A Cost Analysis of Options for Schistosomiasis Control Mda Programs Targeting Children Aged Five Years and Below in Umkhanyakude District of Kwazulu-natal Province, South Africa

Background Schistosomiasis negatively impacts early childhood development. Inclusion of children aged ve years and below in mass drug administration (MDA) programs for controlling schistosomiasis could improve early childhood development in communities where the disease is endemic. We estimated the projected cost of implementing a schistosomiasis control MDA program for children aged ve years and below in the uMkhanyakude district of South Africa. Method We calculated the cost of implementing a schistosomiasis MDA program targeting children aged ve years and below using an economies of scaled based cost function. We further compared different labor composition simulations to determine the most affordable and available human resources to implement the program. We also explored programs to which the MDA program could be integrated; and estimated what the costs for would be. Moreover, we simulated cost-effectiveness and determined the cost drivers for each simulation considered. major cost driver of implementing a schistosomiasis control MDA program where generic forms of praziquantel are prohibited.


Introduction
Schistosomiasis remains prevalent in many countries especially in those economically disadvantaged [1,2]. Globally, several initiatives and programs have been implemented to control the disease but the debate on the costing of such programs remain inconclusive [1,3]. In a quest to promote health for all, the United Nations General Assembly adopted Agenda 2030 with seventeen Sustainable Development Goals (SDGs) in September 2015. Goal three of the SDGs seeks to promote health for all people of all ages [4]. Scholars and practitioners have since identi ed several diseases that need to be targeted to ful l SDG3; one of them schistosomiasis which is widely prevalent in economically disadvantaged communities [5]. The cost of schistosomiasis treatment can be prohibitive to the affected communities [6]., contributes signi cantly to the accessibility of schistosomiasis control interventions to the community [7]. Schistosomiasis and the other neglected tropical diseases (NTDs) require funds for control and prevention programs in order to reach the target populations [8,9].
The source of funding for schistosomiasis control is often external to the communities that are affected. Most of the funds for the controlling schistosomiasis and the other neglected tropical diseases (NTDs) comes from the governments of the United States of America and the United Kingdom [10]. A signi cant portion of funds for NTDs is provided by philanthropic organizations such as the Bill and Melinda Gates Foundation, END FUND [10] and the Schistosomiasis Control Initiative (SCI). Further support for schistosomiasis control has been provided by the pharmaceutical industry through a WHO initiative to mobilize the donation of praziquantel (PZQ). PQZ is the only WHO approved drug for treating schistosomiasis. The cost of schistosomiasis control can be signi cant but varies by region. For instance, in Egypt it cost US$10 million annually [11] and the country successfully controlled the disease using domestic funds [11]. Although philanthropic organizations have taken a keen interest in nancing of schistosomiasis, the World Bank recommends local nancing as the best option for sustainable control of schistosomiasis [12].
Much emphasis has been placed on Schistosomiasis treatment as not treating infected children negatively impacts early childhood development. This is of great concern to multilateral organizations such as the World Bank which recommends countries to invest in early childhood development programs for poverty alleviation and diminishing the probability of acquiring disease at an early age [13]. One such investment is the PZQ mass drug administration for schistosomiasis as preventive treatment in children aged ve years and below [14]. Schistosomiasis is endemic in six provinces of (Limpopo, North West, Mpumalanga, Eastern Cape, KwaZulu-Natal, Gauteng) South Africa. In the uMkhanyakude district of KwaZulu-Natal and other economically disadvantaged communities, schistosomiasis is widespread and as such, children below the age of ve need schistosomiasis preventive treatment [14,15].
Community-wide treatment has been found to be the most cost-effective strategy to control schistosomiasis [16]. To align schistosomiasis control chemotherapy with the WHO strategy for global schistosomiasis control, the inclusion of children aged ve years and below is paramount [17,18]. The actual cost for schistosomiasis control mass drug administration (MDA) programs is di cult to infer from literature because of the variation in methods and cost lines used in the studies. Furthermore, it is only in 2014 that a working group to provide recommendations on the inclusion of children aged ve years and below in schistosomiasis control preventive chemotherapy programs as part of the Integrated Management of Childhood Illnesses (IMCI) [19]. Thus, there is dearth in the discourse of implementing a schistosomiasis control MDA for children aged 5 years and below.
The South African government, through the National Department of Health, plans to implement a nationwide schistosomiasis control MDA program to control and eliminate the disease [20]. To support this government initiative, we therefore sought to carry out an analysis of the cost of implementing a schistosomiasis control MDA program for children ve years old and below. We conducted the study in uMkhanyakude district, KwaZulu-Natal province where the disease is widespread. The uMkhanyakude district has ve sub-districts. Each of the sub-districts has an average of 15 000 children aged ve years and below [21]. Our analysis considered the economies of scale and economies of scope associated with the mass treatment program. This paper, therefore, is intended to inform policymakers on the budgetary requirements for designing and implementing a schistosomiasis control MDA program for children aged ve years and below in uMkhanyakude district.

Methods
Site selection and study design uMkhanyakude district was purposively selected because it is one of the districts where schistosomiasis is endemic in South Africa [22,23]. The child health program in uMkhanyakude district is implemented through the district Department of Health [24]. The cost analysis was done from a government perspective to establish the cost that the healthcare system in the district would bear in implementing a schistosomiasis control MDA program for children aged ve years and below. The data was collected between August and November 2019. All South African rand (ZAR) values were converted to United States Dollars (US$) for ease of reference and international comparability. In this regard, the data point exchange rate as at 30 November 2019 was US$1:ZAR14.1 [25,26].

Cost Classi cations
The costs of implementing the schistosomiasis program were classi ed into two broad categories; xed and variable costs. Fixed costs are those that remain constant despite changes in the number of children treated in the MDA program. Variable costs on the other hand, change with the number of children treated in the MDA program. The cost of clinic infrastructure was treated as sunk costs and hence were regarded as irrelevant for decision-making. Cost of training was also excluded from the costing exercise mainly because training for the MDA program would be done alongside IMCI training thus making it an indirect cost. Consequently, only direct costs to the treatment program were included in the costing exercise for this study. Table 1 shows the classi cation of costs for the MDA program.

Data Analysis
All calculations were done on Microsoft Excel Spreadsheet. The average cost of treatment per child was calculated by dividing the sum of the xed cost and the variable costs (aggregate) by the total number of children treated. Subsequently, differential costing was determined by subtracting the average cost per child or total cost of treatment in a speci c treatment of scenario from an alternative scenario.
Economies of scale were calculated by determining the difference between the average cost of treating children when the only variable in the treatment scenario is the number of children treated. Economies of scope were calculated by dividing the xed costs with the number of programs that shared resources; adding the variable costs that were speci c to the number of children treated in the MDA program and comparing that to a scenario where the MDA program was not integrated with any other program.
The total cost was calculated based on the number of children where the economies of scale were highest. Costeffectiveness was calculated based on the average cost of treating an additional child to achieve the 75% treatment coverage that is recommended by the WHO [28,29]. The cost proportions were calculated by comparing the contributions of each cost line to the total cost of treating 75% of the children aged ve years and below in uMkhanyakude district. We used the cost of branded praziquantel in our cost analysis, therefore, cost sensitivity analysis was calculated based on the cost of using generic praziquantel and on a scenario that praziquantel was donated to the country. We used generic praziquantel as the recommended treatment to lower the cost of schistosomiasis treatment in South Africa. The donated praziquantel was used in the sensitivity determination because WHO is encouraging pharmaceutical companies to donate praziquantel to communities that need to implement schistosomiasis control mass drug administration programs [19].

Resource Planning and Economies of scale
Community healthcare programs in uMkhanyakude are often implemented or supported by ward-based outreach teams (WBOTs) [24]. A WBOT consists of a professional nurse, a staff nurse and community care givers (CCGs) [24]. If the simulation treatment team consists of the ward-based outreach team (WBOT) and one pediatric specialist, the average  If only the WBOT is used to treat 2500 children, and laboratory diagnosis is limited to a sample of 500 children, the average cost of treatment per child in the MDA program is 18% lower than when all the children are screened and treated for schistosomiasis. If 18 500 children are treated by a WBOT based on results of testing 500 samples, the cost of the MDA program would be 26% lower than when all the children are tested. A scenario in which only a sample of 500 children was tested for schistosomiasis using laboratory methods was used for further calculations. Figure 1 depicts economies of scale for each treatment model.
As shown in Fig. 1, the average cost per child in a schistosomiasis control MDA program decreases sharply when the number of children treated are increased from less than 500 children to about 2500 children. A gradual decrease in the cost of treatment per child was observed if the number of children to be treated was increased to more than 2500.
Implementation strategy and economies of scope.
uMkhanyakude district has xed clinics and School Health Mobile Clinics where the vaccination program is done. In comparing the cost of using the xed clinics and mobile clinics, we assumed that four additional community caregivers would be required in the mobile clinics to make up for the absence of the parents at the schools when the School Health Mobile Clinic is used in crèches and early childhood development centers. It cost 14% more per child to treat the children in the School Health Mobile Clinic than it would cost in the xed clinics when 500 children were treated. If 2500 children are treated, it costs 6% more to treat children in the School Health Mobile Clinic compared to the xed clinic. Figure 2 shows the economies of scope for the program based on simulations.
Currently, the vaccination program in uMkhanyakude District is done in the xed clinics while the school health mobile clinics are used for the deworming program. The deworming program is integrated with the vitamin A supplementation program. If the schistosomiasis control MDA is in integrated with the vaccination program, it costs 33% less to treat 500 than it would to treat 500 children if the MDA is implemented independently in a xed clinic. This cost difference reduces to 14% when 2 500 are treated. If the schistosomiasis control program is integrated with the deworming and vitamin A supplementation program, the cost of treating 500 children becomes 47% lower than treating the children in an independent schistosomiasis control MDA program using a mobile clinic. The cost difference reduces to 22% when 2 500 children are treated.
If the schistosomiasis control MDA for children aged ve years and below is integrated to the xed clinic, it costs 9% more to treat 500 children than when the MDA program is integrated with the deworming and vitamin A supplementation program. The cost difference reduced to 3% when 2 500 were treated. There was no difference in the cost of using either of the options to treat at 17 000 or more.

Total Costs and Cost Effectiveness Analysis
Based on the WHO recommended coverage of 75%, the study found that the estimated total cost of implementing the MDA is US$1.087 million when the schistosomiasis control MDA program for children aged ve years and below is integrated with the immunization program. Integrating the schistosomiasis control MDA program for children aged ve years and below with the deworming and vitamin A supplementation programs was US$1.056 million indicating a 3% reduction in costs from the former.
At present, coverage for the immunization program in uMkhanyakude district is 85% while that of the deworming and Vitamin A program is 74% [21]. The cost effectiveness is calculated based on the average cost per additional child above or to reach the 75% treatment coverage recommended by WHO [28,29]. If the schistosomiasis control MDA program is integrated with the immunization program, 75% coverage would require US$100 per additional child to reach that level.
When the MDA program is integrated with the deworming and vitamin A supplementation program, the additional cost per child is US$1,246 to reach the 75% coverage. Assuming that deworming and vitamin A supplementation program reaches the 85% coverage similar to that of the immunization program, the additional cost per child to achieve the WHO coverage of 75% would be US$97. Table 2 shows the total cost and cost effectiveness analysis of the different options for treatment programs.

Cost Ratios
PZQ was the main cost driver of the MDA program and accounted for 59% of the costs of the schistosomiasis control MDA program for children aged ve years and below when either the xed clinics or mobile clinics were used. For both the xed clinics and the mobile clinics, dose syringes contributed 26% of the costs, food contributed 13% while labor and laboratory costs contributed 1% each to the total costs of the schistosomiasis control MDA program as shown in Figs. 3 below.

Sensitivity Analysis
The main cost driver of the treatment program was the cost of PZQ. The sensitivity of the total costs to the use of generic drugs is detailed in Table 3: Table 3 Sensitivity Analysis Percentage difference in cost of PZQ (brand to Generic) 39% 45% 60% 100% Percentage difference in total cost of treatment -23% -27% -35% -59% If the average cost of generic PZQ is less than the branded PZQ, the total cost of the schistosomiasis control MDA program for children aged ve years and below would reduce by 23%. When the cost of generic drugs is less than the branded drug by 45%, there is a 27% reduction in the cost of the MDA program. A 60% reduction in the generic PZQ compared to the branded drug gave a 35% reduction in the total cost of the MDA program. In the case of donated PZQ the treatment program costs 59% less that when branded PZQ is used to treat schistosomiasis in an MDA program targeting children aged ve years and below.

Discussion
Resource planning The Essential Medicines List of South Africa stipulates that a pediatric specialist should be involved in the treatment with praziquantel of children aged below ve years [30]. South Africa is currently experiencing a shortage of specialist skills across all elds of medicine [31]. Rural areas such as uMkhanyakude District are the most affected by the shortage of medical professionals such as pediatric specialists [32]. General practitioners are also inadequate to service the public health needs of South Africa's rural districts including the uMkhanyakude district. As a result, many of the healthcare programs are managed by WBOTs [24]. The inclusion of participation of a pediatrician in the schistosomiasis control MDA program for children aged ve years and below was found to be the most expensive option in the human resources planning choices that are available. Treatment done using a WBOT only is the most affordable and practical option for a schistosomiasis control MDA program for children aged ve years and below.
Economies of scale.
The cost of treating children aged 5 years and below with PZQ in a schistosomiasis control MDA program decreases with an increase in the number of children treated. This decrease in costs is a consequence of economies of scale [33,34].
Economies of scale occur because the xed programmatic costs such as computers, weight scales and labor remain constant even when the number of children increases [33,35]. The economies of scale increase exponentially until the increasing variable costs are signi cant enough to buffer the reduction in the average cost of treatment per child that occurs as result of the xed costs. uMkhanyakude district has ve sub-districts. Each of the sub-districts has an average of 15 000 children that are aged ve years and below. Our ndings indicate that economies of scale will exist when the MDA program is implemented at sub-district level (population: 15 000). In practice, the economies of scale will only exist up to a speci c threshold beyond which diseconomies of scale will cause the average of cost of treating a child in the MDA program will increase [16,36].
Diseconomies of scale occur when the increase in variable costs due to an increase in the number of children being treated outweighs the reduction in the average cost per child that results from the xed costs remaining constant when the number of children that are treated increases [34]. When diseconomies of scale are experienced operational strategies to reduce costs reduce upscaling costs [34]. A point at which economies of scale reduce signi cantly was identi ed as 2500 children receiving treatment under the program. We recommend that each schistosomiasis control treatment station in uMkhanyakude district should target about 2500 children to be cost effective.

Economies of scope
The WHO recommends that schistosomiasis control MDA be integrated with existing child health activities [37].
Healthcare program integration improves the nancial sustainability and operational e ciency that is needed by NTDs control programs [38]. In uMkhanyakude district, the immunization program and the deworming and vitamin A supplementation programs could be used to implement the schistosomiasis control MDA program for children aged ve years and below [24]. When programs are integrated, some of the common costs are shared between more than one program, thus creating economies of scope [34]. Resource planning is essential in the exploitation of economies of scope because the economies of scope will increase when the biggest cost drivers of individual programs are the same in the programs that are being integrated [39].

Cost drivers
The biggest cost driver of the MDA program was the cost of praziquantel, which contributed 59% of the total cost of the MDA program in the model used in this study. In South Africa, only branded PZQ is licensed for use. The cost of PZQ in South Africa is 50 times more that the WHO expected cost [40]. The licensing of generic PZQ could make a signi cant contribution to the control of schistosomiasis in South Africa and in lowering the cost of schistosomiasis control MDA programs in the country. In Canada the use of generics reduces the cost of drugs by an average of 88% [41]. In South Africa, the difference in cost of drugs (brand to generic) was found to average 39%; 45% and 60% for the high cost; medium cost and low cost drugs respectively [42]. We used these parameters in our sensitivity analysis and found that the use of generic PZQ could lower the total cost of the MDA program by between 23%-35%. Donated PZQ could reduce the cost of a schistosomiasis control MDA program for children aged ve years old and below in the uMkhanyakude district by about 59%.
The second largest cost driver was the dose syringes. The dose syringes could be replaced by spoons provided strict infection prevention control protocols are followed. The use of spoons has been successful in clinical studies [18]. Food also contributed signi cantly to the cost of the MDA program. Food is important in the MDA program to improve the systemic assimilation of PZQ and to ameliorate some of the side effects of PZQ. Schistosomiasis affects poor communities; hence it is expected that most of the children affected by schistosomiasis will have inadequate access to food. Food is therefore an integral part of the treatment process. In instances where feeding schemes are part of the crèches and early childhood development programs, 13% of the total cost of treatment could be removed by treating the children immediately after mealtime. Cost driver reduction measures are more effective when they add value to an already cost-effective program.

Total cost and Cost-effectiveness analysis
The cost effectiveness of PZQ MDA programs in the control of schistosomiasis is understood and well document [43]. We considered programmatic cost-effectiveness in our study to determine the best programmatic strategy that could be used to implement a schistosomiasis control MDA program for children aged ve years and below in uMkhanyakude. Based on the existing coverage (85%) of the xed clinic immunization [21], it is estimated that US$100 per child would be required to achieve the WHO recommended 75% coverage when the MDA program is integrated with the immunization program.
When the MDA program is integrated with the mobile clinic deworming and vitamin A supplementation whose coverage is 74%, US$1246 would be required to close the 1% coverage gap towards the WHO recommended coverage. Based on the cost per additional child only, the immunization program is a more cost-effective approach. However, on assumption that the coverage of the MDA programs is the same at 85% the deworming and vitamin A supplementation program is a more cost-effective vehicle for the schistosomiasis control MDA program for children aged ve years and below. The deworming and Vitamin A supplementation program from 46% in 2017 to 74% in 2019 [21,44]. This increase in coverage shows that economic determination of program integration choices requires predictive analysis of program performance when costs are projected. The use of cost functions that could resolve economies of scale and economies of scope illustrates the complexity of the economic evaluations that are required to inform program integration. In addition to economic evaluations, policy and political decisions may in uence the decision making process.

Strengths And Weaknesses Of The Study
The strengths of this study is premised on the fact that it took into account the economies of scale and economies of scope to determine the cost of implementing a schistosomiasis control MDA program for children aged ve years and below. The study therefore aligns to the programmatic decision making process that is involved in implementing PZQ MDA programs. In addition, the processes that govern healthcare interventions in South Africa is similar throughout the country making the study applicable to several districts in South Africa where schistosomiasis is endemic. The ndings of this study also account for the recommended improvement in schistosomiasis treatment coverage to 85% [1]. The limitation of the study is that some of the initial costs such as community healthcare worker training were not considered in this study as this information was not readily available.

Conclusion
The planned schistosomiasis control MDA program for school going children in South Africa could be expanded to include children aged ve years and below in the districts where schistosomiasis is endemic. Implementing a schistosomiasis control MDA program for children aged ve years and below in uMkhanyakude district can be done by a WBOT. Such an MDA program could be integrated with either the immunization program or the deworming and Vitamin A supplementation program. Using a discount rate of 5%, a 3-year budget for a schistosomiasis control MDA program for Economies of Scope of the treatment program Cost ratios for mobile clinic Figure 4 Cost ratios for xed clinic use