Estimating the financial impact of livestock schistosomiasis on traditional subsistence and transhumance farmers keeping cattle, sheep and goats in northern Senegal

Background Schistosomiasis is a disease that poses major threats to human and animal health, as well as the economy, especially in sub-Saharan Africa (SSA). Whilst many studies have evaluated the economic impact of schistosomiasis in humans, to date only one has been performed in livestock in SSA and none in Senegal. This study aimed to estimate the financial impact of livestock schistosomiasis in selected regions of Senegal. Methods Stochastic partial budget models were developed for traditional ruminant farmers in 12 villages in northern Senegal. The models were parameterised using data from a cross-sectional survey, focus group discussions, scientific literature and available statistics. Two scenarios were defined: scenario 1 modelled a situation in which farmers tested and treated their livestock for schistosomiasis, whilst scenario 2 modelled a situation in which there were no tests or treatment. The model was run with 10,000 iterations for 1 year; results were expressed in West African CFA francs (XOF; 1 XOF was equivalent to 0.0014 GBP at the time of analysis). Sensitivity analyses were conducted to assess the impact of uncertain variables on the disease costs. Results Farmers surveyed were aware of schistosomiasis in their ruminant livestock and reported hollowing around the eyes, diarrhoea and weight loss as the most common clinical signs in all species. For scenario 1, the median disease costs per year and head of cattle, sheep and goats were estimated at 13,408 XOF, 27,227 XOF and 27,694 XOF, respectively. For scenario 2, the disease costs per year and head of cattle, sheep and goats were estimated at 49,296 XOF, 70,072 XOF and 70,281 XOF, respectively. Conclusions Our findings suggest that the financial impact of livestock schistosomiasis on traditional subsistence and transhumance farmers is substantial. Consequently, treating livestock schistosomiasis has the potential to generate considerable benefits to farmers and their families. Given the dearth of data in this region, our study serves as a foundation for further in-depth studies to provide estimates of disease impact and as a baseline for future economic analyses. This will also enable One Health economic studies where the burden on both humans and animals is estimated and included in cross-sectoral cost–benefit and cost-effectiveness analyses of disease control strategies. Graphical Abstract Supplementary Information The online version contains supplementary material available at 10.1186/s13071-021-05147-w.


Background
Schistosomiasis is a major neglected tropical disease (NTD), second only to malaria as a parasitic disease of humans in terms of socio-economic impact [1]. The causative agents, Schistosoma spp., are dioecious trematodes which affect both humans and animals and are indirectly transmitted to their mammalian definitive hosts via freshwater molluscan intermediate hosts [2][3][4]. Over 240 million people are estimated to be infected with schistosomiasis caused by Schistosoma haematobium (and hybrids therein), S. japonicum, S. mansoni, S. mekongi, S. guineensis or S. intercalatum [5], with more than 90% of human cases occurring within sub-Saharan Africa (SSA) [3].
Whilst zoonotic transmission of schistosomiasis between humans and over 40 potential mammalian reservoir hosts is fully acknowledged within Asia [6][7][8], there is also an increasingly acknowledged zoonotic role within Africa [9,10], as well as an awareness of the morbidity impact of animal schistosomiasis in general [11,12]. Although the total number of livestock infected globally has not been accounted for [13], schistosomiasis in domestic animals often occurs within the same underprivileged communities most affected by human schistosomiasis [9,11]. Furthermore, in addition to the previously assumed host-specific Schistosoma species, across many parts of SSA in particular, viable hybridised combinations including S. haematobium:S. bovis, S. haematobium:S. curassoni, S. haematobium:S. mattheei and S. bovis:S. curassoni have been reported in humans, while S. bovis, S. curassoni and S. mattheei together with S. bovis:S. curassoni and S. bovis:S. mattheei hybrids have been documented in domestic livestock [2,[14][15][16][17].
Since 2002, large-scale mass drug administration (MDA) with praziquantel (PZQ) as preventative chemotherapy in high-risk groups of children, predominantly school-age children, has been implemented across much of SSA [18]. Morbidity control has been generally successful across many countries [19] and has led to a revision of the World Health Organization's (WHO's) strategic plan for a vision of "a world free of schistosomiasis" in 2012 [20,21], and more recently the new WHO NTD Road Map aimed at achieving elimination as a public health problem (EPHP), i.e. elimination of morbidity where the prevalence of heavy infection intensity in school-age children is less than 1% in all endemic countries by 2030, as well as a complete interruption of transmission (IoT, i.e. reduction in incidence of infection to zero) in selected African regions by the same point [22]. However, the sole focus of MDA in humans without complementary control of the disease in livestock, as well as misuse of the only available drug, PZQ, in animals to control livestock schistosomiasis, continues to frustrate efforts to achieve schistosomiasis control and elimination goals stipulated by the WHO within SSA [11].
Furthermore, schistosomiasis has been reported as one of the NTDs with the greatest unequal socioeconomic distribution [23], posing a threat to public health and having grave economic implications [24][25][26]. The drug PZQ is donated at a large scale by pharmaceutical companies, predominantly Merck KGA, and given for free to school-age children across many SSA countries [27] at an estimated value of $32.5 million annually [28]. Evaluations to date described the cost of the disease in humans in terms of disability-adjusted life years (DALYs), qualityadjusted life years (QALYs), the number of working days lost, and the financial burden of the disease [25]. Redekop et al. [29], for instance, conducted a review of studies on the economic impact of human schistosomiasis in terms of treatment costs and disease costs and estimated the global annual productivity loss associated with schistosomiasis at $5.5 billion from 2011 to 2020, and $11.9 billion from 2021 to 2030.
There is a dearth of studies, in contrast, on the economic implications of animal schistosomiasis [11]. A few studies have reported on the treatment costs for the disease to farmers and the biological effects and productivity impact of livestock schistosomiasis. They found that the different species of schistosomes cause organ pathologies in cattle [30], sheep [31] and goats [32], as well as productivity losses of meat, milk and reproduction [33]. To the authors' knowledge, the only published study estimating the economic impact of schistosomiasis in animals in Africa is a benefit-cost analysis of investing in a potential vaccine for schistosomiasis in cattle in Sudan [33]. In this Sudanese study, the disease costs included production losses and the capital and operating costs of the vaccination programme. The benefit-cost ratios were estimated based on infection probability, vaccine uptake, mortality and vaccine production costs. The study showed that for every $1 spent on bovine schistosomiasis in provinces with a 50% infection probability, lower mortality, low vaccination and high vaccine production costs, the benefit-cost ratio was $0.7. However, in provinces with a high infection probability, high mortality rates, high percentage of vaccinated animals and low vaccine Keywords: Schistosomiasis, One Health, NTDs, Livestock, Subsistence farming, Praziquantel, Financial impact, Partial budget analysis, Senegal, Disease control production costs, the benefits were higher, at $12.7, for every $1 invested [33]. These results showed that the development of cost-effective vaccines would yield high returns on investment.
The lack of economic assessments of livestock schistosomiasis makes decisions on investment in the treatment of livestock schistosomiasis difficult, particularly given the need to balance any potential benefits gained with increased risks in terms of the evolution of PZQ resistance [10], and where there might be other endemic disease priorities for the sector. Livestock schistosomiasis not only affects measures to control or eliminate human schistosomiasis but also causes disease costs for farmers, affects livelihoods and reduces the availability of livestock-derived foods for human consumption. Knowledge of the losses caused by the disease and expenditures needed for diagnosis and treatment enables the generation of a baseline of the current impact of the disease [34]. This baseline can then be used in cost-benefit or cost-effectiveness analyses to estimate the potential value of control strategies (e.g., mass or targeted drug treatment of animals) for individual farmers or the sub-sector.
The aim of this study was to estimate the financial impact of livestock schistosomiasis on traditional subsistence and transhumance farmers in selected villages around the Lac de Guiers and Barkedji town in Senegal. The objectives were to (1) establish herd/flock structures and production parameters for a regular cattle, sheep and goat herd or flock in northern Senegal, and (2) estimate losses and expenditures due to schistosomiasis in these production systems. The findings are discussed in terms of the potential economic impact livestock schistosomiasis can have on the livelihoods of farmers and their communities.

Study sites
This research was carried out in two regions in northern Senegal. Six villages were selected around the town of Barkedji (15.2774° N, 14.8674° W) in the Linguere department of the Louga region in the Vallée du Ferlo, and six villages around the Lac de Guiers (16.2247° N, 15.8408° W) near the town of Richard Toll in the Saint-Louis region in the Senegal River Basin (Fig. 1). The Richard Toll/Lac de Guiers area has undergone significant modifications such as desalination and the creation of irrigation canals, with permanent changes to local ecology, favoring expansion of snail intermediate host habitats, and increased sharing of water contact points by communities with their animals. In Barkedji, temporary ponds are an important source of water for human populations and their animals. These ephemeral water sources disappear completely during the dry season, interrupting transmission of schistosomiasis and necessitating seasonal migration by a large proportion of livestockkeeping communities. In both study areas, water contact points are used simultaneously by people and their livestock, encouraging the transmission of schistosomiasis between and within humans and animals [9]. In the area of Lac de Guiers, human schistosomiasis prevalence in humans can be as high as 88%, and 47% in Barkedji [9]. In Senegal, S. bovis, S. curassoni and hybrids of S. bovis:S. curassoni are the prevalent species causing livestock schistosomiasis [6,12]. Recent work of Léger et al. [9] on livestock schistosomiasis revealed that S. bovis is the primary species causing livestock schistosomiasis in the Lac de Guiers area and S. curassoni in the Barkedji area. The prevalence estimates in slaughtered livestock in the two regions were as high as 85% for Lac de Guiers and 92% for Barkedji [9].

Study overview
First, a generic partial budget model for the estimation of disease costs was conceptualised and data needs identified based on knowledge of the effects of livestock schistosomiasis and variables commonly used in impact studies of livestock disease. Subsequently, protocols were developed for a cross-sectional interview-based survey and focus group discussions (FGDs) with farmers covering questions on knowledge, occurrence and manifestations of livestock schistosomiasis, herd and production data, and management of livestock and disease.
The data collected were analysed and used to develop and parameterise specific production and partial budget models for the two sites and to define scenarios in line with local production and management practices. Secondary data and expert opinion were collated to complement the primary data where needed. Finally, livestock schistosomiasis disease costs were estimated for herds or flocks of cattle, sheep and goats using stochastic simulations in RiskAMP Add-in software for Excel with 10,000 iterations for a time frame of 1 year.

Participant selection
Target participants were subsistence and transhumance livestock farmers, i.e., the predominant ruminant production system in the two regions, rearing cattle, sheep and/or goats whose livestock products are consumed by the farmers' households or sold to neighbours/at the local market. The selling of animals often takes place on a need basis to cover expenditures such as school fees; if there is no need, assets are commonly stored in the form of a herd or flock.

Data collection and analysis
Of the 12 villages selected from Barkedji and the Lac de Guiers regions, eight had previously participated in the Zoonoses and Emerging Livestock Systems (ZELS) project, and four villages (two in each region) were newly recruited. For the cross-sectional survey, questions were encoded in Open Data Kit (ODK) mobile data collection software. The questionnaire covered the following topics: demographics, production and management practices (including disease management and selling of animals and products), impact of livestock deaths on livelihood, prevention behaviour in people and animals, knowledge of disease in humans and livestock, signs of the disease in livestock, and equity. Most questions were closed, while a few were open. The full survey questionnaire is available upon request from the corresponding author(s). Each survey participant was also asked to complete a table about the number of animals owned per species, age group (young, adult), sex and breed (local, exotic or cross-bred); this information can be found in Additional file 1. The survey was translated from English to French and administered by local enumerators following a training session with the researchers leading the fieldwork.
Farmers who participated in the survey were also invited to participate in FGDs and participatory group activities to gather data on general signs of animal disease, signs of schistosomiasis in livestock, selling and buying of animals, milk and meat, feed and medicine including prices. All group activities were facilitated by a local enumerator with one person acting as note taker; the language used was Wolof. The full question guide can be found in Additional file 2. Summary notes were generated, and the discussions were recorded in full. The recordings were transcribed and then translated into English by the Senegalese research collaborators.
Data were collected in August and September 2019. Upon completion of the survey, data were downloaded from ODK and stored as an Excel file on a safe Royal Veterinary College [University of London] (RVC) drive. The tables on livestock numbers were collected as hard copies and manually added to the Excel file using the identifier code given to each participant. The translated transcripts of the FGDs were sent to the research team based at the RVC for storage and analysis.

Consent and ethical approval
For all primary data collection activities, the researchers first explained what the study was about, how the data collection would work and the rights of the participants. Following that, each participant was asked to give their consent, which was either recorded as oral or written consent in the survey software or as written consent for the FGDs. Ethical approval was sought and granted by the (i) Clinical Research and Ethical Review Board at the RVC, approval numbers URN 20151327 and 2019 1899-3; and (ii) the Comité National d'Ethique pour la Recherche en Santé (Dakar, Senegal), approval numbers SEN15/68 and SEN 19/68.

Data cleaning and analysis
Survey data were checked for completeness and cleaned, which entailed mainly harmonisation of spelling in open question fields. Answers available in French in the open comment fields were translated to English by the authors and professional translators. Data on the demographics of respondents, knowledge on schistosomiasis and the economic impact of the disease were analysed. Microsoft Excel was used to calculate summary statistics and to visualise the data. For uncertain variables (e.g. those with skewed distributions, inconsistency or too few responses), probability distributions were assigned. The open questions were read in detail in the search for information that would be relevant for the conceptualisation of the economic models including the definition of scenarios; relevant information was extracted as summary statements. For example, some respondents stated that sick animals in the herd will lose value and condition and explained a need to replace them with new ones; this informed the replacement strategy used in building the models. Data about why livestock are kept, milking animals with schistosomiasis, and which animals are sold and bought were extracted from individual interviews. Data from the group activities were analysed to identify information on daily feed quantity and type of feed consumed by animals, cost of feed, whether or not farmers sell sick animals, and questions on whether animals with schistosomiasis sell differently. Common topics were identified across responses for the FGDs and interviews which were used to inform the structure of the partial budget model and the input variables.

Estimation of the financial impact of livestock schistosomiasis Model development and scenarios
Stochastic models were developed in Microsoft Excel with the RiskAMP Add-in for simulation modelling; they are available on request from the corresponding author.
Programme evaluation and review technique (PERT) distribution was assigned to the identified uncertain parameters. The information gained from the analysis of the primary data collected, available literature and expert opinion was used to decide on what species to include, and to define scenarios for the financial impact analysis. The data were used to model a representative herd or flock for each species including the number of animals per age group and sex. Further, the information was used to define scenarios for the analysis.
Integrated production and partial budget analysis models were set up for 1 year, which is approximately the production cycle of lactating cows in the study populations. Two scenarios were considered based on the most common practices reported by respondents. Scenario 1 was a situation where farmers would test and treat their animals when seeing clinical signs consistent with livestock schistosomiasis. Scenario 2 was a situation where farmers would not test or treat their animals when seeing schistosomiasis in their herds or flocks. Detailed scenario descriptions are given in Table 1.

Partial budget analysis
The financial impact per year was the net value estimated for each species and scenario using the following basic equation: Each of the six models (two scenarios per species, three species in total) had distinct input parameters as listed in Table 2 (general input variables) and Table 3 (scenariospecific input variables).
New costs were additional costs for testing and treatment and replacement of sick animals.
For scenario 1, this included the following costs: where N Y stands for the number of young animals, Mb Y the morbidity rate of young animals, P TS the proportion of sick animals tested, and Pr Te the price of testing per animal.
where N A stands for the number of adult animals, and Mb A the morbidity rate of adult animals. (1) Net value = (Costs saved + Added revenue) − New costs + Revenue forgone (2) Testing of young sick animals Testing of adult sick animals Treatment for sick animals tested where P TT stands for the proportion of tested animals that are treated, and Pr Tr the price of clinical treatment per animal.
where P UTT stands for the proportion of untested animals treated.
For scenarios 1 and 2, this included the following costs: where Pr AHA stands for the market price of an adult healthy animal, Pr YHA the market price of a young healthy animal, P SUT the proportion of sick animals sold among those not treated, and P SAR the proportion of young sick animals sold that are replaced. Revenue forgone stemmed from milk not sold or sold at a lower price and selling animals at a lower market value. For scenarios 1 and 2, this included revenue forgone as follows: where P LF stands for the proportion of lactating females among the adult animals, R LF the rate of reduced lactation duration in sick females, D CI the duration of clinical illness if an animal is not treated, R MY is the rate of reduced milk yield in sick females, M HA the daily milk quantity in healthy animals, and Pr MHA the price of milk per litre for a healthy animal.

(5)
Treatment for sick animals not tested Replacing sick animals sold Milk not sold from sick females (kept in the herd) Milk not sold from sick females (kept in the herd) due to reduced milk production per day where Pr MSA is the price of milk per litre for a sick animal.
where D CIS is the average duration of clinical illness before the animal is sold.
where Pr ASA stands for the market price of an adult sick animal and Pr YSA for the market price of a young sick animal.
where Mt A and Mt Y are the mortality rates for adult and young animals, respectively, among those sick and not sold.

(9)
Milk sold from sick females (kept in the herd) at lower market price Milk sold from sick females (before the sick females are sold) at lower market price Sick animals sold at lower market price Value reduction of sick animals not sold (but alive) Herd value reduction due to sick animals sold and not replaced Value reduction of sick, untreated animals not sold and dead Table 1 Definitions of scenarios for the partial budget analysis. Scenario-specific input parameters are given in Table 3 a Sick animals are animals with clinical signs Replacement strategy Treated animals will recover and not be replaced. The majority of untreated sick animals, irrespective of age, will be sold at a lower market price. A proportion of the animals sold will be replaced with the same type of animal (young for young, adult for adult) The majority of sick animals, irrespective of age, will be sold at a lower market price. A proportion of the animals sold will be replaced with the same type of animal (young for young, adult for adult) Sick animals in the herd will lose value and condition, hence the need to replace them with new ones      Expenditures saved stemmed from saving concentrate feed, supplements and routine treatment. For scenarios 1 and 2, this included expenditures saved from the following: where D S stands for the average duration without the animals sold and not replaced in the herd/flock, F HA the daily concentrate feed quantity in kilograms in healthy animals, and Pr F the price of concentrate feed per kilogram.
where S HA stands for daily supplement quantity in kilograms in healthy animals and Pr Su the supplement price per kilogram.
where Pr RT stands for the price of routine treatment per animal per day. (15) Concentrate feed saved on sick animals sold and not replaced Concentrate feed saved on sick, untreated animals not sold and dead Supplement saved on sick animals sold and not replaced Supplement saved on sick, untreated animals not sold and dead Routine treatment saved on sick animals sold and not replaced Extra revenue comprised the revenue from selling sick animals: The partial budget models did not consider the effect on labour, as these production systems rely predominantly on unpaid family labour. All prices used for the models were in Senegalese currency, i.e., the West African CFA franc; 1 XOF = 0.0014 GBP as at the time of (20) Routine treatment saved on sick, untreated animals not sold and dead Revenue from sick animals sold due to disease analysis (2020). Each partial budget analysis model was run with 10,000 iterations, and the net values were assigned as outputs. Finally, the impact of uncertain variables on the output of models (net value) was conducted using the built-in function performing univariate regression analysis.

Respondent demographics
A total of 92 respondents representing different households participated in the survey; demographic characteristics are shown in Table 4.

Production and disease management
Local, cross and exotic breeds of all three species were kept in the two study areas (Additional file 3). In both study areas, the predominant breeds in all species were local breeds. Cattle were regarded by survey respondents as the most important livestock (49% of respondents), followed by sheep (27% of respondents) and then goats (5% of respondents). The animals were mostly kept for dual production purposes such as meat and breeding, dairy and breeding or meat and dairy, and the triple combination of meat, dairy and breeding (Additional file 4). In the predominant breed, i.e., local breed, cattle, sheep and goats were kept mostly for the triple purpose of meat, dairy and breeding (41%, 34% and 35%, respectively) and the dual purpose of dairy and breeding (30% for cattle, 22% for sheep and 15% for goats). With regard to the treatment of animals, 57/92 respondents (62%) stated that they routinely treated their animals. A total of 84/92 respondents (91%) stated that they routinely gave their animals supplements.

Signs of schistosomiasis in animals and schistosomiasis-related management practices
A total of 81/92 respondents (88%) reported that they knew that animals could be infected with schistosomiasis, while 11/92 respondents (12%) reported not knowing. The most common signs of schistosomiasis reported by survey respondents for cattle, sheep and goats are displayed in Table 5. A total of 48/92 respondents (52%) reported that they would seek advice from local veterinary workers if they thought their livestock had schistosomiasis; 33/92 respondents (36%) had never tested their livestock in the past for schistosomiasis and 28/92 respondents (85%) used a veterinary clinic. With regard to treatment, 35/92 respondents (38%) stated that they had treated their livestock for schistosomiasis in the last 4 years, with 33/92 respondents (36%) using "Tenicure" (PZQ-levamisole combination) to treat.

Net disease value estimated using partial budget analysis
Results for livestock schistosomiasis costs per animal and year in the three species studied are shown in Tables 6, 7 and 8. For cattle ( In all models, the largest contribution to the total net value was caused by replacement of animals, herd value reduction and revenue from young sick animals sold due to disease. Sensitivity analyses showed that the market prices for young and adult healthy and sick animals had the greatest impact on the net value for all species, with the highest regression coefficients for the market price for adult healthy animals (0.355 to 0.542) followed by the market price for adult sick animals (0.253 to 0.381), the market  Not mentioned 2 (2) price for young healthy animals (0.039 to 0.180), the market price for young sick animals (0.016 to 0.099), the daily feed quantity, the rate of reduced feed intake and the rate of reduced lactation (regression coefficients between 0.01 and 0.03). The proportion of untested animals that are treated also had a noticeable influence on the net value in scenario 1, with regression coefficients of 0.092 for goats, 0.069 for sheep and 0.067 for cattle. The morbidity rate in adult animals had regression coefficients of 0.019 (scenario 1, goats), 0.013 (scenario 2, goats) and 0.011 (scenario 1, sheep); the morbidity rate in young animals in goats had a regression coefficient of 0.012. The variable 'sick animals sold that are replaced' had regression coefficients of 0.021 (scenario 1, goats) and 0.013 (scenario 1, sheep). The other uncertain variables all had regression coefficients < 0.01.

Discussion
In this study the financial impact of livestock schistosomiasis on livestock keepers in two regions of Senegal was shown to be substantial, particularly in scenario 2, i.e., a situation where farmers do not test and treat animals. We observed that the median disease costs in a representative herd for the areas studied were between 0.23 and 1.22 of the average annual income in rural Senegal, with the disease costs highest in small ruminants (the average monthly income for people living in rural Senegal is 57,461 XOF [41]). Thus, having schistosomiasis in a herd will reduce the farmers' livelihood and, in some instances, potentially cause a situation where basic needs can no longer be covered.
The survey data showed that farmers consult a veterinarian or veterinary technician for their animals to be tested, although no information was available on the specific diagnostic test(s) used here by the veterinary technicians (considering the setting of these areas, it is very unlikely that advanced diagnostic tests such as molecular tests were used). Because of the existing practice of selling sick animals, the financial impact estimated was caused mainly by the selling and buying of animals and changes in herd value. With weight loss being a prominent sign of schistosomiasis infection reported by respondents, sick animals fetch a lower market price and cause replacement costs for the farmer. Consequently, farmers have an interest in selling sick, untreated animals as soon as possible to avoid a further reduction in market price. With the clinical signs reported including weight loss, hollowing around the eyes and diarrhoea, sick animals are likely recognised as such by potential buyers, and they will only pay the price for a sick animal.
The subsistence and transhumance farmers studied sell animals based only on needs and usually maintain their herd or flock size as a capital asset; thus, the reduction in herd value was modelled explicitly. In partial budget models for farming units where products are sold to make profits, the change in herd value is not commonly incorporated in a partial budget [42,43]. However, in a setting where the herd or flock is not used as a means to make a profit but functions as a social and capital asset, the estimation of its change in value appears justified. Using the models described, the loss in herd value was a major cost to the farmers, caused mainly by a reduction in animals, as it was assumed that not all animals could be replaced. This was also reflected in the sensitivity analysis, where the market prices of animals were shown to have the greatest influence on the financial impact. Because farmers not testing and treating will have a larger number of sick animals (than those that test and Don't know 9 (13) treat), but most likely will not have the means to replace all the animals they are selling, the financial impact for them was highest. This indicates that testing and treating animals has the potential to reduce the financial impact of livestock schistosomiasis in these populations. A previously published study on rural development and poverty reduction reported that most people in Senegal contribute 50% of their family labour to subsistence livestock farming, which accounts for a 23.8% share of their average income [44]. Many of the respondents from the two study areas examined here considered disease in their livestock as a large economic loss. As these farmers place great importance on their livestock, it is not surprising that some of the farmers would test as well as treat, although the cost of the diagnostic test (1050 XOF) is higher than the medication for the disease. The costs of schistosomiasis treatment (567 XOF) seem to be affordable, yet many farmers were not testing or treating their animals. Farmers who do not test and treat could experience a range of constraints and have other economic priorities. In a study on the attitudes of farmers regarding animal welfare, Kauppinen et al. [45] reported that most farmers considered their welfare and that of their animals as mutually dependent. Though the farmers are aware that their animals can be infected with schistosomiasis, they may not understand that treating the animals also confers protection on them by also potentially interrupting the zoonotic transmission of the disease from animals to humans and preventing hybridisation of species. Thus, further studies may need to look in more depth at the health-seeking behaviour and farmers' motivation for disease control.
The availability of the human formulation of PZQ and the lack of accessibility to a suitably dosed veterinary formula of the drug means that farmers may use donated PZQ intended only for human use to treat their livestock [11]. Consequently, a systematic mis-dosing, and particular under-dosing, of the drug in the animals can be identified as one of the factors which have led to the Milk not sold from sick females (kept in the herd) due to reduced milk production per day 168 297 Milk sold from sick females (kept in the herd) at lower market price 10 18 Milk sold from sick females (before the sick females are sold) at lower market price 12 28 Sick animals sold at lower market price 6560 29,116 Value reduction of sick animals not sold (but alive) 388 3763 reported high prevalence of livestock schistosomiasis in the regions examined [9]. This is a One Health concern, as the use and cross-or misuse of PZQ in animals have been reported to potentiate resistance and reduce efficacy of the drug [10,14,[46][47][48]. The People's Republic of China has already employed potential bovine vaccine development for zoonotic S. japonicum in some regions, in addition to controlled PZQ treatment of bovines, setting the pace for an integrated approach to schistosomiasis, simultaneously combining mitigation measures in animals with control measures in humans as part of its national control programme [49]. The multisectoral and inter-ministerial approach used in China leveraged technological advancements and socio-economic changes [50]. For example, one mitigation measure was to detect the intermediate host, Oncomelania snails, through DNA extraction and loop-mediated isothermal amplification (LAMP), and control the snails using mechanised tractor-plough molluscicide dispensers on marshland regions endemic for S. japonicum [51,52]. In addition, treating bovines against schistosomiasis caused by S. bovis can interrupt the transmission of the disease from animals to humans by preventing possible environmental contamination by schistosomal eggs shed in the faeces of buffaloes [53,54]. China's prevalence of schistosomiasis in humans and bovines is now less than 1% [54]. If countries in Africa were to follow the Chinese example of integrated schistosomiasis control, the estimated high prevalence in humans and animals would be expected to decline. Importantly, the current study models the financial impact of livestock schistosomiasis on a representative herd or flock in the study areas. This study is based on common practices as reported by farmers and reflects a common situation in a regular production year, where there are no major droughts, epidemic outbreaks or similar events. Consequently, the models capture only a narrow set of the infinite possibilities of impact defined by a diverse set of farmers, practices, circumstances, and seasonal and annual fluctuations (caused by weather, celebrations, festive periods, etc.). Further, the input values are based on a wide range of sources and assumptions, as the primary data collected did not cover all aspects sufficiently. For example, limitations were encountered when asking questions about herd size, during which several farmers seemed to give inconsistent answers. This was likely because talking about herd size is taboo based on the belief that talking about it may attract bad luck. This was also found in other studies; for example, Parisse encountered a similar problem of receiving inconsistent or approximate numbers with regard to herd size [55].
The respondents in the current study included transhumance subsistence farmers who rarely kept records. For instance, the mortality rate could not be determined, as the farmers gave no or inconsistent answers to this question. Similarly, the effect on feed use remained inconclusive. The milk yield produced with and without schistosomiasis could not be accurately determined, as respondents typically did not measure the quantity of milk their animals produced or that the household consumed. We also recognised, particularly in the northern Richard Toll regions, that Fasciola could be a confounding factor in the diagnosis of the disease, as many of the farmers reported signs that are attributable to liver fluke and other diseases that we could not always identify. To address these limitations in input parameters, other sources were consulted including related studies, scientific literature and expert opinion. Moreover, sensitivity analyses were conducted to assess the influence of uncertain parameters on the financial impact.
Given the limitations of the cross-sectional dataset in this study, we recommend a longitudinal study design with testing of livestock to determine their schistosomiasis status and the recording of the production, treatment and management data. The generation of such baseline data for livestock populations in Senegalese transhumance and subsistence populations can only be achieved with appropriate investment, but funding for NTDs in livestock is scarce [56][57][58]. There seems to be a general lack of studies of production and economic studies in these settings, a problem most likely exacerbated by a shortage of animal health and One Health economists in the region that could generate knowledge on herd and production data, effects of schistosomiasis in livestock, and health-seeking behaviour. This shortage of capability and capacity will need longer-term investment in education, research and development.
Schistosomiasis is a disease that has a dual burden on human and animal health, and several studies have suggested the role the environment plays in the transmission and hybridisation of the species [16,59,60]. A more holistic analysis of the impacts of the disease using One Health economics is recommended in the future to assess the monetary and non-monetary impacts. Practical methods to evaluate the disease costs for zoonotic diseases may include evaluating the net cost of the disease to all sectors, calculating the separable costs for the human health and veterinary sectors, estimating the costs and benefits of an integrated intervention such as treating livestock schistosomiasis, and analysis of the zoonotic disability-adjusted life year (zDALY) [61].
The current study highlights the financial impact livestock schistosomiasis has on traditional subsistence and transhumance farmers keeping cattle, sheep or goats in northern Senegal. The presence of disease and its effects underscore the need to consider livestock schistosomiasis in control programmes. Since the benefits reaped from the treatment of livestock zoonotic infections also spill over into the public health and medical sectors, albeit at a cost to the agricultural sector, multisectoral collaboration will be needed.