Sage Journals: Discover world-class research

Abstract

This study examines whether higher maternity cash incentives under Nepal’s Aama Program—intended to reduce household out-of-pocket (OOP) expenditures for delivery care—are associated with completion of the continuum of maternity care across Nepal’s three ecological regions: mountainous, hilly, and Terai. A difference-in-differences (DID) analysis was conducted using data from the Nepal Demographic and Health Surveys 2011 and 2022. The treatment group included mountainous and hilly districts (n = 55), which received higher cash incentives for facility deliveries, while the control group comprised Terai districts (n = 20). The primary outcome was a binary ‘continuum of care’ indicator, coded as 1 if a woman received 4 or more antenatal care (ANC) visits, at least 1 ANC from a skilled provider, and skilled birth attendance (SBA) during delivery, and 0 otherwise. From 2011 to 2022, continuum-of-care coverage increased from 16.2% to 46.3% in treatment districts and from 18.9% to 42.4% in control districts, yielding an unadjusted DID of 6.7 percentage points. The regression DID analysis found that the higher cash incentive in hilly and mountainous regions was associated with a statistically significant 0.7 percentage point increase in the likelihood of completing the continuum of maternity care (P = .014). While statistically significant, the magnitude is modest, indicating incremental improvement rather than a transformational shift in service utilization. Mother’s age, mother’s education, and husband’s education were positively associated with completion, whereas poverty, rural residence, and higher parity were negatively associated (all P < .001). These findings suggest that targeted financial incentives can positively influence maternal health service utilization, but achieving larger improvements will likely require refining the program and addressing non-financial barriers. Limitations include the two-period DID design, reliance on parallel trends, and potential residual confounding.

Keywords

Nepal continuum of maternity care maternal health program evaluation

Introduction

Despite significant progress in recent decades, Nepal’s Maternal Mortality Ratio (MMR) of 142 per 100 000 live births and Neonatal Mortality Rate (NMR) of 17 deaths per 1000 live births remain well above the 2030 Global Sustainable Development Goal (SDG) targets of 70 and 12, respectively.^1
-5 Although MMR declined from 504 in 2000 to 142 in 2023, experts suggest that the rate of decline has slowed since 2015. This delay is attributed to both inadequate utilization of essential health services and gaps in the quality of care.³ Key supply-side challenges include insufficient and unskilled human resources, inadequate infrastructure, inaccessible health facilities, and negative provider attitudes. On the demand side, low service utilization among poorer families, lack of perceived need, and the persistence of unsafe abortions continue to hinder progress. Addressing these issues requires context-specific strategies, such as improving access to comprehensive emergency obstetric care and establishing maternity waiting homes in hilly and mountain regions.⁶

The maternity continuum of care is emerging as a critical strategy in improving maternal, newborn and child health (MNCH) outcomes through integrated service delivery.⁷ This approach spans from pre-pregnancy through childbirth and the postnatal period, delivered across both community and clinical settings. Evidence from low- and middle-income countries (LMICs) shows that consistent uptake of antenatal care, skilled birth attendance, and postnatal follow-up can significantly reduce neonatal and perinatal deaths.⁷ Recommended standards include 8 antenatal visits, delivery by skilled birth attendants, and postnatal care within the first 6 weeks.^8
-10 Antenatal care plays a crucial role in identifying and managing conditions such as anemia, pregnancy-induced hypertension, pre-eclampsia, and infections.⁹ Consistent antenatal visits also strengthen the connection to formal health services, increasing the likelihood of skilled birth attendance and reducing complications such as hemorrhage, eclampsia, obstructed labor, and sepsis.⁹

Nepal’s Aama Program, officially known as Aama Surakshya Karyakram, reflects the continuity-of-care perspective in maternal health policy. Originally launched in 1998 as the Safe Motherhood Policy, the program aimed to reduce household out-of-pocket (OOP) expenditures for delivery care and promote facility-based deliveries to improve maternal and newborn health outcomes.¹⁰ In 2009, the program was restructured to include 2 key components: (1) a universal free delivery service, and (2) the continuation of the Safe Delivery Incentive Program (SDIP), which provides cash payments to women who deliver in health facilities and incentive payments to health workers conducting home deliveries.^10,11 The cash incentives vary by ecological region: 1500 NPR in mountainous areas, 1000 NPR in hilly regions, and 500 NPR in Terai plains.¹² These differences were designed to account for the varying costs and challenges of accessing health facilities across Nepal’s diverse geography. Women in mountainous regions often face longer travel times, limited transportation infrastructure, and fewer nearby facilities, making institutional delivery more difficult and expensive. A 2006 survey found that women took an average of 8.3 h to reach a birthing facility in mountain areas, compared to 5.6 h in hills and 2.8 h in the Terai plains.¹³ The tiered incentive structure was intended to offset these geographic disparities and encourage facility-based deliveries in harder-to-reach areas.^14,15

Impact evaluations of SDIP found a 4.2% increase in skilled birth attendance (SBA) rates.¹⁶ Another study reported a reduction in neonatal mortality by four percentage points among women with fewer than 2 living children, and by 6.9 percentage points among women from lower castes and indigenous groups in treatment districts compared to control districts.¹⁷ However, the program showed stronger effects in districts with higher Human Development Index (HDI), while its impact was limited in low-HDI districts due to barriers such as poor infrastructure, social and cultural barriers, and low program awareness.¹⁸

Disparities in maternal health service utilization persist across Nepal. Women from advantaged ethnic backgrounds or those with access to bank accounts are more likely to receive optimal MNCH services. In contrast, women who speak Maithili language, have high birth order (4 or more children), or live in remote provinces are less likely to receive quality antenatal care.^3,19
-21 These findings underscore the need for policies that specifically target disadvantaged populations and improve service quality in underserved areas.

Comparative evidence from other countries shows mixed results. India’s Janani Suraksha Yojana (JSY) increased facility deliveries but did not significantly reduce maternal mortality, partly because many deliveries occurred in lower-level facilities unequipped for emergencies.^22,23 In contrast, a randomized controlled trial in Nigeria found that conditional cash transfers (CCTs) led to improvements in delivery care quality and satisfaction.²⁴ Bangladesh’s Maternal Health Voucher Scheme (MHVS) also showed substantial increases in facility deliveries, particularly after 2 years of implementation.²⁵

While previous studies have examined the impact of Nepal’s Safe Delivery Incentive Program (SDIP) and the Aama Program on individual maternal health indicators—such as antenatal care, place of delivery, and postnatal care—this study contributes new insights by modeling the aggregated maternity continuum of care as a binary outcome.^3,17,18,26 This approach allows us to assess not just isolated service use, but the extent to which women receive a full sequence of recommended maternal health services, which is critical for improving outcomes. Although difference-in-differences (DID) methods have been applied in earlier evaluations, this study is among the first to extend the analysis through 2022, leveraging the most recent Nepal Demographic and Health Survey (NDHS) data. By exploring variation in CCT amounts across ecological regions, we provide a more nuanced understanding of how financial incentives interact with contextual barriers to influence continuity of care. A quasi-experimental DID design is appropriate because SDIP incentive levels were set differently across ecological regions, creating a natural policy contrast for examining changes over time.⁵⁷

This study aims to estimate the long-term association between differential SDIP cash incentives and completion of the maternity continuum of care in Nepal, using DID approach with NDHS 2011 and 2022.

Methods

Data

The data utilized for this research paper originated from the Nepal Demographic and Health Surveys conducted in 2011 and 2022. The DHS is a nationally representative household survey that collects data on health, nutrition, empowerment, and other key indicators. These cross-sectional surveys occur approximately every 5 years using consistent sampling methodologies to enable comparisons over time. Each survey employs a stratified two-stage cluster sampling technique to select women of reproductive age (15-49 years). For this analysis, we restricted the sample to women aged 15 to 49 who reported at least 1 live birth in the 5 years preceding each survey.¹¹ This inclusion criterion ensures relevance to maternal health outcomes and aligns with the continuum of care indicators under study. After applying this filter, the final analytic sample included approximately 12 674 women from the 2011 NDHS and 14 845 women from the 2022 NDHS.

The 2011 NDHS was the fourth DHS conducted in Nepal (following 1996, 2001, and 2006), and the 2022 NDHS was the sixth DHS (following 2011 and 2016).^13,29 The survey provides estimates at the national and provincial levels, as well as for urban and rural areas, and the 3 ecological regions of Nepal: mountain, hill, and Terai.¹¹ Data from NDHS 2011 and 2022 were used to conduct a differences-in-differences (DID) analysis to assess the impact of the Aama Program over time. DID estimates the effect of a specific intervention (in this case, the maternity cash incentive component of the Aama Program) by comparing outcome changes over time between an intervention group and a control group. The mountainous and hilly districts were designated as the treatment group, while the Terai region served as the control group. This classification was based on the differential cash transfer amounts provided through the Aama Program, which were higher in mountain and hill regions to account for greater travel and access challenges.

While the program aimed to equalize access by adjusting cash incentives, comparing maternal health outcomes across these regions remains meaningful. If the adjustment was successful, we would expect similar improvements across regions. However, persistent disparities in infrastructure, health service availability, and sociocultural factors may have influenced the program’s effectiveness differently across ecological regions. Thus, comparing these groups allows us to assess whether the Aama Program’s design adequately addressed contextual barriers to care.

Measures

The primary outcome of interest in this study is the continuum of care for maternal health services. The continuum of care indicator used in this study—comprising 4 or more ANC visits, at least 1 ANC visit with a skilled provider, and skilled birth attendance—has been employed in prior research assessing maternal health service quality and outcomes.^27,28 The continuum of care was coded as a binary variable (0 = no continuity, 1 = continuity in care).

The requirement of 4 or more ANC visits reflects the previous World Health Organization (WHO) recommendation that all pregnant women should attend a minimum of 4 focused ANC check-ups with a skilled health worker. Although WHO and the Government of Nepal have recommended 8 ANC contacts since 2016, our analysis uses the earlier four-visit standard to ensure comparability between the 2011 to 2022 NDHS. ANC with a skilled provider refers to pregnancy care received from skilled health personnel, such as doctors, nurses, or auxiliary nurse midwives, as defined by NDHS. Skilled birth attendance (SBA) is defined as births delivered with the assistance of doctors and nurses/auxiliary nurse midwives who have received training as skilled birth attendants, consistent with Government of Nepal guidelines.¹³ Because women were asked about these services only for births occurring within the 5 years preceding the survey interview date, we restricted the analysis to responses from that time frame.

Covariates

Synthesizing findings from diverse studies grounded in the Andersen-Newman Behavioral Model (ANBM) of healthcare utilization guided the selection of predictor and outcome variables for this analysis. The ANBM categorizes factors influencing healthcare utilization into 3 domains: predisposing, enabling, and need factors. While this study does not aim to test the ANBM directly, the framework informed our variable selection and helped conceptualize the pathways through which maternal healthcare utilization may be shaped. A visual representation of this framework is provided in Figure S1, included in Supporting Information.

The covariates included in this study were selected to reflect these domains and to address specific questions about barriers and facilitators to maternal healthcare utilization within the Aama Program in Nepal. Predisposing factors included woman’s age, education level, ethnicity, and religion.^30,31,33 Age and education were included as continuous variables in the analysis. Religion was dichotomized into ‘Hindu’ and ‘Non-Hindu’, with the latter category combining Buddhists, Muslims, Kirat, and Christians. Enabling factors encompassed income, residence location, partner’s education level, and land ownership status.^31,32 Place of residence was categorized as rural or urban. The household wealth index was dichotomized into ‘Poor’ and ‘Non-poor’, where ‘Poor’ combined the 2 lowest wealth quintiles, and the ‘Non-poor’ included the middle, fourth, and richest wealth quintiles. Partner’s education was included as a continuous variable.

The decision to dichotomize certain categorical variables, including religion and wealth index, was made to simplify analyses and enhance interpretability. However, we acknowledge that collapsing categories may obscure important within-group differences. Parity was also included as a covariate and categorized into 4 groups: none, 1 to 2, 3 to 4, and more than 4 children. For all covariates, missing observations and responses coded as ‘don’t know’ were excluded from the analysis.

Initially, we planned to include ethnicity (a predisposing factor) and women’s land ownership (an enabling factor) as covariates. However, during the data exploration phase, we identified a significant correlation between ethnicity and treatment assignment (control vs treatment). Including ethnicity could introduce multicollinearity and confound the interpretation of treatment effects, so it was excluded from the final analysis. Similarly, land ownership was found to correlate with the wealth index. After careful consideration, we retained the wealth index due to its direct relevance to our research question and its unique contribution to understanding the outcome variable (ie, continuum of care) and excluded land ownership to avoid multicollinearity.

Study Districts and Program Period of Analysis

Although Nepal officially has 77 districts under the federal structure established by the 2015 Constitution, the administrative context differed between the 2 DHS surveys. At the time of NDHS 2011, the country was organized into 75 districts; after federalization, 2 of those districts (Nawalparasi and Rukum) were each split into 2 administrative units, resulting in a total of 77 districts by the time of NDHS 2022.³⁵ The NDHS 2022 sampling frame, therefore, included 77 districts, whereas the NDHS 2011 covered 75 districts. For comparability across survey years, we classify districts into the same 3 ecological regions used in the pre-federal structure: mountain (n = 16), hill (n = 39), and Terai (southern plains; n = 20). As illustrated in the timeline of the Aama Program (Figure 1), Nepal initially implemented the Safe Delivery Incentive Program (SDIP), which provided cash transfers to women who gave birth in health facilities. Women in districts with high HDI received standard cash transfers, while those in 25 low-HDI mountain and hill districts received larger cash transfer along with free delivery services.^36,37 Eligibility was initially limited to women with no more than 2 children and/or those visiting a facility due to obstetric complications.

Figure 1.

Timeline of the Aama Program and data sources.

To account for transportation costs, the cash transfer amounts varied by ecological region: women in mountain districts received 1500 NPR, those in hill districts received 1000 NPR, and those in Terai districts received 500 NPR. In November 2007, eligibility was expanded to include all women regardless of parity. In 2009, SDIP was renamed the Aama Program, and the free delivery care component was extended nationwide.

This background informed our decision to use NDHS 2011 as our baseline and NDHS 2022 as the endline to assess the long-term impact of the Aama Program on the continuum of care. In line with our analysis syntax, we restricted both surveys to births occurring in the 5 years preceding each survey. Therefore, NDHS 2011 captures births from approximately 2007 to 2011, with a midpoint around 2009—coinciding with the national rollout of the Aama Program, while NDHS 2022 captures births from approximately 2018 to 2022. Although certain indicators in NDHS 2022 employ shorter reference periods in standard tabulations, we applied a consistent 5-year window across both surveys to ensure comparability in our analysis.^13,29 These timeframes allow us to compare maternal health service utilization before and well after the program’s expansion. Due to the relatively small sample size in mountain districts, these were combined with hill districts to form the treatment group, while Terai districts served as the control group.

Analytical Approach

We used a regression model to estimate the differential effects on the continuum of care between treatment and control districts:

Y_{i j t} = β_{0} + β_{1} P_{j t} + β_{2} T_{t} + β_{3} P_{j t} . T_{t} + β_{4} X_{i j t} + ε_{i j t}

where $Y$ = Outcome variable (continuum of care) for individual $i$ living in district $j$ at time $t$

$P$ = Treatment group (0 = control, 1 = treatment)

$T$ = Time dummies (0 = 2011, 1 = 2022)

$β_{0}$ = Intercept term, representing the average outcome in the control districts before treatment

$β_{1}$ = Average treatment effect in control districts before the treatment

$β_{2}$ = Average effect of time (trend) on the control districts

$β_{3}$ = DID estimate, representing the average treatment effect on the treated relative to control districts

$X$ = Vector of socio-demographic covariates (age, wealth, education, etc.)

$ε$ = the error term.

Statistical Analysis

We used StataSE 18 software to run our analysis.³⁸ For the descriptive analyses, weighted percentages were used. The analysis of demographic, cultural, and socioeconomic characteristics was discussed in the Covariates section above, along with data on our outcome variable (continuum of care) and the variables used to form the continuum of care. As described in the previous section, DID was estimated using a regression model to assess the difference in the continuum of care between the treatment and control regions, both before and after treatment, while controlling for potential explanatory variables. We model the binary outcome using a linear probability model (LPM) estimated by ordinary least squares with robust standard errors and DHS survey weights for the two-stage stratified cluster design in Nepal. Coefficients are directly interpretable as percentage-point differences in predicted probability, matching our interest in differences in the continuum of care coverage between treatment and control districts and easing communication to policy audiences compared with log-odds from logistic regression. Although the LPM allows predictions outside the [0,1] range and implies heteroskedastic errors, these limitations are addressed by focusing on effects within the observed covariate range and using heteroskedasticity- and cluster-robust inference, a strategy widely applied in large-scale survey analyses.^39
-43

This study is reported in accordance with the STROBE guidelines for observational studies (see S2 Checklist included in Supporting Information).

Results

Descriptive Results

Table 1 presents women’s background characteristics and maternal health indicators in treatment and control districts at baseline (2011) and endline (2022). Across both regions, there is a notable improvement in education levels over time. Mothers and their husbands in treatment districts consistently show higher mean years of education compared to those in control districts, with the gap widening by endline.

Table 1.

Descriptive Statistics for All Women in 2011 and 2022 in Treatment and Control Districts.^a,b,c

Background characteristics	Baseline (2011)		Endline (2022)
	(N = 12 674)		(N = 14 845)
	Treatment districts	Control districts	Treatment districts	Control districts
Mother’s education (mean)	4.55 (0.15)	4.27 (0.24)	7.05 (0.13)	5.73 (0.15)
Husband’s education (mean)	6.07 (0.15)	5.66 (0.22)	7.68 (0.11)	6.65 (0.13)
Respondent’s age (mean)	29.06 (0.15)	28.57 (0.16)	30.11 (0.14)	29.49 (0.11)
Religion
Non-Hindu	16.91	14.83	20.08	13.85
Hindu	83.09	85.17	79.92	86.15
Wealth Index
Poor	50.42	22.72	48.89	27.22
Middle-class/rich	49.58	77.28	51.11	72.78
Residence
Urban	16.09	12.84	63.40	72.77
Rural	83.91	87.16	36.60	27.23
Parity
None	30.21	29.35	30.00	27.55
1-2	3.27	33.62	44.59	41.43
3-4	23.85	25.92	19.65	25.16
>4	13.24	11.11	5.76	5.87
Outcome variables
ANC visits
<4 ANC visits	50.84	49.06	12.27	22.31
≥4 ANC visits	49.16	50.94	87.73	77.69
Skilled birth attendance
Yes	36.6	43.43	68.79	29.67
No	63.40	56.57	31.21	70.33
Skilled ANC
Yes	36.60	43.43	68.79	70.33
No	63.40	56.57	31.21	29.67
Continuum of care
Yes	16.16	18.93	46.29	42.41
No	83.84	81.07	53.71	57.59

Treatment districts refer to hilly and mountainous districts, and Control districts refer to Terai districts.

Standard errors in parentheses.

Estimations are sample-weighted percentage point distributions.

Additionally, there is a consistent increase in respondent age in both treatment and control regions over the study period, with slightly higher average ages observed in the treatment region. In terms of religious affiliation, there are more Hindu participants in both regions; however, the proportion of Hindu respondents declined slightly in treatment districts while increasing in control districts. Wealth distribution shifted over time, with treatment districts experiencing a modest increase in middle-class/richer households. This trend is accompanied by a significant rise in urban residence, particularly in treatment districts, suggesting broader socio-economic changes.

Changes in demographic characteristics are mirrored by improvements in maternal health outcomes. Treatment districts show substantial gains in the proportion of women receiving 4 or more ANC visits, skilled ANC, and skilled birth attendance. The continuum of care indicator also improved markedly in treatment districts, rising from 16.2% to 46.3%. In contrast, control districts show more modest improvements or even declines in some indicators, such as skilled birth attendance.

Table 2 and Figure 2 present changes in the continuum of maternity care between 2011 and 2022 for treatment and control districts. At baseline (2011), 16.2% of women in treatment districts received the full continuum of care, compared to 18.9% in control districts. By endline (2022), both groups showed substantial improvement, with 46.3% in treatment districts and 42.4% in control districts. The absolute increase over time was greater in the treatment group (30.1 percentage points) than in the control group (23.5 percentage points). The DID estimate—calculated as the change in treatment minus the change in control—was 6.7 percentage points, suggesting a positive impact of the higher cash incentives provided in treatment districts. Figure 2 visually illustrates these trends, showing a steeper upward trajectory for the treatment group compared to the control group.

Table 2.

Comparison of Continuum of Care Coverage Between Treatment and Control Districts at Baseline and Endline.

Districts	Baseline (2011)	Endline (2022)	Difference (2022-2011)
Treatment	16.2%	46.3%	30.1 Percentage points
Control	18.9%	42.4%	23.5 Percentage points
Difference (treatment − control)	−2.8 Percentage points	3.9 Percentage points	6.7 Percentage points

Figure 2.

Trends in the continuum of care in the control and treatment districts between 2011 and 2022.

DID Estimates

Table 3 presents the regression results from the DID estimation with covariates, examining factors associated with the continuum of care outcome. The coefficient for the treatment variable is −13.65 (P = .014), indicating that, at baseline, women in treatment districts had a significantly lower prevalence of receiving full continuum of care compared to those in control districts. The interaction term between treatment and time is positive and statistically significant (coefficient = 0.007, P = .014), suggesting a 0.7 percentage point increase in the likelihood of receiving continuum of care in treatment districts relative to control districts between 2011 and 2022, after adjusting for covariates. In Table 2, the DID estimate was 6.7 percentage points because it is based on a simple 2 × 2 comparison and does not adjust for differences in participant characteristics between regions or changes in those characteristics over time. In the regression DID model, we adjusted for measured sociodemographic and other relevant factors, and the DID coefficient was 0.7 percentage points. This suggests that much of the unadjusted difference was due to baseline and time-varying differences in these factors, rather than the maternity cash incentive differential itself.⁴⁴ The round variable, representing time, is also significant (coefficient = 0.009, P < .001), reflecting an overall increase in continuum of care coverage over time across both groups.

Table 3.

Regression Results: DID Estimation with Covariates.^a

Continuum of care	Coefficient	Standard error	t-Value	P-value	95% CI
Treatment	−13.65	5.532	−2.47	.014	−24.514	−2.792
Round	0.009	0.002	4.74	<.001	0.006	0.014
Treatment × time	0.007	0.003	2.47	.014	0.001	0.012
Mother’s age	0.005	0.001	4.39	<.001	0.003	0.007
Mother’s education level	0.017	0.002	8.92	<.001	0.013	0.021
Husband’s education level	0.006	0.002	3.43	.001	0.003	0.009
Religion	0.029	0.016	1.84	.067	−0.002	0.060
Wealth Index	−0.101	0.015	−6.86	<.001	−0.131	−0.072
Place of residence	−0.123	0.017	−7.06	<.001	−0.158	−0.089
Parity
3-4	−0.061	0.014	−4.43	<.001	−0.088	−0.034
>4	−0.092	0.021	−4.48	<.001	−0.133	−0.052
Constant	−19.40	4.127	−4.70	<.001	−27.501	−11.298

The R-squared value for this model was .187; F-statistic = 121.61; P-value associated with the F-statistic = < .00001.

Several covariates were significantly associated with the outcome. Mother’s age, mother’s education, and husband’s education were all positively associated with receiving continuum of care. Being classified as ‘poor’ was associated with a 10.1 percentage point decrease in the likelihood of receiving continuum of care (P < .001), and living in a rural area was associated with a 12.3 percentage point decrease (P < .001). Higher parity was also negatively associated with the outcome: women with 3 to 4 children had a 6.1 percentage point lower likelihood, and those with more than 4 children had a 9.2 percentage point lower likelihood of receiving continuum of care (both P < .001). Although being Hindu was associated with a higher likelihood of receiving continuum of care, this association was not statistically significant (P = .067).

The model has an R-squared value of .187, indicating that approximately 18.7% of the variation in continuum of care is explained by the included covariates. The F-statistic is significant (P < .001), confirming the overall model fit.

Discussion

The study found a small but statistically significant effect of the higher cash incentives provided to the hilly and mountainous (treatment) regions under Nepal’s Aama Program on the completion of the continuum of maternity care. While modest, this effect suggests that financial incentives may play a role in improving maternal health service utilization. Importantly, our findings pertain to the maternity incentive component of the Aama Program (ie, differential cash payments by ecological region), rather than the full Aama package of interventions. These findings build upon existing research examining the Aama Program and related Safe Motherhood initiatives in Nepal. Previous studies have evaluated the effects of user fee removal and comprehensive maternal health programs on delivery practices and facility utilization. For example, 1 study found a significant positive impact of user fee removal on deliveries by skilled birth attendants and increased use of public health facilities.¹⁹ Similarly, an impact evaluation of SDIP in Nepal highlighted the role of financial incentives in promoting facility-based deliveries.¹⁶ Another study examined the cash transfer component of SDIP, revealing nuanced findings about women’s responsiveness to program eligibility based on socioeconomic status and the district-level HDI.¹⁸ Evidence from longer-term reviews of Nepal’s maternal financing policies also suggests that demand-side incentives can increase service utilization, but that gains are often constrained by geography, service readiness, and differential access across ecological regions.¹⁵

Despite improvements over time, our findings also underscore the persistent issue of low completion rates of the continuum of maternity care in Nepal. In 2011, only 16.2% of women in treatment districts and 18.9% in control districts completed the full continuum of care. By 2022, these rates increased to 46.3% and 42.4%, respectively. While this progress is encouraging, it remains concerning that half of the women received comprehensive maternal health services. These results are consistent with other studies, including 1 based on the Nepal Multiple Indicator Cluster Survey, which found that only 41% of women received antenatal care, skilled birth attendant delivery, and postnatal care during their most recent birth.³⁴ Recent analyses on NDHS 2022 similarly report completion around one-half nationally, with larger drop-offs around postnatal care (PNC) and persistent socioeconomic and geographic inequities.⁴⁵ This is particularly alarming given the critical role of the continuum of care in preventing maternal morbidity and mortality. As discussed earlier, ANC visits are essential for identifying and managing pregnancy-related complications and fostering continued engagement with the health system, which increases the likelihood of skilled attendance at birth and reduces the risk of adverse outcomes.^7
-9

There are several plausible reasons the observed association is modest despite statistically significant incentive differentials by ecological region. First, service readiness and quality constraints limit the returns to increased demand: recent facility assessments indicate low or stagnant functionality of emergency obstetric and newborn care (EmONC) signal functions between 2015 and 2021, and health-facility survey analyses identify shortfalls in staffing, competencies, and respectful care that shape women’s satisfaction and willingness to return.^12,46 Second, operational bottlenecks within Aama, including administrative complexity and reimbursement lags, have been documented in program reviews and early facility monitoring.⁴⁷ These reimbursement challenges can affect both clients (eg, delayed or incomplete payments) and providers (eg, delayed facility reimbursement for free delivery care), potentially weakening trust in the program and disrupting continuity across ANC–delivery–PNC. Third, procurement delays and stock-outs of essential medicines, especially at peripheral levels, erode both perceived and actual value of repeated ANC and timely PNC, thereby reducing completion.⁴⁸ Because Aama Program also includes free delivery services (and associated reimbursements to facilities), the continuity of care is likely influenced not only by the cash incentive but also by whether public facilities can reliably provide the ‘free’ package—including essential medicines and supplies—without informal payments or repeated stock-outs.⁴⁹

Policy Implications

The findings carry significant policy implications. The observed increase in continuum of care completion in regions receiving higher cash incentives suggests that the government may consider revising the incentive design; however, blanket increases in the Terai region alone are unlikely to yield proportionate gains unless paired with supply-side strengthening. Comparative analyses indicate that demand-side financing can increase service utilization, meaningful improvements in quality and continuity typically require addressing structural constraints.⁵⁰ If policymakers wish to test higher Terai incentives, a more pragmatic approach would be to implement time-bound pilots coupled with co-investments in facility readiness—including assured availability of essential medicines, adequate staffing, respectful care training, and explicit evaluation of continuity outcomes rather than delivery alone.^48,51

Additionally, factors such as socioeconomic status, urban residence, media access, and lower parity have been associated with higher engagement in the continuum of care.³⁵ Targeting marginalized populations and tailoring Aama Program services to address these disparities could enhance program effectiveness. Furthermore, ensuring that health facilities are adequately staffed and equipped to provide quality ANC and delivery services is essential. Strengthening infrastructure and service delivery may encourage women to return for follow-up care and foster trust in the health system.

Because the Aama Program operates as a cash transfer scheme, questions about how to sustainably finance maternity incentives will remain important. With Nepal’s transition to federalization, sub-national governments now play an important role in planning and financing basic health services.^52,53 In this context, future adjustments to maternity incentives and supply-side investments may be most sustainable if they are jointly financed by provinces and municipalities, and guided by clear service standards and equitable resource allocation mechanisms. Given budget uncertainty and competing priorities, embedding Aama-related reimbursements and essential-medicine budgets into more predictable sub-national budgets and public reporting on stock-out rates may reduce operational uncertainty and support sustained improvement.^52,53

Strengths and Limitations

This study contributes to the literature by evaluating the long-term impact of the Aama Program using 2 recent nationally representative datasets (NDHS 2011 and 2022). To our knowledge, this is the first impact evaluation of the program incorporating NDHS 2022 data. Additionally, our use of a composite outcome variable —defined as receiving 4 or more ANC visits, ANC from a skilled provider, and skilled birth attendance—offers a novel approach to assessing the continuum of care in the context of maternal health policy evaluation.

To ensure robust estimation using regionally representative data, we assessed key assumptions underlying the DID approach. The common support and stable unit treatment value assumptions are likely to hold given the targeted nature of the Aama Program and the overlap in covariate distributions between treatment and control groups.^54,55 Additionally, there were no interference or spillover effects from the treatment.³⁹ We tested the exogeneity assumption by examining baseline covariate balance and adjusting for imbalances in the regression model.^56,57 The parallel trends assumption was evaluated by comparing pre-intervention trends in ANC and skilled birth attendance between 1996 and 2006 (see Figure S3 in Supporting Information).³⁹

A key limitation is that our continuum of care measure uses the former WHO standard of 4 or more ANC visits. Although the WHO and the Government of Nepal have recommended 8 ANC contacts since 2016, we applied the four-visit threshold to maintain comparability across NDHS 2011 and NDHS 2022. As a result, our estimates may understate gaps relative to the current eight-contact benchmark and may not fully capture improvements concentrated in later ANC visits. While combining hilly and mountainous districts into a single treatment group improved statistical power, it may have masked differences in program impact between these regions, especially given the variation in cash incentive amounts (NPR 1500 for mountain, NPR 1000 for hill, and NPR 500 for Terai).¹² Additionally, dichotomizing certain covariates simplified the analysis but may have reduced statistical power and obscured important within-group variation. As this study used secondary NDHS data, the sample size was determined by the survey design rather than by an a priori power calculation. However, because the NDHS is representative at the provincial level and our analyses were conducted at the ecological-region level (a higher level of aggregation), the risk of low statistical power due to small sample sizes is expected to be minimal. Finally, inference is constrained by the two-period DID design and its reliance on the parallel trends assumption, as well as potential residual confounding from unmeasured regional changes over time.

Conclusion

This study provides valuable insights into demographic trends, maternal healthcare utilization, and their relationship with the continuum of care in Nepal. Descriptive analyses revealed notable shifts in education levels, age distribution, religious affiliation, wealth index, and urbanization—particularly within the treatment regions. Over time, both treatment and control regions experienced significant improvements in the continuum of care indicators, with the treatment group showing greater progress.

Regression results further highlighted that covariates such as maternal and paternal education, wealth status, place of residence, and parity were significantly associated with the likelihood of completing the continuum of care. While the negative coefficient for the treatment variable reflects lower baseline levels of care in treatment regions, the positive and statistically significant interaction term indicates a modest, incremental improvement over time. In this DID analysis, higher cash payments under the maternity incentive component of the Aama Program were associated with a higher likelihood of completing the full continuum of maternity care in hilly and mountainous regions compared with the Terai. However, the magnitude of this association was modest, and substantial gains in continuity of maternity care likely require more than financial incentives alone—such as improvements in service availability and quality, transportation and geographic access, and demand-side barriers related to information and social norms. Accordingly, our findings should be interpreted as evidence regarding the maternity incentive component of the Aama Program, not the program as a whole.

Overall, the results suggest that while maternity cash incentives can support incremental improvements in service uptake, the Government of Nepal should prioritize integrated approaches that pair demand-side incentives with supply-side investments in service readiness, quality of care, and equitable access, particularly in underserved regions.

Future studies should explore the effects of the Aama Program under the updated eight-ANC-contact guideline and examine how program performance varies with measures of facility readiness, quality of care, and timeliness of cash disbursement. Additionally, mixed-methods research could provide deeper insight into women’s experiences with the incentive process, informal costs, and social or cultural determinants that influence completion of the continuum of care.

Supplemental Material

sj-docx-1-inq-10.1177_00469580261441754 – Supplemental material for Examining the Effect of Maternity Cash Incentives Under Nepal’s Aama Program on the Continuum of Maternity Care: A Difference-in-Differences Analysis of NDHS 2011 to 2022

Supplemental material, sj-docx-1-inq-10.1177_00469580261441754 for Examining the Effect of Maternity Cash Incentives Under Nepal’s Aama Program on the Continuum of Maternity Care: A Difference-in-Differences Analysis of NDHS 2011 to 2022 by Poonam Rawat, Jon M. Hussey and Karar Zunaid Ahsan in INQUIRY: The Journal of Health Care Organization, Provision, and Financing

Footnotes

Acknowledgements

The authors would like to thank the Department of Maternal and Child Health at UNC Gillings School of Global Public Health for their support and guidance during the development of this research paper. We acknowledge the institutions and individuals who made the Nepal Demographic and Health Surveys (NDHS) in 2011 and 2022 possible. The NDHSs were implemented by New ERA under the aegis of the Ministry of Health and Population, Nepal. Funding for the NDHSs was provided by the United States Agency for International Development (USAID), and ICF provided technical assistance through The DHS Program, a USAID-funded project supporting population and health surveys worldwide. Finally, we extend our gratitude to the participants of the NDHSs for their valuable contributions to this study.

ORCID iD

Karar Zunaid Ahsan

Ethical Considerations

Demographic and Health Survey Data collection procedures for the 2011 and 2022 NDHS were reviewed and approved by the ICF Institutional Review Board (IRB), MD, USA, and the Nepal Health Research Council (NHRC), Kathmandu, Nepal. Informed, written consent was obtained from the household head to conduct the interviews, and separately before obtaining biomarker and anthropometric measurements. Respondents who did not provide consent were excluded from the final dataset.

Author Contributions

JMH provided overall guidance for the analyses. KZA conceived and designed the study. PR conducted the literature review and data analyses, with support from KZA. PR drafted the initial manuscript, which was further developed by JMH and KZA. All authors critically reviewed and revised the manuscript. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. All authors read and approved the final manuscript.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The survey data reported in this manuscript are publicly available via the DHS Program. The NDHS 2011 and 2022 datasets are freely available from the following website: .

Supplemental Material

Supplemental material for this article is available online.

References

Amouzou

Leslie

Ram

, et al. Advances in the measurement of coverage for RMNCH and nutrition: from contact to effective coverage. BMJ Glob Health. 2019;4(Suppl 4):e001297. doi:10.1136/bmjgh-2018-001297

Barros

Ronsmans

Axelson

, et al. Equity in maternal, newborn, and child health interventions in countdown to 2015: a retrospective review of survey data from 54 countries. Lancet. 2012;379(9822):1225-1233. doi:10.1016/S0140-6736(12)60113-5

Khatri

Durham

Karkee

Assefa

High coverage but low quality of maternal and newborn health services in the coverage cascade: who is benefitted and left behind in accessing better quality health services in Nepal?

Reprod Health. 2022;19(1):163. doi:10.1186/s12978-022-01465-z

World Health Organization (WHO). Maternal mortality ratio (per 100,000 live births). Data retrieved from World Bank Gender Data Portal. 2023. Accessed June 1, 2024. https://genderdata.worldbank.org/en/indicator/sh-sta-mmrt

United Nations Children’s Fund (UNICEF). UNICEF Data warehouse. 2023. Accessed September 24, 2025. https://data.unicef.org/resources/data_explorer/unicef_f/

Karkee

Tumbahanghe

Morgan

Maharjan

Budhathoki

Manandhar

DS.

Policies and actions to reduce maternal mortality in Nepal: perspectives of key informants. Sex Reprod Health Matters. 2021;29(2):1907026. doi:10.1080/26410397.2021.1907026

Kikuchi

Ansah

Okawa

, et al. Effective linkages of continuum of care for improving neonatal, perinatal, and maternal mortality: a systematic review and meta-analysis. PLoS One. 2015;10(9):e0139288. doi:10.1371/journal.pone.0139288

Shibanuma

Ansah

Kikuchi

, et al. Evaluation of a package of continuum of care interventions for improved maternal, newborn, and child health outcomes and service coverage in Ghana: a cluster-randomized trial. PLoS Med. 2021;18(6):e1003663. doi:10.1371/journal.pmed.1003663

Greenfield

Lynch

Maharjan

, et al. Antenatal care in Nepal: a qualitative study into missed opportunities in the first trimester. AJOG Glob Rep. 2022;2(4):100127. doi:10.1016/j.xagr.2022.100127

10.

World Health Organization. WHO Recommendations on Antenatal Care for a Positive Pregnancy Experience. World Health Organization; 2016.

11.

Aksünger

De Sanctis

Waiyaiya

van Doeveren

van der Graaf

Janssens

What prevents pregnant women from adhering to the continuum of maternal care? Evidence on interrelated mechanisms from a cohort study in Kenya. BMJ Open. 2022;12(1):e050670. doi:10.1136/bmjopen-2021-050670

12.

Witter

Khadka

Nath

Tiwari

The national free delivery policy in Nepal: early evidence of its effects on health facilities. Health Policy Plan. 2011;26 Suppl 2(suppl_2):ii84-ii91. doi:10.1093/heapol/czr066

13.

Ministry of Health and Population [Nepal], New ERA, and ICF. Nepal demographic and health survey 2022. Published online June 15, 2023. Accessed June 1, 2024. https://dhsprogram.com/publications/publication-FR379-DHS-Final-Reports.cfm

14.

Nepal Department of Health Services Annual. Report 2016-2017 | IHME|GHDx. 2018. Accessed June 1, 2024. https://ghdx.healthdata.org/record/nepal-department-health-services-annual-report-2016-2017

15.

Ensor

Bhatt

Tiwari

Incentivizing universal safe delivery in Nepal: 10 years of experience. Health Policy Plan. 2017;32(8):1185-1192. doi:10.1093/heapol/czx070

16.

Khanal

GN.

Conditional cash transfer policies in maternal health service utilization in Nepal: analysis of safe delivery incentive program (Aama Surakshya Karyakram) using Kingdon’s multiple streams framework. Int J Health Plann Manage. 2019;34(1):e131-e141. doi:10.1002/hpm.2691

17.

Zhang

Liu

Hutton

Koirala

Geographical dynamics of poverty in Nepal between 2005 and 2011: where and how?

Sustainability. 2018;10(6):2055. doi:10.3390/su10062055

18.

Powell-Jackson

Hanson

Financial incentives for maternal health: impact of a national programme in Nepal. J Health Econ. 2012;31(1):271-284. doi:10.1016/j.jhealeco.2011.10.010

19.

Lamichhane

Sharma

Mahal

Impact evaluation of free delivery care on maternal health service utilisation and neonatal health in Nepal. Health Policy Plan. 2017;32(10):1427-1436. doi:10.1093/heapol/czx124

20.

Tiwari

Financial incentives and delivery care: evidence from the Safe Delivery Incentive Program in Nepal. Health Econ. 2023;32(10):2372-2389. doi:10.1002/hec.4732

21.

Sapkota

Gyawali

Thapa

Inequalities in antenatal care service utilization in Nepal: evidence from nationally representative Nepal multiple indictor cluster surveys 2014 and 2019. BMC Public Health. 2025;25(1):1727. doi:10.1186/s12889-025-22897-9

22.

Shreezal

Adhikari

Decomposing inequality in maternal and Child Health (MCH) services in Nepal. BMC Public Health. 2023;23(1):995. doi:10.1186/s12889-023-15906-2

23.

Huang-fu

Wang

Karmacharya

, et al. Spatial profiling of geographical accessibility to maternal healthcare and coverage of maternal health service utilisation in Nepal: a geospatial analysis based on demographic and health survey. BMJ Glob Health. 2025;10(2):e017229. doi:10.1136/bmjgh-2024-017229

24.

Lim

Dandona

Hoisington

James

Hogan

Gakidou

India’s Janani Suraksha Yojana, a conditional cash transfer programme to increase births in health facilities: an impact evaluation. Lancet. 2010;375(9730):2009-2023. doi:10.1016/S0140-6736(10)60744-1

25.

Randive

Diwan

De Costa

India’s Conditional Cash Transfer Programme (the JSY) to promote institutional birth: is there an association between institutional birth proportion and maternal mortality?

PLoS One. 2013;8(6):e67452. doi:10.1371/journal.pone.0067452

26.

Okeke

Wagner

Abubakar

IS.

Maternal cash transfers led to increases in facility deliveries and improved quality of delivery care in Nigeria. Health Aff. 2020;39(6):1051-1059. doi:10.1377/hlthaff.2019.00893

27.

Nandi

Charters

Quamruzzaman

, et al. Health care services use, stillbirth, and neonatal and infant survival following implementation of the Maternal Health Voucher Scheme in Bangladesh: a difference-in-differences analysis of Bangladesh demographic and Health Survey data, 2000 to 2016. PLoS Med. 2022;19(8):e1004022. doi:10.1371/journal.pmed.1004022

28.

Sharma

Teijlingen

Belizán

Hundley

Simkhada

Sicuri

Measuring what works: an impact evaluation of women’s groups on maternal health uptake in rural Nepal. PLoS One. 2016;11(5):e0155144. doi:10.1371/journal.pone.0155144

29.

UNFPA Nepal. Nepal Demographic and Health Survey (NDHS). 2011. Accessed January 14, 2026. https://nepal.unfpa.org/en/publications/nepal-demographic-and-health-survey-ndhs-2011

30.

Tiyare

Dimore

Marine

BT.

Assessing antenatal care service satisfaction and associated factors among pregnant women at health facilities in East Shewa Zone, Ethiopia: a facility-based cross-sectional study. Discov Med. 2025;2(1):122. doi:10.1007/s44337-025-00325-6

31.

Pavalagantharajah

Negrin

Bouzanis

, et al. Facility-based maternal quality of care frameworks: a systematic review and best fit framework analysis. Matern Child Health J. 2023;27(10):1742-1753. doi:10.1007/s10995-023-03702-8

32.

Alkhawaldeh

ALBashtawy

Rayan

, et al. Application and use of Andersen’s behavioral model as theoretical framework: a systematic literature review from 2012-2021. Iran J Public Health. 2023;52:1346-1354. doi:10.18502/ijph.v52i7.13236

33.

Rutaremwa

Wandera

Jhamba

Akiror

Kiconco

Determinants of maternal health services utilization in Uganda. BMC Health Serv Res. 2015;15(1):271. doi:10.1186/s12913-015-0943-8

34.

Ngwira

Lulin

Factors associated with post-natal care (PNC) services utilization among adolescent mothers in rural northern Malawi. Am J Public Health Res. 2021;9(1):38-47. doi:10.12691/ajphr-9-1-5

35.

Agergaard

Subedi

Brøgger

Political geographies of urban demarcation: learning from Nepal’s state-restructuring process. Polit Geogr. 2022;96:102605. doi:10.1016/j.polgeo.2022.102605

36.

Powell-Jackson

Morrison

Tiwari

Neupane

Costello

AM.

The experiences of districts in implementing a national incentive programme to promote safe delivery in Nepal. BMC Health Serv Res. 2009;9:97. doi:10.1186/1472-6963-9-97

37.

Köhler

Cali

Stirbu

Social Protection in South Asia: A Review. UNICEF; 2009.

38.

StataCorp. Stata Statistical Software: Release 18. 2023. Accessed August 9, 2024. https://www.stata.com/

39.

Angrist

Pischke

JS.

Mostly Harmless Econometrics. Princeton University Press; 2009. Accessed August 9, 2024. https://press.princeton.edu/books/paperback/9780691120355/mostly-harmless-econometrics

40.

Hellevik

Linear versus logistic regression when the dependent variable is a dichotomy. Qual Quant. 2009;43(1):59-74. doi:10.1007/s11135-007-9077-3

41.

Mood

Logistic regression: why we cannot do what we think we can do, and what we can do about it. Eur Sociol Rev. 2010;26(1):67-82. doi:10.1093/esr/jcp006

42.

Zou

A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159(7):702-706. doi:10.1093/aje/kwh090

43.

Greenland

Interpretation and choice of effect measures in epidemiologic analyses. Am J Epidemiol. 1987;125(5):761-768. doi:10.1093/oxfordjournals.aje.a114593

44.

Zeldow

Hatfield

LA.

Confounding and regression adjustment in difference-in-differences studies. Health Serv Res. 2021;56(5):932-941. doi:10.1111/1475-6773.13666

45.

Pandey

Adhikari

Sangroula

, et al. Continuum of care for maternal and newborn health services in Nepal: an analysis from demographic and health survey 2022. PLoS One. 2025;20(3):e0319033. doi:10.1371/journal.pone.0319033

46.

Tuladhar

Rehfuess

Delius

Siebeck

Paudel

Oberhauser

Emergency obstetric and newborn care signal functions: the situation in health facilities in Nepal between 2015 and 2021. Health Care Women Int. 2025;46(9):985-1002. doi:10.1080/07399332.2024.2313659

47.

Nepal Health Sector Support Program (NHSSP), Ministry of Health and Population (MOHP)/Nepal, UK Aid. Review of Aama Surakshya Programme in Nepal. Published online January 2020. Accessed January 29, 2026. https://www.nhssp.org.np/Resources/PPFM/Review%20of%20Aama%20Surakshya%20Programme%20in%20Nepal%20-%20January%202020.pdf

48.

Adhikari

Ranabhat

Khanal

, et al. Procurement process and shortages of essential medicines in public health facilities: a qualitative study from Nepal. PLOS Glob Public Health. 2024;4(5):e0003128. doi:10.1371/journal.pgph.0003128

49.

Målqvist

Bhandari

Gurung

Basnet

Sunny

AK.

Payment mechanism for institutional births in Nepal. Arch Public Health. 2021;79(1):163. doi:10.1186/s13690-021-00680-7

50.

de Walque

Kandpal

. Reviewing the evidence on health financing for effective coverage: do financial incentives work? BMJ Glob Health. 2022;7(9):e009932. doi:10.1136/bmjgh-2022-009932

51.

Chalise

Bista

Pandey

Thapa

Correlates of continuum of maternal health services among Nepalese women: evidence from Nepal Multiple Indicator Cluster Survey. PLoS One. 2019;14(4):e0215613. doi:10.1371/journal.pone.0215613

52.

Saugat Pratap

Sharma

Pandey

Baral

SC.

Nepal health financing in light of federalism and pandemic. Perspect Public Health. 2023;143:249-251. doi:10.1177/17579139231180750

53.

Chen

Thapa

Weissglass

Karmachaya

Impact of federalization for health financing and workforce in Nepal. Glob Health Res Policy. 2023;8:19. doi:10.1186/s41256-023-00304-3

54.

Lechner

The estimation of causal effects by difference-in-difference methods. Found Trends Econom. 2011;4(3):165-224. doi:10.1561/0800000014

55.

Abadie

Semiparametric difference-in-differences estimators. Rev Econ Stud. 2005;72(1):1-19.

56.

Cunningham

. Causal Inference: The Mixtape. Yale University Press; 2021.

57.

Wing

Simon

Bello-Gomez

RA.

Designing difference in difference studies: best practices for public health policy research. Annu Rev Public Health. 2018;39(2018):453-469. doi:10.1146/annurev-publhealth-040617-013507

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.28 MB