Received: 22 December 2021 Revised: 23 April 2022 Accepted: 5 May 2022 DOI: 10.1111/tbed.14590 ORIGINAL ARTICLE Evaluation of the economic impact of classical and African swine fever epidemics using OutCosT, a new spreadsheet-based tool Jordi Casal1 Damián Tago2 Pilar Pineda1 Blagojcho Tabakovski3 Imelda Santos4 Carolyn Benigno5 Tran Huynh6 Giovanna Ciaravino1 Daniel Beltran-Alcrudo7 1 Department Sanitat i Anatomia Animals, Universitat Autònoma de Barcelona, Abstract Barcelona, Spain African swine fever (ASF) and classical swine fever (CSF) are two major transboundary 2 Food and Agriculture Organization (FAO), Regional Office for Asia and the Pacific, animal diseases of swine with important socioeconomic consequences at farm, subna- Bangkok, Thailand tional and national level. The objective of this study was to evaluate the direct cost of 3 Food and Veterinary Agency of the Republic outbreaks and their control at country/regional level in four countries: namely CSF in of North Macedonia, Skopje, North Macedonia 4 Colombia in 2015–2016, the retrospective cost of ASF in the Philippines in 2019 and in Bureau of Animal Industry, Department of Agriculture, Quezon City, Philippines a province of Vietnam in 2020 and a hypothetical ASF scenario in one region in North 5 Philippine College of Veterinary Macedonia, using the newly developed Outbreak Costing Tool (OutCosT). The tool cal- Epidemiologists, Quezon City, Philippines culates the costs of 106 different items, broken down by up to four types of farms, and 6 Food and Agriculture Organization (FAO), by who assumes the cost (whether veterinary services, farmers or other stakeholders). Rome, Italy 7 Food and Agriculture Organization (FAO), The total cost of CSF in Colombia was US$ 3.8 million, of which 88% represented the Regional Office for Europe and Central Asia, cost of the vaccination campaign. For ASF, there were wide differences between coun- Budapest, Hungary tries: US$ 8,26,911 in Lao Cai (Vietnam), US$ 33,19,666 in North Macedonia and over Correspondence US$ 58 million in the Philippines. While in the Philippines and Vietnam, 96–98% of the Jordi Casal, Department Sanitat i Anatomia cost occurred in the affected farms, the highest expenditure in North Macedonia sce- Animals, Universitat Autònoma de Barcelona, 08193 Bellaterra, Barcelona, Spain. nario was the movement control of the neighbouring and at-risk farms (77%). These Email: jordi.casal@uab.cat important differences between countries depend on the spread of the disease, but also Funding information on the production systems affected and the measures applied. Apart from the financial Food and Agriculture Organization of the cost, these diseases have other negative impacts, especially in the livelihoods of small- United Nations, Grant/Award Number: TCP/RER/3704; United States Agency for holder farms. The OutCosT tool also allows users to evaluate qualitatively other impor- International Development (USAID), tant aspects related to the epidemics, such as the impact on human health, the environ- Grant/Award Number: 720FDA19IO00092 ment, animal welfare, socioeconomic vulnerability, trading and political response. OutCosT, which is a FAO corporate tool (available online at: https://www.fao.org/ fileadmin/user_upload/faoweb/animal-health/OutCosT_PIG.xlsx), can be an impor- tant tool to support country authorities to rapidly respond to a swine disease outbreak This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. © 2022 The Authors. Transboundary and Emerging Diseases published by Wiley-VCH GmbH. Transbound Emerg Dis. 2022;1–11. wileyonlinelibrary.com/journal/tbed 1 2 CASAL ET AL . by estimating the associated costs and for advocacy purposes to mobilize resources at national or international levels. KEYWORDS African swine fever, classical swine fever, Colombia, economic cost, North Macedonia, the Philip- pines, tool, Vietnam 1 INTRODUCTION Today, ASF is without doubt the most important and economically dev- astating disease of swine. Present in the five continents, the disease Animal diseases, especially transboundary animal diseases (TADs), can had never infected and killed so many animals as it does today, seri- have important economic consequences at farm, regional and national ously becoming endemic in affected nations and threatening to con- level due to livestock production losses and the high costs of pre- tinue spreading into still unaffected countries. vention, control or eradication measures. Even if these costs are usu- In Africa, the economic cost of ASF had been estimated at US$ ally supported by governments, they still represent an important bur- 15,13,340 in Benin between 2014 and 2018 (Ohouko et al., 2020). The den for livestock producers. The economic consequences of TADs cost of an epidemic in Nigeria in 2001, just due to the high mortal- present important differences across regions and countries (Marsh ity (91%) in 306 farms, was US$ 9,41,492 (Babalobi et al., 2007). An et al., 2017), which depend on the disease, its level of spread before epidemic in 219 households in Tanzania translated into a US$ 41,065 being first diagnosed, the structure of the livestock industry, the con- (Kivumbi et al., 2021). Outside of historically endemic regions, ASF trol and/or eradication measures applied and the duration of the epi- caused devastating effects on the swine production in China, tripled demic. the price of live finishers from about 13 yuan/kg to 38 yuan/kg (Huang Besides the economic cost, the introduction of TADs has also an et al., 2021). In Vietnam, 20% of pigs died or were culled within the important social effect, especially in smallholder farms in developing first 5 months after the onset, with an economic impact only in 2019 countries, pushing families into poverty by reducing their purchasing of between US$ 880 million and 4.4 billion (Nguyen-Thi et al., 2021). In capacity, their sources of protein and even their capacity to pay health India, the direct cost due to the loss of animals from April 2020 until and education expenses (Chenais et al., 2017; Cooper et al., 2019). June 2021 were estimated at US$ 37.32 million (Mohan et al., 2021). In the long term, the occurrence of such diseases might shift those Caused by a Pestivirus of the family Flaviviridae, CSF is in regres- affected from livestock farming to other sources of income, change sion worldwide due to the extended campaigns based on vaccination, their social standing and reduce the public confidence on the author- depopulation of infected farms, movement restrictions and surveil- ities (Mohan et al., 2021). Small producers may have varied sources lance. The World Organisation of Animal Health (OIE) recognises as of income (i.e., diversified activities), which can partially compensate disease-free most countries from Europe and Oceania, Kazakhstan, the the losses (Nguyen-Thi et al., 2021). The occurrence of such diseases USA, Canada, Mexico, Costa Rica, Argentina, Chile, Paraguay, Uruguay may create concerns on food safety and drop the consumption of the and some areas of Brazil and Colombia (OIE, 2021). The CSF epidemics affected species because of the fear of zoonotic transmission, which in the 1990s caused high economic losses to affected countries. In will affect not only farmers but also other actors along the value chain, the Netherlands, 429 farms were affected between 1997 and 1998, such as traders, slaughterhouses and retailers (Nguyen-Thi et al., 2021). with a total cost of US $ 2.3 billion and the destruction of 10 million Knowing the cost of diseases outbreaks and applied control mea- pigs (Elbers et al., 1999). This wave also affected eight farms in Bel- sures is critically important for veterinary services to prioritise gium, with a total cost for the country of €11 million (Mintiens et al., resource allocation and to prepare and plan future possible events and 2001). The estimated impact of animal mortality due to CSF in Mexico, interventions (Brown et al., 2021). Unfortunately, the success of con- Brazil and Dominican Republic between 1997 and 2001 ranged from trol strategies depends on many aspects that may differ from one out- US$ 1,40,000 in Brazil to US$ 1 million in Mexico (FAO, 2003). break to another, making it difficult to provide a clear overview of their Only two tools have been described to calculate the cost of differ- economic impact (Horst et al., 1999). Despite the importance of know- ent human or zoonotic diseases at regional/country level including the ing the economic effects of the diseases, there are a few papers describ- Outbreak Costing Tool (OCT) developed by the US CDC (Bodenham ing the cost of TADs. et al., 2021) and a module of the Be-FAST model that allows to calculate African swine fever (ASF) and classical swine fever (CSF) are two of the cost of different swine diseases (Fernández-Carrión et al., 2016). the most important TADs that affect swine production. ASF is caused The objective of this work is to quantify the economic impact and quali- by a DNA virus, member of the Asfarviridae family. The disease was first tatively evaluate other consequences of ASF and CSF outbreaks in four described in Africa in 1921, where it has remained endemic. In the mid- scenarios in three continents: (a) CSF in the Atlantic Coast region of 20th century, ASF was first detected outside of this continent, in the Colombia in 2015–2016; (b) ASF in the Philippines in 2019; (c) ASF in Iberian Peninsula, with some limited spread throughout Europe and the the province of Lao Cai in Vietnam in 2020 and (d) hypothetical ASF Americas, but since 2007, ASF has spread at an unprecedented rate. outbreak in North Macedonia. The calculations were made through the CASAL ET AL . 3 TA B L E 1 Main characteristics of the epidemics for which the economic cost has been calculated Colombia (Atlantic Vietnam Country Coast) North Macedonia The Philippines (Lao Cai) Disease CSF ASF ASF ASF Year 2015/2016 Hypothetical 2019 2020 scenario (2019) Farms in affected area Affected farms (Total) 96,606 2,889 8,62,200 55,647 Population (sows and boars) 1,51,785 17,451 23,65,780 43,999 Population (fatteners) 3,95,938 1,15,931 1,03,43,468 2,30,723 Outbreaks Total confirmed 91 18 18,221 976 Population in outbreaks (sows and boars) 706 151 28,913 833 Population in outbreaks (fatteners) 1,675 28 1,79,681 2697 Affected animals (including deaths) 1,141 150 2,08,595 3530 Dead animals (total) 961 36 10,708 na Stamping out No Yes Yes Yes Neighbouring and at-risk farms Investigated 1,720 549 na na Stamped out No No- na No Immobilized 1,720 549 na 976 General population Increase of surveillance (farms) 77 628 2,423 N/A Increase of surveillance (animals) 215 10,982 na N/A Surveillance in wildlife No Yes No No Number of vaccinated farms 2,64,778a N/A N/A N/A Number of vaccinated animals 25,07,389a N/A N/A N/A Abbreviations: N/A, not applicable; na, not available data. a Numbers refer to the sum of the 2015 and 2016 OUTbreak COSting Tool (OutCosT), a new developed spreadsheet tool tant population of free-range animals. The disease appeared in 2013 that allows to quantify the cost of swine diseases and related control and it remains endemic in the zone (Pineda et al., 2020). We calculated measures, whether at national or subnational level. the cost for the region in the period 2015–2016, when 63 and 28 out- breaks were declared, respectively, mainly in backyard premises. The control of the disease was based on the immobilisation of affected and 2 MATERIAL AND METHODS neighbouring farms, surveillance and vaccination of the whole region (Pineda et al., 2020). This case was used to validate the tool by compar- 2.1 Case studies ing the results with those obtained by Pineda (2021), who calculates the cost for the period 2013–2020. Data from 2015 and 2016 were The main characteristics of the four cases evaluated in this study are selected and grouped using the same headings as OutCosT. The differ- described below and in Table 1. Tables S1, S3, S5 and S7 includes more ences between both methods are presented as a proportion. detailed information of the cases. 2.1.2 ASF epidemic in the Philippines, 2019 2.1.1 CSF in the Atlantic Coast Region of Colombia, 2015–2016 The Philippines reported its first case of ASF in July 2019 in Rodriguez (Rizal Province) affecting several backyard farmers (i.e., raising pigs at a The disease affected an economically deprived region at the north of maximum of 20 fattener heads and/or 10 breeder pigs per household as the country with over 96,000 pig farms, mostly small family or back- defined by the Registry System for Basic Sectors in Agriculture under yard premises. The region has abundant waterbodies with an impor- the Philippine Crop Insurance Corporation). By the end of 2019, ASF 4 CASAL ET AL . had spread to a total of 10 provinces from five regions, affecting more includes options not usually applied to ASF or CSF outbreaks, like treat- than 18,000 farms belonging mostly to the backyard sector. Cost esti- ment or partial stamping outs. mates refer to the calendar year 2019. - Population: the number of farms and animals. The tool allows divid- ing the pig sector into up to four different types of farms, which can 2.1.3 ASF epidemic in Lao Cai Province of be defined by the number of animals and/or by the production sys- Vietnam, 2020 tem. As the characteristics of swine production vary greatly between and within countries, the farms’ classification is not predefined, and Lao Cai is a province of the mountainous Northwest region bordering farm types are not fixed categories. By default, the tool suggests China, covering an area of 6,384 square kilometres. The pig produc- the following categories: (i) industrial; (ii) commercial; (iii) family; (iv) tion in Lao Cai Province is dominant by small-scale producers with less backyard (or any other outdoor or alternative production systems). than 30 pigs per household. First confirmed in Lao Cai Province in June Although they can be modified according to the characteristics of the 2019, by the end of 2019, ASF was already reported in all nine districts pig sector, the country-specific definitions, or the availability of data. of the province, causing the loss of 36,811 pigs (Animal Health from Lao For each farm type, the model considers two subtypes, either sows or Cai, 2019). In 2020, the disease was better controlled with only 3,530 fatteners. By sows, we refer to farms with adult animals, their piglets pigs lost due to ASF (both direct deaths and the result of culling). We and the weaners (i.e., weaning units are included within this group calculated the cost in the province for calendar year 2020. to simplify the model). Farrow-to-finish farms should be considered twice: as sow farms and fattener units. - Data about the epidemic: The number of suspected (by passive surveil- 2.1.4 A hypothetical scenario of ASF in North lance) and confirmed farms for each farm type are defined, including Macedonia the total number of animals in each group. Data can be filled as total numbers in a real situation or as percentages of the population when The country has mainly family farms with low number of animals. In the calculating the cost of a hypothetical scenario. hypothetical scenario, the outbreak occurred in the North-eastern and - Control measures in the confirmed farms: The model allows for up to Eastern regions, those with the highest pig population of the country three stamping out strategies: all farms, only some farms and all (O’Hara et al., 2021), and it was eradicated before its spread to other farms but just some animals, for example, partial stamping out. If ani- regions. The scenario assumed that ASF was introduced into domes- mals are not culled and some treatment is applied, the tool allows tic pigs, affecting 18 farms: nine farrow-to-finish family farms (for a specifying up to two different drugs. For each treatment, collected total of 25 sows and 226 fatteners) and nine backyard premises (for information refers to the animals that were treated (e.g., all animals a total of 16 fatteners); while no wild boar were affected. The evalua- present in the farm, only affected animals, sows or fatteners, only), tion is based on population, costs and prices from calendar year 2019 the price of the drug and that of its application. Cleaning and disin- as part of the follow-up to the survey on ASF conducted in the coun- fection and insect control are other measures that can also be evalu- try (O’Hara et al., 2021). Besides, the choice of North Macedonia as ated in affected farms. case study country was justified by the high risk for ASF introduc- - Control measures in neighbouring and at-risk farms: This accounts for tion, given the presence of active outbreaks in two bordering coun- measures that can be applied to nonaffected farms that are at high tries, Serbia and Bulgaria. In fact, two ASF outbreaks were reported in risk of becoming infected, like those subjected to an outbreak inves- North Macedonia in 2021, after the present economic evaluation was tigation and testing, stamping out, immobilisation, disinfection and conducted. increased surveillance. The main control measures applied in the three ASF cases - Control measures applied in other farms beyond the affected or the were depopulation, outbreak investigations, cleaning and disinfection, neighbouring/at-risk farms: Vaccination and other measures may be surveillance activities and public awareness campaigns. applied either at the whole population or just in a buffer zone, a region or a certain production system. In either case, the number of animals and farms should be specified to calculate the costs. Control 2.2 The tool measures applied to wildlife can also be included. - Active surveillance: The tool allows to apply two different active The OUTbreak COSting Tool (OutCosT) was built in an Excel spread- surveillance strategies in a given number of farms and animals. sheet for the purpose of evaluating the financial cost of swine diseases, Surveillance activities applied to wildlife or vectors are also consid- including the related control measures, both in a real situation or a sce- ered. nario. - Production parameters: Duration of lactation, pigs weaned per sow, Calculations are based on a deterministic model, and the inputs feed intake, mortality at the different ages are some parameters required to run the model are listed below (while details about the not directly related with the disease but required to calculate the formulas are included in the Supporting Information 2). The tool was value of sacrificed/died animals and costs related to the waiting time developed to be applicable to all swine diseases, which is why it before resuming operations. CASAL ET AL . 5 - Production losses due to the disease includes abortions, losses of body and the veterinary services assumed 14.6% and 6.7% of the total cost, weight for fatteners (measured by the increase in days to reach the respectively (Figure 1). slaughtering weight, additional days in which sows remain open and an ‘other loses’ item if needed). - Costs and prices: this section covers prices associated with the swine 3.2 The ASF epidemics of the Philippines (2019) production (e.g., replacement animals, feed and fixed costs per ani- and Lao Cai, Vietnam (2020) and the hypothetical mal), and the labour costs (time spent and salaries) of farmers and scenario in North Macedonia veterinarians for the extra work due to outbreak-related activi- ties (e.g., visiting farms, sampling and testing, culling and carcass The cost of ASF epidemics ranged widely from US$ 8,26,911 in Lao disposal, disinfection, treatment of diseased animals, vaccination, Cai (Vietnam) to more than US$ 58 million in the Philippines (Table 3). surveillance, and so forth). Expenses for the organization of training The average cost per outbreak varied also from US$ 1,84,426 in North and awareness campaigns are also included. Tables S2, S4, S6 and S8 Macedonia to US$ 847 in Vietnam. includes the prices used to evaluate the four cases. Regarding the cost distribution, in the Philippines and Vietnam, more than 95% of the total cost was associated to affected farms (i.e., As model outputs, OutCosT returns the cost estimates of 106 differ- diseased animals and control activities). On the other hand, in North ent items (summarised by sections in Table 2). All costs are also broken Macedonia, the cost attributed to affected farms was very low (about down by farm type. An indication about who assumes each cost (veteri- 2%), while 84% and 10% of the total cost were due to control mea- nary services, farmers and other stakeholders, according to the infor- sures implemented in connected farms and surveillance, respectively mation given by the user for each item) allows assessing how the total (Table 3). costs are allocated between sectors. In all ASF scenarios, most of the costs were paid by the veterinary Besides, the OutCosT dedicates a specific section on a simple qual- services (ranging between 72% and 88%), while farmers covered the itative assessment of indirect and nonmonetary costs providing a remaining part (between 12% and 28%) (Figure 1). framework to evaluate the implications of a disease on human and For ASF, the qualitative cost assessment indicated great concern animal welfare, the sociopolitical context and the environment. The about the effect of ASF on wildlife and about the increase in the price of qualitative assessment is adapted from the FMD impact calculator pork (they were considered extreme in North Macedonia and Vietnam, (available online at: https://www.eufmd.info/impactcalculator), and it respectively, and high in the other countries). Other points that were includes 22 new items grouped into six main categories: socioeco- considered of high concern in at least two of the four countries were nomic vulnerability; human health; environment; animal welfare; social the possibility that people are forced to leave swine farming due to behaviour; trading and political response. For each item, the model economic reasons (Vietnam and Colombia), the timing to pay compen- returns a qualitative estimation of the level of concern (i.e., null, low, sation (the Philippines and Colombia) and the decreased confidence in moderate, high and extreme). This is obtained by combining the scores the government (Vietnam and Colombia) (Table 5). attributed (by the user) to the importance of consequence and to the likelihood of occurrence through a semiquantitative method (from 0 to 4). 4 DISCUSSION This work estimates the cost of different outbreaks of CSF and ASF 3 RESULTS using the OUTbreak COSting Tool (OutCosT), a novel tool which allows quantifying the financial cost of swine disease outbreaks and their con- 3.1 CSF in the Atlantic Coast Region, Colombia trol in a territory, whether national or subnational. (2015–2016) Despite the high burden of some TADs, the number of studies cal- culating the economic impact of swine diseases and their control is The total cost of CSF in Colombia in 2015/2016 was US$ 3.8 million. quite scarce. Rushton (2009) provided an exhaustive review of the Approximately 86% of the total cost was attributed to the vaccina- cost of outbreaks, analysis of disease losses and cost-benefit analyses tion campaign, while the cost due to the disease and control measures of control strategies regarding the most relevant swine diseases (e.g., implemented in the affected and connected farms was 4.3% and 2.3%, ASF, CSF, Aujeszky’s disease and Porcine Reproductive and Respira- respectively, and the cost of surveillance represented 6.9% of the total tory Syndrome – PRRS). (Table 3). The total cost was 11% higher than the cost obtained using a More recently, Brown et al. (2021) published a scoping review partial budgeting analysis (Pineda, 2021) (Table 4). on the economic impact of ASF, CSF and foot and mouth disease The mean cost of the disease and control activities per outbreak was (FMD), examining scientific papers which included an analysis of out- US$ 2,923 (with US$ 975 of them for the activities in neighbouring and break costs/losses. The authors identified 14 studies for CSF and ASF. at-risk farms). Most of the cost (78.7%) was covered by the pig indus- Eight papers reported an economic evaluation of CSF (one retrospec- try through a national pig farming fund, which is financed by a tax that tive study from The Netherlands, and nine scenarios in Denmark, commercial farmers pay per each slaughtered animal. Affected farmers the Netherlands, Belgium and Australia) with costs ranging from US$ 6 CASAL ET AL . TA B L E 2 OutCosT results: Summary of the outputs given by the model Items/activities included Sections Categories (N) Presence of the disease Direct costs of the disease 7 Treatments costs 4 Control activities in suspected Visits to suspected farms 5 outbreaks Immobilisation of farms 4 Eradication/control measures in Value of culled animals 6 outbreaks Culling and carcass disposal 8 Cleaning & disinfection/insect 6 control Indirect costs due to loss of 6 opportunity Measures in neighbouring and Visiting farms 5 at-risk farms Stamping out 8 Cleaning and disinfection 2 Immobilisation of farms 5 Tests to confirm if farms are 2 disease free Indirect costs due to loss of 6 opportunity in case of culling animals Measures in nonaffected farms Vaccination 6 Other measures 2 Surveillance in domestic animals Active surveillance – strategy 3 1 Active surveillance – strategy 3 2 (if applied) Wildlife and vectors Control measures in wildlife 4 Wildlife and vector 6 surveillance Other costs Coordination and 2 bureaucratic tasks Training and awareness 6 campaigns 58,338 in Australia to US$ 3.7 billion in the Netherlands. The economic demiology, location of outbreaks, trade implications, consumer reac- impact of ASF has been assessed in six studies. Two used retrospec- tion and especially to the disease management and control practices tive data (both from Nigeria) and four studies forecasted a hypothet- (Brown et al., 2021; Marsh et al., 2017). ical scenario (from Denmark, Nigeria, Spain and the USA). The costs In this paper, the economic impact of CSF and ASF was assessed ranged from US$ 6,49,000 of annual costs in Nigeria to US$ 94.5 mil- using the same methodology. Besides, in the four analysed case stud- lion of total losses of a swine depopulation in Spain. Considering the ies, diseases mainly affected family and backyard farms; however, significant number of outbreaks reported worldwide, the authors high- there were important differences in the total cost and its distribu- lighted the low number of papers with retrospective economic stud- tion. While in the Philippines (ASF) and Vietnam (ASF), 96–98% of ies on both diseases. Divergences in the economic impact of the dis- the cost occurred in the affected farms, in Colombia (CSF) and North eases were attributed to methodological differences among studies, Macedonia (ASF) affected farms represented only a small proportion and especially to discrepancies in the items included in the analyses and of the total costs (4.6% and 2.2%, respectively), and the major cost fac- a lack of consistency in definitions and descriptions of the used data. tors were the vaccination and measures applied in connected farms Besides, the observed variability can be due to the virus strains, epi- (especially the movement control), respectively. Such differences are CASAL ET AL . TA B L E 3 Cost of the outbreaks of African and classical swine fever (in US$) CSF in Colombia (Atlantic Country Coast) ASF in North Macedonia ASF in The Philippines ASF in Vietnam (Lao Cai) Year of reference 2015/2016 2019 (Hypothetical scenario) 2020 2020 Outbreaks 166,171 (4.3%) 87,800 (2.2%) 57,616,000 (98.1%) 791,223 (95.7%) a b Disease cost 92,770 (2.4%) 17,533 (0.7%) 1,472,720 (2.5%) 0 Outbreak investigation 54,304 (1.4%) 16,220 (0%) 13,030,140 (22.2%) 18,303 (2.2%) Stamping out 0 52,447 (1.1%) 31,976,638 (54.5%) 736,990 (89.1%) c Indirect costs 19,097 (0.5%) 1,601 (0.3%) 11,136,503 (19%) 35,930 (4.3%) Connected farms 88,740 (2.3%) 5,851,038 (83.9%) na 34,625 (4.2%) Immobilisation 49,518 (1.3%) 4,850,782 (77%) 34,625 (4.2%) Other 39,223 (1.0%) 1,000,256 (6.8%) na Vaccination /other 3,358,176 (86.3%) N/A N/A N/A Application 2,710,957 (69.7%) Cost of vaccine 647,220 (16.6%) Surveillance 266,798 (6.9%) 131,071 (10.2%) 981,711 (1.7%) N/A Awareness 9,269 (0.2%) 126,850 (3.8%) 119,136 (0.2%) 1,063 (0.1%) Total: 3,889,155 6,196,760 58,716,847 826,911 Number of outbreaks (N) 91 18 18,221 976 Average cost per outbreak: 42,734 344,264 3,222 847 Abbreviations: “N/A”, Not applicable;“na”, not available data. a The disease cost includes the cost of dead animals and losses of production, including abortion and suboptimal growth (i.e., extra days open in sows, and cost of extra days due to reach market weight). b Disease costs are included in the stamping out cost due to data limitations that do not allow to differentiate between culled and dead animals. c The indirect costs includes the cost of treatments and costs related to the waiting time before resuming operations (i.e., interrupted production). Cost of maintaining empty the farm. 7 8 CASAL ET AL . TA B L E 4 Cost of the outbreaks of classical swine fever in Colombia estimated using OutCosT compared to the costs calculated by Pineda (2021) Type of cost OutCosT (Total country cost) Pineda (2021) Difference (%) Outbreaks 1,77,273 1,32,194 25.7 Connected farms 88,740 85,997 3.2 Vaccination/treatments 33,58,176 30,16,125 11.3 Surveillance (farms) 1,89,163 2,50,538 6.5 Wildlife 0 0 Awareness 9,269 3,219 188.0 Total 38,22,622 34,88,073 11.5 FIGURE 1 Proportion of costs assumed by each stakeholder (i.e., veterinary services, farmers and industry) by country mainly explained by the timeframe, the different magnitudes of the epi- 33,977) and vaccination (difference of US$ 3,42,051) can be explained demics (i.e., number of outbreaks and/or affected animals), the mea- by a lower discrimination of OutCosT between the cost of salaries of sures applied and to a lesser extent by the production systems affected. the farmers and workers from the veterinary services. The main advan- Since the timeframe used for each case is different (as well as the dis- tage of this tool in comparison with the more detailed approach pro- ease in some cases), we cannot draw direct comparisons. However, posed by Pineda is its ease of use (i.e., how easily and quickly it can be our purpose here was to evaluate the tool’s flexibility that allows its performed). use under different situations rather than making direct comparisons The cost of ASF in the Philippines was the highest of the evaluated between countries. ASF epidemics due to the high number of affected premises (18,221) In Colombia (2015/2016), the vaccination campaigns against CSF and the application of stamping out measure in all these farms, which had a cost of US$ 3.3 million (out of a total of US$ 3.8 million). The represented 55% of the total cost. Unfortunately, we could not include campaign was mainly financed (79%) by the national pig farming fund, the cost in neighbouring (noninfected) farms, because reliable esti- through a tax paid by commercial pig producers and managed by the mates were not available in the Philippines, which implies that the total Colombian Pork Producers Association (Porkcolombia). In addition, the cost would be even higher. cost of time spent by animal owners was evaluated in US$ 3,80,133. In Lao Cai Province (Vietnam), the cost was mainly incurred in The presented results have been validated with the results of the affected farms (96%), especially due to the cost of culling and destroy- study from Pineda (2021). Estimations from both studies (Table 4) show ing the animals (89% of the total cost). Due to the unavailability of that results obtained with both methods differ only by 11.5%. Main reliable data, we included the number of dead pigs within the num- discrepancies were observed in the evaluation of the impact of CSF ber of compensated pigs; therefore, the cost of mortality was covered in the outbreaks, the vaccination campaign and the surveillance. The by the stamping out cost. This indicates that the proportion for the higher costs attributed by the OutCosT to outbreaks (difference of US$ cost of culling and destroying should be lower, since it also includes CASAL ET AL . 9 TA B L E 5 Main results of the qualitative estimation of other noneconomic costs evaluated using OutCosT Colombia (Atlantic North Vietnam (Lao Coast) Macedonia Philippines Cai) Socioeconomic vulnerability: • People can be forced to leave the activity Moderate High Null High Human health: • The disease or the control measures increases mental Moderate Moderate Moderate High disturbances of farmers and related people Environment: • Disease can also affect wildlife animals High High Extreme Moderate • Disease can modify the ecological balance or affect an High Null Low Low endangered species • The control/eradication measures affect the environment High Null Null Null Animal welfare: • The disease has an important effect on the welfare of High Low Low Moderate affected animal • The control measures have an important effect on the High Moderate Low Moderate welfare of affected animal Social behaviour: • Deficient management of public opinion related to High Null Low Low response measures Political response: • Timing for paying the compensations for the carried out High Moderate Moderate High measures. • Decrease in farmers confidence in relation to the Moderate High Moderate High administration Trading: • Increase in prices due to a reduction of the production High Extreme High Moderate some direct cost of disease. The same pattern of cost distribution was tool to rapidly evaluate the financial cost of epidemic and endemic observed in Vietnam and the Philippines: the vast majority of costs swine diseases in different situations, and at subnational or national were related to the direct costs due to ASF and the control measures level. The tool allows users to consider different elements involved in the outbreaks, whereas very few costs were attributed to surveil- in the management of swine diseases. It is flexible by allowing the lance activities (1.7% or less) and almost none to trainings and aware- evaluation of real cases or hypothetical scenarios, the inclusion of ness campaigns (0.2% or less). input data in alternative ways and not introducing some data items in Finally, in the hypothetical scenario of an ASF epidemic in North case of lack of information. OutCosT evaluates the costs of diseases Macedonia, most costs were related to the activities and measures and the cost of activities associated with their control, including applied in neighbouring farms (84% of the total cost). This was a conse- measures applied at farm level, in neighbouring and at-risk farms, but quence of the defined scenario, which considered the region with high- also in wildlife. Moreover, the cost of applying prevention and control est number of farms in the country. As a result, the occurrence of 18 measures in the general population (i.e., vaccination, awareness cam- outbreaks implied the immobilisation of 549 connected farms, which paigns, and surveillance in domestic species, wildlife and vectors) are translated into a cost of US$ 2.5 million. The cost pattern differed also also considered. As with any simulation model, the results will depend in the proportion attributed to surveillance strategies and awareness only on the inputs (provided by the users), which should therefore be campaigns, which represented 10% and 4% of the total cost, respec- defined carefully. tively. Depending on the scenarios, some data may be unknown or present OutCosT is to become a corporate tool of the Food and a high uncertainty. In these cases, users can insert average values or use Agriculture Organization (FAO), and it is available online at: different scenarios to calculate an approximate range of the costs. https://www.fao.org/fileadmin/user_upload/faoweb/animal- Another strength of OutCosT beyond its ease of use is that it allows health/OutCosT_PIG.xlsx. Through this validation process, Out- a detailed evaluation of the cost of each activity, differentiating who CosT was demonstrated to be a useful, powerful, flexible, yet simple pays for them (famers, veterinary services and other stakeholders), and 10 CASAL ET AL . breaking down these costs by up to four types of farms, which can be of the four case studies showed big differences depending on the type defined differently depending on the swine production characteristics of disease, incidence and epidemiology of the disease, the density of and the country-specific definitions. The differentiation between types farms and the measures applied in the country. The tool can be applied of farms allows a more accurate evaluation of the total cost and can following the introduction of a TAD, as in the ASF cases, or for endemic provide relevant information for the decision makers at country level. diseases, as the CSF in Colombia. In summary, OutCosT can be used (1) Although due to the different definitions used for each case study, the to estimate the cost of an outbreak that already occurred; (2) to esti- cost estimate by type of farms is not comparable. The knowledge about mate the potential cost of a future outbreak (with best and worst case which groups bear the burden of the costs and which groups receive scenarios) and alternative response strategies and (3) to estimate the the benefits is especially important to plan compensation programs cost of an endemic disease. Moreover, by comparing different scenar- (Marsh et al., 2017). Besides, the tool also allows to make a qualita- ios (e.g., with or without vaccination, or different stamping out strate- tive evaluation of other important aspects related with the epidemics, gies), the tool can inform decisions around control options and guide such as the effect of the epidemics on the human health, environment, decision makers on the optimal strategy. However, this requires con- animal welfare, socioeconomic vulnerability and trading and political sultation with local experts to make the disease-spread scenarios real- response. istic and in alignment with the different strategies assessed. Countries OutCosT focuses on materialized costs due to ASF outbreaks to be that decide to institutionalize the use of this tool will generate substan- bear by the pig sector or the government (central, regional and local). tial data over the years that can then be used as inputs for realistic On the other hand, it is not designed to conduct complex analyses; scenarios. in fact, its simplicity is one of its key advantages. OutCosT does not address macroeconomic impacts. Indirect costs due to trade restric- ACKNOWLEDGEMENTS tions or modification in the consumer habits were out of the scope. The authors acknowledge the national agencies of Colombia, North Other indirect costs such as the restrictions in movements of veterinar- Macedonia, the Philippines and Vietnam for providing the outbreak ians between farms and the stricter measures for animal transporta- and cost data. Acknowledgements also go to the Food and Agricul- tion outside the affected zone have not been calculated either. These ture Organization of the United Nations (FAO), which financed the indirect costs can indeed have an important impact. For example, the study through the Technical Cooperation Project (TCP/RER/3704) reduction in pig meat price in Vietnam reduced from US$ 1.61/kg of live on ‘African swine fever emergency preparedness in the Balkans’. weight to 1.39/kg after the first outbreaks in 2019 (Nguyen-Thi et al., United States Agency for International Development (USAID) grant 2021). In North Eastern India, the foregone export revenue after the (720FDA19IO00092). irruption of ASF has been estimated in US$ 2.47 million (April 2020 to June 2021) (Mohan et al., 2021). CONFLICT OF INTEREST The main purpose of OutCosT is to support authorities to rapidly None. The views expressed in this publication are those of the author(s) respond to ASF outbreaks by estimating the associated costs. Depend- and do not necessarily reflect the views of FAO. ing on how responsibilities are distributed between central and local authorities, an adequate level of resources must be allocated for high- ETHICAL APPROVAL risk areas to ensure that the response to outbreaks is not delayed. No ethical approval was required as this is a research employing sec- Authorities can use the tool to create scenarios to estimate the bud- ondary data. getary needs for an imminent disease incursion or in case the disease spreads to other regions. The budgetary needs coming out from sce- DATA AVAILABILITY STATEMENT narios can be compared with the costs of implementing more radical The data that supports the findings of this study are available in the measures to contain the outbreak to estimate the benefits of doing so supplementary material of this article. and advocate in case the benefits outweigh the costs. The second objective of the tool is raising awareness about the ORCID importance of implementing control efforts to prevent further spread. Jordi Casal https://orcid.org/0000-0002-6909-9366 Being able to provide such monetary figures can prove extremely use- Pilar Pineda https://orcid.org/0000-0001-9752-3438 ful for advocacy purposes to mobilize resources at national or interna- Giovanna Ciaravino https://orcid.org/0000-0002-5796-8093 tional levels (e.g., donors, development agencies). The estimated cost for producers can be used to raise awareness and promote collabora- REFERENCES tions (including resource mobilization) with the private sector for con- Animal Health from Lao Cai. (2019). Results of livestock and veterinary work in trolling the outbreaks. 2019. Orientation and tasks in 2020. Lao Cai Sub-department of Livestock As a conclusion, the application of OutCosT allowed the evaluation and Animal Health. of the financial cost of swine diseases in different situations and at dif- Babalobi, O. O., Olugasa, B. O., Oluwayelu, D. O., Ijagbone, I. F., Ayoade, G. O., & Agbede, S. A. (2007). Analysis and evaluation of mortality losses of ferent levels. Three retrospective economic analyses (CSF in Colombia, the 2001 African swine fever outbreak, Ibadan, Nigeria. Tropical Animal ASF in the Philippines and Vietnam) and a simulation of a hypothetical Health and Production, 39, 533–542. https://doi.org/10.1007/s11250- ASF epidemic in North Macedonia were conducted. The financial cost 007-9038-9 CASAL ET AL . 11 Bodenham, R. F., Mtui-Malamsha, N., Gatei, W., Woldetsadik, M. A., Cassell, Mintiens, K., Deluyker, H., Laevens, H., Koenen, F., Dewulf, J., & De Kruif, C. H., Salyer, S. J., Halliday, J. E. B., Nonga, H. E., Swai, E. S., Makungu, A. (2001). Descriptive epidemiology of a Classical Swine Fever out- S., Mwakapeje, E., Bernard, J., Bebay, C., Makonnen, Y. J., & Fasina, F. break in the Limburg Province of Belgium in 1997. Journal of Veterinary O. (2021). Multisectoral cost analysis of a human and livestock anthrax Medicine, Series B, 48, 143–149. https://doi.org/10.1046/j.1439-0450. outbreak in Songwe Region, Tanzania (December 2018-January 2019), 2001.00429.x using a novel Outbreak Costing Tool. One Health, 30(13), 100259. https:// Mohan, N. H., Misha, M. M., & Gupta, V. K. (2021). Consequences of African doi.org/10.1016/j.onehlt.2021.100259 swine fever in India: Beyond economic implications. Transboundary and Brown, V. R., Miller, R. S., McKee, S. C., Ernst, K. H., Didero, N. M., Maison, R. Emerging Diseases, Advance online publication. https://doi.org/10.1111/ M., Grady, M. J., & Shwiff, S. A. (2021). Risks of introduction and economic tbed.14318 consequences associated with African swine fever, classical swine fever Nguyen-Thi, T., Pham-Thi-Ngoc, L., Nguyen-Ngoc, Q., Dang-Xuan, S., Lee, H. and foot-and-mouth disease: A review of the literature. Transbound- S., Nguyen-Viet, H., Padungtod, P., Nguyen-Thu, T., Nguyen-Thi, T., Tran- ary and Emerging Diseases, 68, 1910–1965. https://doi.org/10.1111/tbed. Cong, T., & Rich, K. M. (2021). An assessment of the economic impacts of 13919 the 2019 African Swine Fever Outbreaks in Vietnam. Frontiers in Veteri- Chenais, E., Boqvist, S., Emanuelson, U., von Brömssen, C., Ouma, E., Aliro, nary Science, 8, 686038. https://doi.org/10.3389/fvets.2021.686038 T., Masembe, C., Ståhl, K., & Sternberg-Lewerin, S. (2017). Quantitative O’Hara, K. C., Beltrán-Alcrudo, D., Hovari, M., Tabakovski, B., & Martínez- assessment of social and economic impact of African swine fever out- López, B. (2021). Descriptive and multivariate analysis of the pig sector in breaks in northern Uganda. Preventive Veterinary Medicine, 1(144), 134– North Macedonia and its implications for African Swine Fever Transmis- 148. https://doi.org/10.1016/j.prevetmed.2017.06.002 sion. Frontiers in Veterinary Science, 8, 733157. https://doi.org/10.3389/ Cooper, D. S. T., Pereira, A., & da Costa Jong, J. B. (2019). Counting the cost: fvets.2021.733157 The potential impact of African Swine Fever on smallholders in Timor- Ohouko, O. F. H., Koudouvo, K., Dougnon, T. J., Agbonon, A., Karim, I. Y. Leste. One Health, 18(8), 100109. https://doi.org/10.1016/j.onehlt.2019. A., Farougou, S., & Gbeassor, M. (2020). African swine fever in Benin 100109 and prevalence of the disease in Southern Benin: A retrospective study Elbers, A. R., Stegeman, A., Moser, H., Ekker, H. M., Smak, J. A., & Pluimers, F. (2014-2018). Journal of Advanced Veterinary and Animal Research, 7, 464– H. (1999). The classical swine fever epidemic 1997–1998 in the Nether- 470. https://doi.org/10.5455/javar.2020.g442 lands: Descriptive epidemiology. Preventive Veterinary Medicine, 42, 157– OIE. (2021). https://www.oie.int/en/disease/classical-swine-fever/#ui-id-2 184. https://doi.org/10.1016/S0167-5877(99)00074-4 (last accessed 1st December 2021). FAO. (2003). Estimación del impacto de la Peste Porcina Clásica en sistemas Pineda, P., Deluque, A., Peña, M., Diaz, O. L., Allepuz, A., & Casal, J. (2020). productivos porcinos en América Latina: Estudios de casos en tres países lati- Descriptive epidemiology of classical swine fever outbreaks in the period noamericanos – Plan Continental para la Erradicación de la Peste Porcina 2013–2018 in Colombia. Plos One, 15, e0234490. https://doi.org/10. Clásica en las Américas. 1371/journal.pone.0234490 Fernández-Carrión, E., Ivorra, B., Martínez-López, B., Ramos, A. M., & Pineda, P. (2021). Epidemiology of classical swine fever in Colom- Sánchez-Vizcaíno, J. M. (2016). Implementation and validation of an eco- bia. Description of outbreaks, economic impact and epidemiologi- nomic module in the Be-FAST model to predict costs generated by live- cal surveillance in free zones [PhD thesis, Universitat Autònoma de stock disease epidemics: Application to classical swine fever epidemics Barcelona]. in Spain. Preventive Veterinary Medicine, 126, 66–73. https://doi.org/10. Rushton, J. (2009). The economics of animal health and production. Cabi. 1016/j.prevetmed.2016.01.015 Horst, H. S., de Vos, C. J., Tomassen, F. H. M., & Stelwagen, J. (1999). The economic evaluation of control and eradication of epidemic livestock dis- eases. Revue scientifique et technique (OIE), 18, 367–379. https://doi.org/ SUPPORTING INFORMATION 10.20506/rst.18.2.1169 Additional supporting information can be found online in the Support- Huang, Y., Li, J., Zhang, J., & Jin, Z. (2021). Dynamical analysis of the spread ing Information section at the end of this article. of African swine fever with the live pig price in China. Mathematical Bio- sciences and Engineering, 18, 8123–8148. https://doi.org/10.3934/mbe. 2021403 Kivumbi, C. C., Yona, C., Hakizimana, J. N., & Misinzo, G. (2021). An assess- How to cite this article: Casal, J., Tago, D., Pineda, P., ment of the epidemiology and socioeconomic impact of the 2019 African swine fever outbreak in Ngara district, western Tanzania. Veterinary Tabakovski, B., Santos, I., Benigno, C., Huynh, T., Ciaravino, G., & and Animal Science, 24, 14,100198. https://doi.org/10.1016/j.vas.2021. Beltran-Alcrudo, D. (2022). Evaluation of the economic impact 100198 of classical and African swine fever epidemics using OutCosT, a Marsh, T. L., Pendell, D., & Knippenberg, R. (2017). Animal health economics: new spreadsheet-based tool. Transboundary and Emerging An aid to decision making on animal health interventions – Case studies in the United States of America. Revue Scientifique et Technique (OIE), 36, Diseases, 1–11. https://doi.org/10.1111/tbed.14590 137–145.