Incremental cost of increasing access to maternal health care services: Perspectives from a demand and supply side intervention in Eastern Uganda

Introduction: High maternal and infant mortality continue to be major challenges to the attainment of the Millennium Development Goals for many low and middle-income countries. There is now evidence that voucher initiatives can increase access to maternal health services. However, a dearth of knowledge exists on the cost implications of voucher schemes. This paper estimates the incremental costs of a demand and supply side intervention aimed at increasing access to maternal health care services.Methods: This costing study was part of a quasi-experimental voucher study conducted in two districts in Eastern Uganda to explore the impact of demand and supply – side incentives on increasing access to maternal health services. The provider’s perspective was used and the ingredients approach to costing was employed. Costs were based on market prices as recorded in program records. Total, unit, and incremental costs were calculated.Results: The estimated total financial cost of the intervention for the one year of implementation was US$525,472 (US$1 = 2200UgShs). The major cost drivers included costs for transport vouchers (35.3%), health system strengthening (29.2%) and vouchers for maternal health services (18.2%). The average cost of transport per woman to and from the health facility was US$4.6. The total incremental costs incurred on deliveries (excluding caesarean section) was US$317,157 and US$107,890 for post natal care (PNC). The incremental costs per additional delivery and PNC attendance were US$23.9 and US$7.6 respectively.Conclusion: Subsidizing maternal health care costs through demand and supply – side initiatives may not require significant amounts of resources contrary to what would be expected. With Uganda’s Gross Domestic Product (GDP) per capita of US$55′ (2012), the incremental cost per additional delivery (US$23.9) represents about 5% of GDP per capita to save a mother and probably her new born. For many low income countries, this may not be affordable, yet reliance on donor funding is often not sustainable. Alternative ways of raising additional resources for health must be explored. These include; encouraging private investments in critical sectors such as rural transport, health service provision; mobilizing households to save financial resources for preparedness, and financial targeting for the most vulnerable. © 2014 Mayora et al.; licensee BioMed Central Ltd.

This costing study was part of a larger study whose implementation design details are extensively described in Ekirapa et al. [26]. Two districts in Eastern Uganda, namely; Kamuli and Pallisa, were included in this study. The districts were selected because they were comparable in terms of their poor maternal heath indicators. Both are rural (geographical location), and have limited capacity to offer maternal health services. Each district had three Health Sub districts (HSD) and one of these was randomly selected as an intervention HSD and one of the remaining two (which most closely reflected similar demographic composition and availability of health services infrastructure) was selected as a control HSD. A HSD is a health administrative area with population ranging from 30,000 to 100,000 with up to 10 health facilities. In the intervention arm, transport and maternal care service vouchers were distributed to pregnant mothers at the ANC clinic during their first visit irrespective of the trimester of the pregnancy (Dec 2009 – March 2010 and June 2010 to June 2011). The transport voucher entitled a pregnant woman to obtain locally available transportation (motor cycle or bicycle) to and from an accredited health facility within their catchment area for four ANC visits, delivery and one PNC visit (hereafter called full package), while the service voucher entitled the pregnant mother to maternal health services at an accredited health facility of their choice. Eligible mothers received both transport and service vouchers. Upon utilizing the transport and maternal services, a mother submitted the voucher to the service provider (transporter or health worker). These service providers later presented the vouchers to the study team for cash reimbursement. Payments were made every three – four weeks as agreed with all service providers. During the pilot period, a full package of services was provided (4 ANC sessions, delivery and one PNC session). In the initial period of the pilot, there was an unanticipated rise in demand for vouchers, and so the transport voucher costs increased. To ensure that these costs were contained, the service package to be covered was scaled down. Hence during the implementation, only delivery care and postnatal care services for those with complications were provided. The selection of delivery and PNC services was informed by available evidence that reveals that most maternal deaths occur during and soon after delivery [28]. However those who had received vouchers for the full package during the pilot period continued to receive all service entitlements including ANC. Women who were referred from a lower level facility to a higher level facility e.g., for a caesarean section or due to other pregnancy related complications, received a ‘special’ voucher for transport that entitled them to use other transport arrangements, usually public taxi or ambulance. Although transport and service vouchers were provided only in the intervention arm, health facility strengthening (health worker training, provision of basic equipment and supplies, and support supervision) was done across both intervention and control arms. The study team from Makerere University School of Public health was responsible for the overall management and administration of the scheme. The team was headed by a study coordinator who liaised with transporters (through organized associations), the health facilities and the district health team. On behalf of the study team, the coordinator distributed vouchers to health facilities where mothers accessed them on their first ANC visits. Upon offering the service (transport and MHC), service provider’s submitted vouchers to the study team (coordinator) and upon verification, the coordinator effected their reimbursement.Figure  1, is a schematic presentation of the management and administrative structure of the voucher scheme that was part of the implemented demand and supply side intervention. Design and implementation structure of the voucher scheme. The costing analysis reported in this paper was done for a one year period (June 2010 to May 2011). Whereas research costs were not included, start-up costs for training and sensitization of communities and other stakeholders were included. For costs that were incurred in the previous year, adjustments were made for inflation. Program costs were analyzed using ingredients approach to costing [29]. Costs were estimated from the provider’s perspective and measured retrospectively based on project accounting, financial and administrative records. The objective of focusing on only the provider perspective was to allow for an understanding of the cost implications of such an intervention in terms of any additional burden to a national health care budget. Thus, economic costs and user contributions are excluded from this analysis. Vouchers were valued according to the reimbursement agreements made between the program and transport and health service providers. Other procurements for the program such as basic clinical supplies distributed across health facilities were valued according to market prices. Costs were classified according to major activity and or source. The different activities and all inputs used for those activities in the intervention area during the implementation period of the scheme were identified and respective costs obtained. The voucher costs included in this study reflect the rates competitively negotiated by the study team and the transport and health service providers. Transport voucher rates ranged between 0.9 USD and 2.3 USD, depending on how far the facility was located, and fuel prices at the time. Payment rates for the public and private facilities were different. The public health facilities received 75 per cent of the rates paid to Private-not-for Profit (PNFP) facilities (Table  1). This is because public facilities are fully financed by government, while the PNFPs receive only partial government support or no support at all. Payment rates for service vouchers also varied during the implementation period. This variation was due to the changing market conditions for the service inputs (supplies), and the initial recorded unanticipated surge in demand for the vouchers. Table  1 provides details of the service voucher fee schedules and fee changes over the implementation period. Fee schedule of health services for different regimes (US$) US$1 = 2,200Ug.Shs. To avoid duplication, forgery and possible fraud, the vouchers contained key client identification information (serialized) that was also recorded in the voucher book and the facility registers. These records were used as the basis for reimbursement for maternal health service vouchers after verification for authenticity. Health facility strengthening costs were associated with training of health workers at health facilities, procurement and distribution of basic drugs, equipment and other supplies, as well as provision of support supervision. Health system strengthening was conducted in both the control and intervention areas for both public and PNFP facilities. These costs were captured separately for the intervention area and control area. In this paper, we report costs for health system strengthening in both areas. Administration costs for the scheme were estimated through reviews of project documents. Administration costs related to expenditures on personnel such as field coordinators and field supervisors. Other costs here included transport for regular monitoring, supervision, coordination, communication, and data collection. Expenditures on other overhead costs such as space, management and finance support were excluded since they were not considered to have been core to the running of the program and their omission could not make a significant difference. Costs incurred during the voucher scheme implementation consisted of both fixed costs and variable costs. Fixed costs included those incurred on administration, sensitization, mobilization, procurement of equipment, and supervision of health facilities. On the other hand, variable costs depended on the number of women utilizing the services. These included cost of transport vouchers, service vouchers, and drugs and supplies. Costs for additional utilization were not calculated separately for two main reasons. It was assumed that costs for additional supplies and equipment which could have been required as a result of the increased utilization had been captured in the costs for the additional supplies and equipment which were supplied as part of the intervention. On the other hand, additional time spent by health workers, was compensated through allowances provided through the service vouchers, implying that these costs are already counted under the service voucher costs. However it is important to mention that the costs could have been underestimated since we did not directly calculate the increase in inputs. Unit costs were estimated for each service provided. This was measured as the total costs per service divided by the number of service users. We separately estimated the average transport costs for each service (for example average total costs for deliveries, ANC and PNC). This was estimated by dividing the total costs incurred in transporting mothers by the number of women transported for these services. Note that there is an overlap for women who received all three services. The data available could not allow us to segregate such women. Women who underwent caesarian section were often referred to a higher level hospital usually outside the catchment area. In such cases, costs for referral transport were higher than the costs for routine transport, and so it was not appropriate to calculate an average cost that would combine routine and referral transport. The authors excluded these in this calculation. Incremental costs for delivery and PNC were later obtained by estimating the additional services induced bythe intervention program. These estimates of incremental effects were obtained by calculating the difference in service utilization (institutional deliveries and PNC) one year prior to the program implementation (December 2008 – November 2009), and during the one year of implementation (June 2010 – May 2011) for both control and intervention areas. The incremental services induced in the intervention and control areas were then added together. These incremental effects were adjusted for underlying service utilization trends in the country based on data from the Uganda Demographic and Health Survey [30] that estimated a 3% annual increase in institutional deliveries from the year 2005 to 2010. The final estimation is based on the assumption that after adjusting for secular trends, the difference in utilization is attributed to the intervention. To understand whether there were other factors that could have influenced the observed difference in service utilization, we collected information on other maternal and child health (MCH) programs in the program area. We found evidence of two other programs (irregular supply of birth kits by the government and provision of ultra sound services in one health facility). Our assessment showed that they were not likely to explain the changes in MCH service utilization observed. Incremental costs in this context were defined as the additional costs that a health system would incur in order to achieve additional utilization (through increased access) of maternal services, beyond what the system currently provides. The additional costs for each service utilized were obtained by attributing major activity costs to each service, based on the proportion of patients who utilized the service.

N/A

• Dima Health