• safe motherhood;
  • skilled attendance;
  • costs;
  • financing
  • Maternité saine;
  • assistance qualifiée;
  • coûts;
  • financement
  • maternidad segura;
  • atención cualificada;
  • costos;
  • financiamiento


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Objective  To measure costs and willingness-to-pay for delivery care services in 8 districts of Nepal.

Method  Household costs were used to estimate total resource requirements to finance: (1) the current pattern of service use; (2) all women to deliver in a health facility; (3) skilled attendance at home deliveries with timely referral of complicated cases to a facility offering comprehensive obstetric services.

Results  The average cost to a household of a home delivery ranged from 410 RS ($5.43) (with a friend or relative attending) to 879 RS ($11.63) (with a health worker). At a facility the average fee for a normal delivery was 678 RS ($8.97). When additional charges, opportunity and transport costs were added, the total amount paid exceeded 5300 RS ($70). For a caesarean section the total household cost was more than 11 400 RS ($150). Based on these figures, the cost of financing current practice is 45 RS ($0.60) per capita. A policy of universal institutional delivery would cost 238 RS ($3.15) per capita while a policy of skilled attendance at home with early referral of cases from remote areas would cost around 117 RS ($1.55) per capita. These are significant sums in the context of a health budget of about 400 RS ($5) per capita.

Conclusions  The financial cost of developing a skilled attendance strategy in Nepal is substantial. The mechanisms to direct funding to women in need must to be improved, pricing needs to be more transparent, and payment exemptions in public facilities must be better financed if we are to overcome both supply and demand-side barriers to care seeking.

Objectifs  Mesurer les coûts et la volonté de payer pour les soins d'accouchement dans huit districts.

Méthode  Les dépenses familiales ont été utilisées pour estimer les ressources totales requises pour financer: (1) le profil en cours des services utilisés, (2) l'accouchement de toutes les femmes dans un service de santé et (3) l'accouchement à domicile avec une assistance qualifiée, incluant un transfert à temps des cas compliqués dans un cadre offrant des services obstétriques approfondis.

Résultats  Le coût moyen par foyer pour l'accouchement à domicile variait de 410 RS ($5,43) (avec assistance d'un amie ou d'un parent) à 879 RS ($11,63) (avec un professionnel de la santé). L'accouchement normal dans un service spécialisé revenait à 678 RS ($8,97). Quand on ajoute les frais supplémentaires, l'opportunité et le transport, le coût total pour le foyer revenait à plus de 5,300 RS ($70). Pour une césarienne le coût total par foyer dépassait 11,400 RS ($150). Sur base de ces données, le coût pour le financement des services en cours était de 45 RS ($0,60) par capita. Une politique pour un accouchement institutionnel universel coûterait 238 RS ($3,15) par capita alors que l'accouchement à domicile assisté d'une personne qualifiée avec le transfert des cas des zones éloignées coûterait environ 117 RS ($1,55) par capita. Ces sommes sont importantes dans le contexte d'un budget pour la santé d'environ 400 RS ($5) par capita.

Conclusion  Le coût financier pour le développement d'une stratégie d'accouchement avec une assistance qualifiée est important. Les mécanismes pour l'orientation des financements vers les femmes dans le besoin doivent être améliorés. Les tarifications doivent être transparentes et les exemptions au paiement dans les services publiques doivent être mieux équilibrées si nous sommes amenés à vaincre les barrières à la fois du côté de l'offre et de la demande dans les recours aux soins.

Objetivo  Medir el costo y la voluntad de pago por servicios de cuidado durante el parto en ocho distritos.

Método  Se utilizaron los costos familiares para estimar los recursos totales requeridos para financiar: (1) el patrón actual de servicio utilizado; (2) todas las mujeres que dan a luz en una unidad sanitaria; y (3) atención cualificada en partos domiciliarios con transferencia oportuna de los casos complicados a un centro que ofrezca servicios obstétricos completos.

Resultados  El costo familiar promedio de un parto domiciliario estaba entre 410 RS ($5.43) (con un amigo o familiar atendiéndolo) y 879 RS ($11.63) (con un trabajador sanitario). En una unidad sanitaria, el precio promedio de un parto normal era de 678 RS ($8.97). Cuando se añadían cargos adicionales, costos de oportunidad y transporte, el total excedía los 5,300 RS ($70). Para una cesárea, los costes familiares totales eran de más de 11,400 RS ($150). Basándose en estos números, el costo de financiar las prácticas actuales es de 45 RS ($0.60) per capita. Una política institucional universal de partos costaría 238 RS ($3.15) per capita, mientras que una política de atención domiciliaria cualificada con transferencia temprana de casos en áreas remotas costaría alrededor de 117 RS ($1.55) per capita. Estos son números significativos en el contexto de un presupuesto sanitario de aproximadamente ($5) per capita.

Conclusiones  El costo financiero de desarrollar una estrategia de atención cualificada en Nepal es sustancial. Deben mejorarse los mecanismos para dirigir los fondos hacia las mujeres que lo necesitan; la política tarifaria tendría que ser más transparente; y las exenciones en los pagos en los centros públicos tendrían que estar mejor financiados si se han de superar tanto las barreras de oferta como de demanda en la búsqueda de ayuda sanitaria.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Improving maternal and infant health is a major focus of the current national development plan in Nepal (HMGN 2002), yet extension of safe motherhood programmes has made limited headway and measures of mortality and use of services have improved little since the beginning of the 1990s. Nepal's maternal mortality ratio is among the highest in Asia, estimated at around 539 per 100 000 live births (Pradhan et al. 1997). The government is committed to reducing maternal mortality. The recent Health Sector Programme Implementation Plan set a target rate of 300 per 100 000 by 2009 (HMGN 2003).

Despite international recognition of the importance of skilled attendance in maternal mortality reduction (Graham et al. 2001), more than 90% of women in rural Nepal deliver at home with relatives or alone (Pradhan et al. 1997; Osrin et al. 2002). Socio-cultural factors impeding use of obstetric services include high rates of illiteracy, especially in rural areas where less than 27% of women can read or write (HMGN 2001). Many women also prefer the home environment and to deal with problems within the community rather than seek help from outside (Mesko et al. 2003). Limited geographic access to health services is a further barrier particularly in hill and mountain districts (Furber 2002). Although the limited quality of health services in rural communities is an even greater deterrent to service use (Acharya & Cleland 2000; Jahn et al. 2000; Hotchkiss 2001). All these factors are compounded by the ongoing disruption caused by the Maoist insurgency, which affects service delivery in many areas (Thapa 2003).

The affordability of obstetric health services is a further determinant of care seeking (Mbuga et al. 1995; Owa et al. 1995; Wilkinson et al. 2001). Evidence from sub-Saharan Africa and South Asia indicates that households often spend significant amounts for delivery care (Nahar & Costello 1998; Levin et al. 2000), especially if complications arise. Hospital costs associated with a dystocia case accounted for 34% of annual household income in Benin (Borghi et al. 2003). Whilst cost is known to impact on demand for maternity care, little is know about how charges are set within facilities, and if and how exemptions for the poor are enforced. Distance from a facility adds to the financial burden facing households through transport charges and time spent away from productive activity (Kowalewski et al. 2002), although indirect costs have received less attention in the literature.

Much of the past strategy to improve maternal services in Nepal has focused on increasing the volume of service providers, and hence reducing the time and transport costs facing women (Hotchkiss 2001). This is still a core part of current activity although more emphasis is now being placed on demand side factors that impede access to services. The importance of addressing both supply and demand-side barriers simultaneously has been suggested in a number of international studies (Maine 1997; Koblinsky 2003).

Against this background, the DFID-financed Nepal Safer Motherhood Project (NSMP), run by Options, has supported the Government's own National Safe Motherhood Programme since 1997 by seeking to improve service quality in facilities and communities and to reduce access barriers. While cost has been identified as a major barrier to care seeking, there have been no detailed or systematic attempts to measure their full impact in Nepal.

We report on a recent study undertaken for the NSMP to quantify the financial barriers to delivery care seeking for households in Nepal and predict the resource implications of increasing skilled attendance at delivery in different settings. More detailed information is provided in the main report (Borghi et al. 2004).

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References


To measure the household cost of delivery care and better understand household coping strategies, we conducted a study between September and November 2003 in eight districts of Nepal1. The objectives were:

  • 1
    to assess the financial implications of delivering with different levels of assistance: alone or with an untrained attendant at home (a relative or untrained traditional birth attendant), with a trained traditional birth assistant or a skilled health professional, or in a hospital;
  • 2
    to determine household willingness-to-pay for each type of service;
  • 3
    to determine user charges and exemption policies within health facilities;
  • 4
    to estimate the cost to the government of providing delivery care in one zonal hospital.

We distinguished carefully between demand-side costs, incurred outside of the facility during the process of care seeking, and facility-based costs, payments made by patients at the facility during the treatment process. This distinction is important since most public systems finance at least some facility-based costs while often neglecting costs incurred outside, although these can constitute a large financial barrier.

Study design

Two household surveys were carried out estimating both the costs of delivery care (1) and willingness-to-pay (2). Two samples of a total 720 women were selected for interview. Within each district, households were selected from three village development committees2 stratified by distance from the district centre (close, medium distance and far3), as described in Borghi et al. (2004). Briefly, women who had given birth in the previous year were identified with the assistance of community health workers. All those who had given birth in a hospital were interviewed. In addition, 10 women who gave birth at home with a trained attendant and 10 who delivered with an untrained attendant were chosen at random from each VDC, totalling 90 women per district. Facility-based costs were derived from patient records at hospitals. The estimation of all other costs relied upon household recall over a period of between a month and a year (Table 1)4. In order to inform public sector policy, we present here the costs incurred in public hospitals for vaginal deliveries and caesarean sections.

Table 1.  Classification of household costs
Type of costFacilityHome
Facility-based feesRegistration, delivery fee, bed charge, laboratory tests, laundry, food, drugs and medical suppliesNot applicable
Additional chargesGifts to staff and medicines and other items purchased by patients together with the value of food and washing materials brought in from outside the facilityGifts to attendant and medicines, food and washing materials and a safe delivery kit where relevant
Transport feesTo and from the facilityNot applicable
Opportunity cost of timeValuation of the time of those accompanying the woman to the facilityAssumed to be zero since attendants can continue with other activities

For the willingness-to-pay survey thirty married women of reproductive age were selected randomly from each village development committee. Women were first asked to express a preference for place of delivery and attendant. Options considered were: a comprehensive obstetric facility; a basic obstetric facility; a trained attendant at home; an untrained attendant at home. They were asked for the maximum they would be willing to pay to give birth for each option. The survey employed an open-ended question format5 that was felt to be suitable in a context where households are used to paying for medical services. The scenario referred to a future (hypothetical) pregnancy, and therefore, women did not know whether the outcome would be a normal vaginal or a surgical delivery and so were required to make an implicit judgment about relative risk.

To estimate socio-economic status for households in both surveys, we used the asset index approach, which ranks households on the basis of ownership of key durable items (Gwatkin 2002). The durable items selected were those used in the demographic and health survey 2001 (HMGN 2001). Quintiles were derived by ordering households by their total asset score and dividing then into five groups of equal size.

A facility survey was conducted with key informants from each of the district hospitals, to assess how fees were set for maternity care and exemption policy and practice for the poor.

To estimate the recurrent costs to government of providing maternity care in a zonal hospital, we conducted a small study using step-down accounting methods based on an adapted version of the WHO Mother and Baby Costing Package6. This study indicated that household direct payments for delivery care amounted to more than 90% of the unit cost of providing these services. Costs derived from the household survey were then used to estimate resource requirements for financing the demand-side and recurrent facility-based costs of three scenarios for skilled attendance coverage in Nepal. Given the variation in household costs between geographical areas, two typical district ‘types’, plain and mountain/hill, were considered, each with an average crude birth rate of 3.4%. A strategy of skilled attendance at delivery includes both availability of skilled attendants (suitable to the level of care provided), necessary supplies and equipment and provision of referral mechanisms to get women to more comprehensive facilities when required (Graham et al. 2001). Currently use of health facilities in Nepal is low (around 56% bed occupancy in 2002 which still leaves around 0.6 million bed-days available). This suggests that some increase in health service use could be absorbed within the existing inpatient infrastructure. For example, providing 4 days of inpatient care to emergency cases (15% of all delivery cases) would require around 0.5 million bed-days and could be met within the existing infrastructure. With this in mind, we examined total and per capita costs of several scenarios.

Current situation.  The first scenario comprises current coverage levels with around 93% of deliveries taking place at home, caesarean section rates of 0.8%7. and other complications, including sepsis, haemorrhage and eclampsia accounting for a further 1.2% of deliveries in facilities.

Institutional delivery.  The second scenario considers that all women deliver at a health facility, with a complication rate of 15%. This provides an indication of the upper limit cost of increasing skilled attendance coverage. However, implementation in the short term would be impractical given insufficient human and infrastructural capacity8. and lack of community acceptance given the preference for home deliveries.

Skilled attendance at home with early referral of remote cases.  This scenario is based on skilled attendants for home deliveries with referral to a comprehensive essential obstetric care facility in the event of complications. Skilled attendants are defined here as a medical professional with midwifery training. The unit cost of a ‘skilled’ home delivery includes the attendant's time based on the charges reported by surveyed women plus the value of a safe delivery kit (950 RS). We assume that all complicated cases (15%) would be treated in a hospital and that women living in mountain districts and more than 2 hours away from a facility (approximately 40%) would move closer to the facility prior to the onset of labour, either to stay with a relative, or to stay in a basic obstetric care facility or a waiting home9. Transport to the health facility and delivery care costs but not ‘waiting costs’ of these 40% are included in the calculations.

All costs are presented in Nepali Rupees (RS) and US Dollar equivalents using the average official exchange rate for 2003, $1 = RS 75.55. Data were analysed using SPSS.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Costs of delivery care and willingness-to-pay

The household survey found that the average cost for a home delivery was 693 RS ($9.17) mostly comprising payments to attendants. The difference between the cost of a delivery attended by a medical professional (879 RS/$11.63) and one assisted by a traditional birth attendant, trained or untrained (759 RS/$10.05), was small. The cost of a delivery attended by a friend or relative was much lower (410 RS/$5.43).

The average facility-based fee for a normal delivery was not significantly different from the total expenditure incurred by those delivering at home (678 RS/$8.97, Table 2). However, once the cost of transport (2812 RS/$37.22), additional items (1354 RS/$17.92) and companion time (492 RS/$6.51) are added, the facility becomes much more expensive (total RS 5336/$70.63).

Table 2.  Household costs of institutional delivery care (Nepali Rupees), 95% confidence intervals in brackets
 Home n = 470(%)Vaginal n = 114(%)C-section (n = 12)(%)
Facility-based fees  678 (539–817)12.7%5500 (2697–8304)48.1%
Additional charges693 (592–795)100%1354 (899–1810)25.4%1469 (−229–3166)12.8%
Transport fees  2812 (1968–3656)52.3%2812 (1968–3656)24.6%
Opportunity cost of time  492 (368–616)9.2%1660 (15–3305)14.5%
Total693100%5336100%11 441100%

Women undergoing caesarean section incurred significantly higher costs in the facility with an average charge of 5500 RS ($72.80) excluding transport and other costs of care. While there was little difference in facility-based costs between geographic areas, the costs of transport varied widely from an average of 1155 RS ($15.29) in flat plains to 3100 RS ($41.03) in hills or mountains. In mountainous districts women spent an average of 8.3 hours reaching a comprehensive essential obstetric care facility; in hilly districts, 5.6 hours and in flat districts, 2.8 hours. Mountain dwellers were generally carried by stretcher or doko (cane basket) although some travelled by plane.

Not only were the costs of facility-based delivery considerable but, unlike home delivery, they varied considerably. Variation in costs was greatest for transport and additional items, where the 95% confidence interval exceeded 30% of the mean (compared to less than 20% for the facility-based fee). The confidence interval for the cost of a c-section exceeded 50% of the mean.

For those in the poorest fifth of households, the cost of a normal delivery in a facility represented 3 months of household income compared to just over 1 month in the richest group of households (Table 3). In case of obstetric complications, seeking care for the poorest has catastrophic financial implications. More than a fifth of women delivering at home said that cost was the main reason for not delivering at a facility. The facility survey indicated that most public hospitals claim to fully or partially exempt poor women from charges. In practice, however, the actual cost to households was found to vary little by economic group.

Table 3.  Cost as a proportion of monthly cash income by wealth group*
Place of deliveryPoorest groupWealthiest group
  1. *Based on wealth groups derived from an index of asset ownership.

Public hospital vaginal366%113%

The willingness-to-pay survey indicated that most women (56%) preferred to give birth at home, in the absence of complications. The main reasons given were: the low cost and flexible payment mechanism allowed by informal attendants, no need to travel, and the familiarity of attendants and the home environment. A third of all women preferred to deliver at a comprehensive obstetric facility. Safety and staff experience were highlighted as the main reasons for this preference. 34% of these women were from the highest wealth quintile, compared to only 8% from the lowest.

On average women were willing to pay up to 733 RS ($9.70; median 500 Rs/$6.62) for a delivery at home with a trained attendant. Those who preferred to deliver at a comprehensive essential obstetric care facility (one-third) were willing to pay RS 4886 ($64.67) on average. Willingness-to-pay for a basic obstetric care facility was much lower: 1452 RS ($19.22). Fifteen per cent of women with preference for delivery in a health facility were not willing-to-pay anything.

Costs of extending skilled attendance

The per capita costs of each scenario described in the previous section are shown in Figure 1 for plain and mountain/hill districts. The cost of financing the current situation (Scenario 1) was estimated at an average 45 RS ($0.60) per capita (weighted average of costs in mountain and plain districts). A policy of fully institutional delivery (scenario 2) was estimated to cost an average 238 RS ($3.15) per capita while the skilled attendance at home scenario (scenario 3) was estimated to cost 117 RS ($1.55). These are substantial sums, considering the total per capita public sector expenditure on health of around 400 RS ($5.29) (HEFU 2003).


Figure 1. Costs of alternative service delivery scenarios. Scenario 1: existing pattern of home and institutional delivery; scenario 2: fully institutional delivery; and scenario 3: skilled attendance with early referral of remote cases.

Download figure to PowerPoint

The cost of developing blood banking capacity and demand-creation within communities have been estimated at around $0.37 per capita (Ensor 2004; Poudyal 2004). When these are included, average per capita cost of the skilled attendance at home strategy (Scenario 3) increases to $1.90. This equates to more than $40 million per year across the country or almost 40% of total public spending on health care between 2001 and 2003 (HEFU 2003).


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Nepalese women delivering in health facilities face significant costs both in accessing services (demand-side) and at the facility. The costs incurred represent a considerable proportion of household resources, especially for the poorest. To improve accessibility and equity of service provision, the full extent of the financial burden needs to be taken into account.

While facility-based costs of normal delivery were relatively modest, those for emergency care (caesarean section and other complications10) imposed a heavy burden on households. Obtaining sufficient cash to cover these costs inhibits many from accessing institutional care and delays the decision to seek care: 34% of those attending a facility reported delaying their decision to seek care by an average of 8 hours.

Furthermore, households face considerable uncertainty regarding the total amount of money they are likely to need, due both to uncertainty of clinical need and also because facilities tend not to publish tariffs. None of the facilities examined provided package services with a standard price for the entire episode of care. There is also little consistency in pricing between pharmacies; similar prescriptions cost vastly different amounts (MacDonagh & Neupane 2003). Encouraging or even requiring public facilities to develop standard charges for services that are then widely publicised could help substantially. The charges would have to go beyond the facility-based fee to include at least some of the additional items paid for by households. Current informal incentives are an important obstacle to this. It is well known that many medical staff who practise privately receive rent-free premises from pharmacies in return for referring patients for prescriptions. The lack of transparency in what is required and included in the standard facility bill provides ample opportunity for this informal activity. Childbirth is an expense that households have time to plan for, but to do this they require a better idea of the amount of money that is required.

The study suggests that there is little difference, particularly at the facility level, between the costs of care for the rich and the poor, despite national guidelines that require public facilities to exempt the most vulnerable. These guidelines appear to fail partly because exemptions are ad hoc and partly because there is little external financing. Public health facilities are increasingly forced to rely on user charges to finance services. Global experience suggests that exemption mechanisms are only effective when external funding is provided that is transparently earmarked for this purpose (Bitran & Giedion 2003). However, even if the exemption system worked effectively, it would only apply to a relatively small portion of the total cost incurred by households (facility-based fees accounting for 13% of the total cost of a normal delivery and 48% of the cost of a caesarean section, Table 2).

Our measurement of socio-economic status was based on a weighted sum of household assets, which may be seen as an imperfect proxy for permanent income. However, this approach has been validated in Nepal, where little difference was found between estimated consumption and the asset index as a means of measuring socio-economic inequality in malnutrition amongst under-five children (Wagstaff & Watanabe 2003). The asset approach was also found to be well correlated to the actual income reported by the households in our survey suggesting it is a valid measure of socio-economic status.

Many of the costs to the household are incurred outside the facility and before treatment is obtained. Our study indicates that the fees charged by health facilities for a normal delivery are roughly comparable to the total expenditure incurred for a home delivery. The difference lies in transport and additional costs incurred together with the opportunity costs of companions. If the state is to cover some or all of the fees for hospital services then either resources must be provided to women in communities or some guarantee of funding these additional demand-side costs. Without this guarantee households are unlikely to take the major step of seeking institutional care.

Funds at the community level have been set up by women's groups and cooperatives in a number of districts to alleviate some of the demand-side costs. However, like micro-finance initiatives in other countries, these funds tend not to be used by the poor who cannot afford to repay the loans provided (Neupane 2004). The funds’ impact seems to be limited as less than 2% of women we interviewed reported them to be their main source of finance used to pay for care. This could be due either to insufficient cash within the funds, or lack of management capacity to stop defaulters (Shehu et al. 1997).

Other mechanisms for solving the problem of limited cash availability in poor households are therefore needed to reduce demand-side barriers, especially for those in remote areas. While there are a number of possible modalities, such as vouchers and external subsidies for community funds, it is important to develop mechanisms to reach those most in need and to overcome demand-side barriers within communities. For women in remote communities free facility-based care is unlikely to be a sufficient incentive to use them.

While it is already part of the Government's strategy to increase the number of health facilities, reducing transport costs, relatively little attention has been paid to other demand side costs. Ways of tackling them are usually relegated to sub-components of donor-financed projects. Yet these costs represent the largest component of cost for women having a normal delivery in a facility. An important decision is the extent to which accessibility should be improved by extending the supply network or by improving referral to existing facilities. This is not an either/or choice; a combination of improved financing for demand-side costs and early referral in remoter areas together with an increase in the number of comprehensive essential obstetric care facilities is likely to be desirable. This suggests the need for cost-effectiveness analyses comparing a marginal increase in facilities to demand-side subsidies. It may also require a policy reconsideration to allow Government to invest in demand creation as well as service delivery.

In Nepal, where accessibility is a major barrier, such a strategy cannot be confined to speedy referral of complications once a woman is in labour. The problem with a more ‘conservative referral’ strategy is that there is much evidence that approaches for identifying high risk women through antenatal care prior to labour are substantially flawed (Carroli et al. 2001). Scenario 3 tackles this differently, by attempting to ensure that all women in remote areas are within reach of a CEOC facility prior to the onset of labour.

Another approach would be to bring comprehensive essential obstetric care services closer to remote populations. The principal way of doing this is by upgrading basic obstetric facilities with operative and blood transfusion capabilities. Women might also be treated at a basic care facility in a way that at least provides first aid for complications through, for example, the use of misoprostal to control haemorrhage. Given the reluctance of many women in Nepal to leave their homes and seek early facility-based care, this may well be more culturally appropriate. Such changes do, however, have substantial investment implications and the relative cost-benefit of demand vs. supply-side interventions requires further investigation.

What is the correct mix of public and private funding for services?

An extremely limited public budget together with significant demand-side costs means that developing a comprehensive financing strategy for safe motherhood will be difficult. Our scenarios indicate the high costs of extending skilled attendance coverage even when based on home delivery for most uncomplicated births. This amounted to almost 40% of total public spending on health care between 2001 and 2003 (HEFU 2003). This is only the cost of the safe-motherhood component of an essential package and not the entire cost of a mother and baby package. Cost sharing between households and government is therefore inevitable if rates of skilled attendance at delivery are to increase in a financially sustainable way. One way to consider this issue is to decide on those costs that should properly be financed from public vs. private sources. While there are no clear rules on how different items should be financed some possible guidance is presented in Table 4.

Table 4.  Costs to be financed publicly or privately
Type of costHousehold type
Home based deliveryPaid for by the household – based on WTP. Possible public provision of safe delivery kitsSubsidised by government or through a system of community financing
Facility-based delivery
 Facility-based feesFully subsidised for complications and emergencies, part-subsidy for normal deliveriesFully subsidised
 Additional chargesDrug and unofficial charges absorbed into facility tariff, costs of food and washing materials financed by householdsDrug and unofficial costs absorbed into facility tariff, costs of food and washing materials financed by households
 Transport feesPooled community financing (loan funds, insurance)Grant subsidy (vouchers?) especially in more remote areas
 Opportunity costsHouseholdMay require a financial incentive

The study indicates that most households are willing and, since contributions appear to vary with income, able to finance home based costs of delivery care. Most women preferred to deliver at home with a trained attendant. Public funding of safe delivery kits and a greater provision of skilled attendants with midwifery training is required, but all but the poorest households might be expected to contribute to costs.

At the facility level the willingness-to-pay study again shows that cost sharing is accepted for the least poor. The high value ascribed to comprehensive compared to basic obstetric care (in many cases exceeding ability to pay) indicates the recognition of the necessity of life saving treatment in the case of real need. However, it is important to note that the valuations given by the willingness-to-pay survey overall closely mirrorred the actual costs of care, suggesting that in contexts where households are used to paying for health care, expected price will influence valuations and not necessarily reflect the service's real worth to the individual.

Payment for additional services could be dealt with in several ways. Items, which are essential to the delivery but currently procured by households, such as antibiotics or oxytocin, might be included in a fixed tariff set by the health facility so that the poor can be exempt from these costs. Other costs, such as food and washing materials, could be left to households.

Transport costs are an important barrier where public intervention is required. Public funding for transport costs for low-income households is necessary since community schemes cannot, on their own, provide sufficient cross-subsidy. It remains an open and researchable question which mechanism should be used to channel purchasing power to those requiring assistance.

Public financing for opportunity costs is likely to be difficult and perhaps contentious. These are not costs that are readily amenable to reimbursement, since there is no bill and costs vary from household to household. Those with highest costs may come from wealthier households. From this perspective it would seem justified that they be financed by households. Yet, for some households, the lost production may be a substantial obstacle to using services. One approach would be to provide grants (or payments) to low-income families to cover the costs of transport and other expenses such as opportunity costs. This amounts to a payment for treatment which has been tested with some success in other countries, for example, increasing compliance with directly observed therapy for tuberculosis, childhood vaccinations and schooling (Giuffrida & Gravelle 1998; Gertler 2000).

The selected approach to cost sharing will impact on the estimated resource requirements for each scenario to increase skilled attendance coverage. If, for example a policy of free delivery care were pursued that finances all facility charges based on Scenario 3, the resource requirements would amount to some $10 million a year. Subsidising transport costs by 50% for all those falling below the poverty line (around 48% of the population) would cost a further $2.5 million. Further costs may be incurred in training and supervising midwives to provide skilled attendance particularly at home where supervision is much more difficult. These costs are not included here but would need to be. It is likely that the cost of the most realistic strategy will lie somewhere between Scenario 1 and 3. Additional public money would be required to fund such a strategy, although the resource requirement would increase only as fast as services were extended to more remote districts. When formulating policy, governments must consider both short and longer run costs of any strategy to increase the coverage of skilled attendance at delivery and ensure that these can be financed from domestic or external sources.

Our results give baseline (Scenario 1) and upper (Scenario 2) estimates on the cost of delivery care together with a possible transitional stage that is costed to allow for more women in remote areas to obtain access to services. There are clear limitations with this third scenario. The cultural preference for home delivery may mean that women are reluctant to move closer to a facility before a delivery even with financial incentives. The 15% estimate of complications is based on post-delivery experience and in practice a larger number of women may need to be referred to a facility based on antenatal or early intra-partum indications. If this were so, the costs could approximate Scenario 2 as the number of facility-based deliveries increases. A final issue with promoting this strategy is that women may have difficulty differentiating between formally trained health workers and other types of attendant such as trasditional birth attendants.

The Nepalese Government is now moving ahead with a scheme to provide financial incentives for women delivering at a health facility and additional incentives for health workers to attend deliveries at home. This is proposed at a national level although it remains to be seen to what extent the mechanisms can be extended to the more remote districts in a way that has a significant impact on the rate of skilled attendance at delivery.

  • 1

    Stratified in terms of NSMP presence and topography: NSMP - Kailali (plain), Sukhet & Baglung (hill), Jumla (mountain) and Non-NSMP - Jhapa (plain), Gulmi & Bhojpur (hill) and Dolpa (mountain)

  • 2

    An administrative region covering around 60 square kilometres and an average population of 7000.

  • 3

    For the plain districts, near was defined as less than an hour walking; medium between 1 - 2 hours; and far, more than 2 hours. In hill and mountain districts, near was defined as less than 2 hours walking or local transport; medium between 2-8 hours, and far more than 8 hours.

  • 4

    We were unable to interview women at facilities, because it would have taken too long to obtain a sufficient sample size and because women may have been reluctant to respond honestly about informal costs whilst in the facility. However, our findings indicated little difference between reported facility-based costs and hospital bills, suggesting that recall bias was minimal.

  • 5

    How much would be the maximum you would be willing to pay for?

  • 6
  • 7

    Based on estimates from the 2001 Demographic and Health Survey (HMGN 2001).

  • 8

    Over 1.3million bed days would be required which is more than twice the existing spare capacity.

  • 9

    Evidence suggests that the maximum time that women should be from a CEOC facility in the event of haemorrhage is around two hours (UNICEF 1997).

  • 10

    Due to small numbers, we could not estimate the costs of assisted deliveries, but they are assumed to lie somewhere between that of a normal delivery and a c-section.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This work was carried out for the NSMP managed by Options Consultancy Services as part of HMG Nepal Safe Motherhood Programme with the financial support of DFID. We are grateful for the guidance of Melissa Cole, Alison Dembo Rath, Hom Nath Subedi, Greg Whiteside and Sandra MacDonagh. We are also grateful to all those that participated in the household and facility surveys and the work of the field interviewers. None of the above is responsible in any way for errors of fact or interpretation in this article of the underlying report.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Acharya L & Cleland J (2000) Maternal and health services in rural Nepal: does access or quality matter more? Health Policy and Planning 15, 223229.
  • Bitran, R & Giedion U (2003) Waivers and exemptions for health services in developing countries. Social Protection Discussion Paper Series No. 0308. World Bank, Social Protection Unit, Washington, DC.
  • Borghi J, Ensor T, Neupane BD & Tiwari S (2004) Coping with the Burden of the Costs of Maternal Health. Nepal Safer Motherhood Project, part of HMGN Safe Motherhood Programme, Options, DFID and HMGN, Kathmandu.
  • Borghi J, Hanson K, Adjei A et al. (2003) Costs of near-miss obstetric complications for women and their families in Benin and Ghana. Health Policy and Planning 18, 383392.
  • Carroli G, Rooney C & Villar J (2001) How effective is antenatal care in preventing maternal mortality and serious morbidity? An overview of the evidence. Paediatric and Perinatal Epidemiology 15 (Suppl. 1), 142.
  • Ensor T (2004) Financing the Costs of Maternal Care: Cross-Read of Recent Studies. Prepared for Nepal Safer Motherhood Project, Options Ltd, Nepal.
  • Furber AS (2002) Referral to hospital in Nepal: 4 years experience in one rural district. Tropical Doctor 32, 7578.
  • Gertler P (2000) The Impact of Progresa on health: Final Report. International Food Policy Research Institute, Washington.
  • Giuffrida A & Gravelle H (1998). Paying patients to comply: an economic analysis. Health Economics 7, 569579.
  • Graham WJ, Bell JS & Bullough CHW (2001). Can skilled attendance at delivery reduce maternal mortality in developing countries. Studies in Health Services Organisation & Policy 17, 97130.
  • Gwatkin DR, Johnson K, Pande RP & Wagstaff A (2002) Socio-Economic Differences in Health, Nutrition and Population in Nepal, 2000. The World Bank, HNP/Poverty Thematic Group.
  • HEFU (2003) Public Expenditure Review of the Health Sector. Health Economics and Financing Unit, Ministry of Health, HMG Nepal, Kathmandu.
  • HMGN (2001) Nepal Demographic and health Survey. Kathmandu, with New ERA, Nepal and ORC Macro, MD, USA.
  • HMGN (2002) The Tenth Plan 2002–2007. National Planning Commission, His Majesty's Government of Nepal, Kathmandu: .
  • HMGN (2003) Nepal Health Sector Programme – Implementation Plan (NHSP-IP) 2003–2007. Ministry of Health, Kathmandu.
  • Hotchkiss DR (2001) Expansion of rural health care and the use of maternal services in Nepal. Health & Place 7, 3945.
  • Jahn A, DarIang M, Shah U & Diesfeld HJ (2000) Maternity care in rural Nepal: a health service analysis. Tropical Medicine and International Health 5, 657665.
  • Koblinsky M (2003) Reducing maternal mortality: learning from Bolivia, China, Egypt, Honduras, Indonesia, Jamaica and Zimbabwe. Human Development Network, Health, Nutrition and Population Series, The World Bank, Washington.
  • Kowalewski M, Mujina P & Jahn A (2002) Can Mothers afford maternal health care costs? User costs of maternity services in rural Tanzania. African Journal of Reproductive Health 6, 6573.
  • Levin A, McEuen M, Dymatraczenko T, Ssengooba F, Mangani R & Van Dyck G (2000) Costs of Maternal Health care Services in three Anglophone African Countries. Special Initiatives Report 22, Partnerships of Health Reform, Abt Associates, Bethseda.
  • MacDonagh S & Neupane R (2003) Private for Profit Maternity Services, Nepal Case Study. Final Report, Ref AG 3128. Options & Kings College London for DFID, UK and Department of Health Services, HMGN, London.
  • Maine D (1997) Lessons for program design from the PMM projects. International Journal of Gynecology and Obstetrics 59, S259S265.
  • Mbuga JK, Bloom GH & Segall MM (1995) Impact of user charges on vulnerable groups: the case of Kibwezi in rural Kenya. Social Science & Medicine 41, 829835.
  • Mesko N, Osrin D, Tamang S et al. (2003) Care for perinatal illness in rural Nepal: a descriptive study with cross-sectional and qualitative components. Biomedical Central International Health and Human Rights 3.
  • Nahar S, Costello A (1998) The hidden cost of ’free’ maternity care in Dhaka, Bangladesh. Health Policy and Planning 13, 417422.
  • Neupane BD (2004) Emergency Fund Study. NSMP/Options for Nepal Safe Motherhood Programme, HMGN, Kathmandu.
  • Osrin D, Tumbahangphe K et al. (2002) Cross sectional, community based study of care of newborn infants in Nepal. British Medical Journal 325, 10631066.
  • Owa A, Osinaike I & Makinde O (1995) Trends in utilization of obstetric care at Wesley Guild Hospital, Ilesa, Nigeria. Effects of a depressed economy. Tropical and Geographical Medicine 47, 8.
  • Poudyal SR (2004) Costing analysis. NSMP/Options for Nepal Safe Motherhood Programme, Kathmandu.
  • Pradhan A, Aryal R, Regmi G, Ban B & Govindasamy P (1997). Nepal Family Health Survey 1996. His Majesty's Government, Nepal. Ministry of Health, New ERA; Macro International Inc, Calverton, Maryland and Kathmandu.
  • Shehu D, Ikeh AT & Kuna MJ (1997) Mobilizing transport for obstetric emergencies in northwestern Nigeria. International Journal of Gynecology & Obstetrics 59, S173S180.
  • Thapa D (2003) A Kingdom under siege: Nepal's Maoist Insurgency, 1996–2003. The Printhouse, Kathmandu.
  • UNICEF (1997) Guidelines for Monitoring the Availability and Use of Obstetric Services. United Nations Children's Fund, New York: .
  • Wagstaff A & Watanabe N (2003) What difference does the choice of SES make in health inequality measurement. Health Economics 12, 885890.
  • Wilkinson D, Gouws E, Sach M & Karim S (2001) Effect of removing user fees on attendance for curative and preventive primary health care services in rural south Africa. Bulletin of the World Health Organisation 79, 665671.