengthened. Expanding community health insurance schemes or implementing a national health insurance system tailored to the needs of low-income populations could provide crucial financial security. Second, reducing indirect costs associated with public healthcare services—such as expanding transportation subsidies and ensuring public facilities are adequately stocked with essential drugs and supplies—could significantly alleviate financial burdens to the poorer households. Third, social safety nets, including cash transfers and voucher schemes for maternal healthcare, should be reinforced to mitigate the financial vulnerability of the most disadvantaged women, ensuring equitable access to essential health services.

Our study has several limitations. First, the use of out-of-pocket expenditure (OOPE) data from private facilities as a proxy for public facility costs for child delivery might introduce bias, as public and private healthcare services differ in cost, quality, and availability. Besides, the objective of profit-maximisation and inadequate regulation of the private healthcare sector in Bangladesh might result in an overestimation of public sector service costs compared to actual delivery care costs, resulting in an over-estimation of public subsidy. Second, reliance on self-reported data may have led to recall bias, particularly in reporting healthcare expenditures and informal payments. Third, OOPE as defined in our study might include demand-side expenditures, such as the cost of transport from home to the health facility. Typically, government allocations do not subsidise such expenditures, which may introduce bias in reporting OOPE and consequently affect the estimation of public subsidies. However, since we utilised secondary data, and the survey methodology did not explicitly report whether participants were provided a clear operational definition of OOPE, we had limited ability to minimise this bias, which remains a significant limitation of the study. Fourth, the study may have oversimplified the complex interactions between public and private healthcare providers in Bangladesh, which could influence the distribution of subsidies. Additionally, gender dynamics were not fully explored. Women’s healthcare access is often shaped by social and cultural barriers, including limited decision-making autonomy, mobility restrictions, and the influence of male family members on health-seeking behaviour. These factors likely restricted women’s ability to seek care, especially at higher-level facilities. Besides, we observed an uneven distribution of women across wealth quintiles which is likely because the quintiles were defined based on the overall household population, rather than exclusively among households where a woman delivered in the last two years. As a result, the wealth distribution among the subset of women who recently gave birth might not perfectly align with the overall population distribution. This distinction is important when interpreting the equity patterns in service utilisation and benefit incidence.

While public healthcare services in Bangladesh have shown an equitable utilisation, particularly at the primary care levels, significant disparities remain in access to tertiary care facilities. Our findings highlight that women with higher income disproportionately benefit from public subsidies at higher-tier facilities, while women with lower income rely more on lower-tier public services. Additionally, the poorer socioeconomic groups continue to face a high burden of distress financing. Targeted policy interventions, such as strengthening referral systems, are necessary to ensure equitable access to tertiary care facilities for the poorer portion of the society. Financial risk protection mechanisms, such as community health insurance schemes, are crucial to help the most vulnerable populations access maternal care services without incurring financial hardships.

Contributors

Conceptualisation: MAR, USA; Data curation: MAR; Formal analysis: MAR, USA, MGH; Investigation: MAR, USA, DTA; Methodology: MAR, USA, SNS; Project administration: MAR.

Resources: DTA, MJH; Software: MAR, USA; Supervision: MGH; Validation: SNS; Visualisation: MAR; Writing – original draft: MAR, USA; Writing – review & editing: MAR, USA, DTA, SNS, MJH, MGH. All authors have read and agreed to the final version of the manuscript.

Data sharing statement

The datasets used in this study are available on request from the DHS Program (https://dhsprogram.com/methodology/survey/survey-display-584.cfm).

Declaration of interests

We declare no competing interests.