- Email: [email protected]
Economic analysis of vaccination to control bovine brucellosis in the States of Sao Paulo and Mato Grosso, Brazil
Preventive Veterinary Medicine 118 (2015) 351–358
Contents lists available at ScienceDirect
Preventive Veterinary Medicine journal homepage: www.elsevier.com/locate/prevetmed
Economic analysis of vaccination to control bovine brucellosis in the States of Sao Paulo and Mato Grosso, Brazil A.J.S. Alves a , F. Rocha a , M. Amaku a , F. Ferreira a , E.O. Telles a , J.H.H. Grisi Filho a , J.S. Ferreira Neto a , D. Zylbersztajn b , R.A. Dias a,∗ a b
Department of Preventive Veterinary Medicine and Animal Health, School of Veterinary Medicine, University of Sao Paulo, Brazil Department of Business, School of Economics, Business and Accounting, University of Sao Paulo, Brazil
a r t i c l e
i n f o
Article history: Received 5 March 2014 Received in revised form 17 November 2014 Accepted 10 December 2014 Keywords: Brucellosis Bovine Economic analysis Vaccination
a b s t r a c t Brucellosis is a zoonotic disease that causes important economic losses in Brazil, and the country has therefore established a national program for its control and eradication. Using data generated in the last national brucellosis survey, we conducted an economic analysis in two Brazilian States with different brucellosis status, Mato Grosso (with high prevalence) and Sao Paulo (with low prevalence). The economic analysis was based on the calculation of the additional benefits and costs of controlling bovine brucellosis through the vaccination of heifers aged between 3 and 8 months with S19 vaccine, considering maximal and minimal impacts of the disease. The analysis showed that vaccinating 90% of the replacement heifers aged 3–8 months of age offers the best economic performance in a vaccination program against bovine brucellosis if compared to vaccination rates of 70% and 80%. Moreover, regions with higher prevalences of bovine brucellosis would experience significant economic advantages when implementing a vaccination strategy to control the disease. This economic analysis will allow decision makers to plan more economically effective vaccination programs. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The bacterium Brucella abortus, which is responsible for brucellosis in cattle, is transmitted through abortion products and vaginal discharge. The main symptom of the disease is abortion. Brucella melitensis can also cause brucellosis in cattle, although has not been isolated in Brazil (Poester et al., 2002). Brucellosis is a zoonotic disease, so it can be transmitted from animals to man (Acha and Szyfres, 1986). The negative impacts of brucellosis in livestock include reduced milk production, reduced feed
∗ Corresponding author. Tel.: +55 11 3091 7700; fax: +55 11 3091 7928. E-mail address: [email protected] (R.A. Dias). http://dx.doi.org/10.1016/j.prevetmed.2014.12.010 0167-5877/© 2014 Elsevier B.V. All rights reserved.
conversion, abortion, infertility and mortality in aborting females, perinatal mortality, increased calving intervals and an increased need for animal replacement (Pacheco and Mello, 1956; Sheperd et al., 1980; Faria, 1984; Bernués et al., 1997). Although national programs against brucellosis were established worldwide since 1896, just a few countries have reached the elimination of the circulation of the B. abortus in their herds. With the exception of the Western European countries and Canada, most of them are islands (Paulin and Ferreira Neto, 2003). The National Program for the Control and Eradication of Animal Brucellosis and Tuberculosis (PNCEBT) was established in 2001 by the Brazilian Ministry of Agriculture, Livestock and Food Supply (MAPA), with the aim of reducing the negative impacts of this
352
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
disease on human health and promoting the competitiveness of the national livestock industry (MAPA, 2006). The PNCEBT introduced compulsory vaccination against brucellosis in bovine and buffalo females aged between 3 and 8 months with the S19 vaccine throughout the country and implemented a strategy for certifying brucellosis-free farms (MAPA, 2006). An epidemiological study was conducted in 15 Brazilian States based on sampling farms displaying reproductive activity from 2001 and 2004. Results of that study were published as a special issue of the Brazilian Journal of Veterinary and Animal Sciences (61(1) (2009) 1–141). Most of the control measures are paid for by farmers, such as vaccination of heifers, diagnostic tests and veterinary services. The official veterinary service is only responsible for auditing these activities and for the certification of brucellosis-free farms. Although the vaccination of heifers aged between 3 and 8 months is compulsory in Brazil (except in the State of Santa Catarina), the vaccination coverage is below 100%, and few farms have been certified as being brucellosis free thus, making the demand for official auditing variable and difficult to plan. In decision making on animal health problems, economic methods can be of great support. In this area, also referred as “Economics of Animal Health”, quantification of the economic effects of a disease, optimization of decisions to be made when the disease is present and determination of costs and benefits when preventive measures are being implemented represent important decision-making tools in disease control and eradication programs (Dijkhuizen et al., 1995; Otte and Chilonda, 2000; James, 2004; Vanni et al., 2009). With the major expansion of the Brazilian beef production, losses caused by infectious agents are growing in importance. Increasing organization of the production sector along with the gradual organization of veterinary services has resulted in higher productivity and increased credibility of the country as a beef exporter. However, infectious diseases still circulate in the Brazilian territory, including bovine brucellosis. For the decision to vaccinate in order to control bovine brucellosis, knowledge on the economic effects is useful. In Brazil, a single paper about economic losses caused by brucellosis was identified by the time of writing (Santos et al., 2013), estimating annual losses of US$448 million and the variation of US$78 million for each change of 1% of the prevalence of the disease. Considering that there are still high prevalences of bovine brucellosis in Brazil and vaccination as an effective control strategy for this disease, the aim of the present work was to conduct an economic evaluation of the adoption of brucellosis control measures from the perspective of the private sector, as government participation is restricted to vaccination and certification auditing. We considered only control measures based on vaccination because at the time of writing, most Brazilian States continued to exhibit a high brucellosis prevalence, which does not justify the application of eradication measures.
2. Materials and methods 2.1. Study area For the economic analysis of brucellosis control measures, we selected two Brazilian States in which various factors, such as the epidemiological status of the disease, the structure and operational logistics of the official veterinary services and production systems differ: Sao Paulo (SP) and Mato Grosso (MT) (Table 1). The choice of such different scenarios was made to evaluate the robustness of the results. SP is the most populous state in Brazil (IBGE, 2010). Despite its smaller herd size, it presents significant dairy production and is the main beef exporter in Brazil (IBGE, 2011). SP receives live animals to be slaughtered from other Brazilian States. Moreover, SP has been shown to have low brucellosis prevalence, estimated in 3.8% of females over two years of age, in 2001 (Dias et al., 2009). In MT, the livestock system is mainly extensive, with intermediate to low technology levels, low stocking densities and properties of large areas. This state exhibits the largest livestock herd (IBGE, 2011) and the highest brucellosis prevalence among the Brazilian States, estimated in 10.2% of females over two years of age, in 2003 (Negreiros et al., 2009). Each of the selected states (SP and MT) was considered as a single production unit, and the outcomes were evaluated in terms of the benefits and costs for the meat and milk production chains. 2.2. General structure of the model The model was constructed to calculate two economic indicators (net present value – NPV – and payback period) for different vaccination strategies against brucellosis in two Brazilian States using S19 vaccine in two distinct epidemiological situations of bovine brucellosis. To achieve that, both costs and benefits of the implementation of vaccination strategies were calculated as present values (PV) for different durations of vaccination programs (Assaf Neto, 2012). The mathematical model proposed by Amaku et al. (2009) was used to re-calculate the decrease in prevalence under vaccination efforts of 70%, 80% and 90% to determine the duration of the vaccination programs in both states. The explanation of this model can be found in the Supplementary material. Calculations were performed in Matlab software, version R2013a. According to Amaku et al. (2009), a threshold of 2% prevalence was defined as the criteria for changing the control strategy for eradication. Therefore, the simulations were run until 2% prevalence was reached. The period of time of the shortest vaccination program was used to compare NPV among the vaccination strategies in both states, since different periods of vaccination programs were expected. This fixed time horizon used for the model were during that time, animals would be vaccinated was called “payback period”. The additional benefits achieved with the adoption of each vaccination strategy were compared with respective additional costs. The herd sizes were kept constant to allow the comparison of the effects of the vaccination strategies
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
353
Table 1 Epidemiological characterization of the States of Sao Paulo and Mato Grosso, Brazil. State
SP
Year/reference
MT
Year/reference
Cattle herd Overall mortality rate Overall birth rate Number of farms with reproductive activity Brucellosis prevalence (95% CI) Total road length (km) Area (km2 ) Human population
13,257,952 2.8% 16.95% 159,999 3.81% (0.72–6.90%) 195,078 248,209 41,262,199
2001 (Dias et al., 2009) 2006 (IBGE, 2011) 2006 (IBGE, 2011) 2001 (Dias et al., 2009) 2001 (Dias et al., 2009) 2005 (MT, 2005) 2010 (IBGE, 2010) 2010 (IBGE, 2010)
24,613,718 1.36% 15.93% 82,474 10.25% (7.44–13.06%) 53,810 903,357 3,035,122
2003 (Negreiros et al., 2009) 2006 (IBGE, 2011) 2006 (IBGE, 2011) 2003 (Negreiros et al., 2009) 2001 (Negreiros et al., 2009) 2005 (MT, 2005) 2010 (IBGE, 2010) 2010 (IBGE, 2010)
between the different vaccination rates and periods of vaccination programs. Finally, the years in which samples were collected in the last national brucellosis survey (2001 for SP and 2003 for MT) were considered the years for which all benefits and costs are expressed and discounted to present values. 2.3. Formal description of the model 2.3.1. Calculation of benefits The benefits of the implementation of brucellosis vaccination per year consist of a surplus in meat, milk and animal sale revenue and a reduction in replacement costs. The additional benefits from surplus meat were calculated as the increased carcass weight associated with the decrease in brucellosis prevalence, multiplied by the average price of meat in the year in which the samples were collected in the last national brucellosis survey (base year): 2001 (for SP) and 2003 (for MT) (Dias et al., 2009; Negreiros et al., 2009). The values were re-calculated for each year of the vaccination program, considering the same number of animals being slaughtered. The income from milk was calculated as the increased milk yield associated with the decreased brucellosis prevalence multiplied by the average price of milk in the base year. The values were re-calculated for each year of the vaccination program. The additional benefits from animal sales were calculated as the surplus of born animals associated with a decrease in prevalence, considering the reduction in abortions and peri-natal mortality multiplied by the mean price of calves in the base year. The reduction in replacement costs was calculated as the surplus of cows associated with a decrease in prevalence, due to decreased mortality and sterilization, both abort-related, multiplied by the mean price of cows in the base year. The values were recalculated for each year of the vaccination program according to the current average meat, milk and calf prices. 2.3.2. Calculation of costs The additional costs associated with the brucellosis vaccination program included the costs of the travel of the veterinarians to each farm, the visit itself, the depreciation of the vaccination gun and the vaccine. We assumed that all farms with reproductive activity would be visited, independent of the vaccination effort. To conduct one visit to all farms with reproductive activity, we assumed that all of the roads within each state would have to be traveled, since the exact location of each
farm was not available. To calculate the costs of veterinary travel, this distance (MT, 2005) (Table 1) was multiplied by 30% of the mean gasoline liter price for the base year in each state (DIEESE, 2006) and then multiplied by 2, considering round trips. This multiplier is based on the mean gasoline consumption per kilometer, car maintenance and car price depreciation. The additional costs of veterinary visits were calculated as the number of farms with reproductive activity (Dias et al., 2009; Negreiros et al., 2009) (Table 1) multiplied by 20% of the official Brazilian minimum wage (Reis and Firpo, 2007) for the base year in each state. This is a common practice of budgeting veterinary visits in Brazil. We assumed a 10% depreciation of the value of the vaccination gun per year (Beef Point, 2012) and a single vaccination gun for each farm in each state in the calculations. The number of heifers to be vaccinated was calculated as the number of heifers born that year multiplied by the vaccination rate (70%, 80% or 90%). The number of heifers born was calculated as the number of uninfected calves (both male and female) plus the infected calves that did not die after birth divided by two (assuming half of the calves are female) according to Eq. (1). In general, one vial of the B-19 strain brucellosis vaccine includes 15 doses in Brazil. We assumed a 3% loss of the vial volume during the vaccination process (Costa et al., 2006). Thus, the cost of the vaccine was calculated as the number of heifers to be vaccinated divided by 15 (one vial has 15 doses) multiplied by the price of a vial of the vaccine, multiplied by 1.03 (representing 3% loss of the vaccine during vaccination). 2.3.3. Maintenance of the herd size constant The total herd size for each state was kept constant throughout the duration of the vaccination programs, considering the base year. The herd sizes were 13,257,952 and 24,613,718 animals for SP and MT, respectively (Dias et al., 2009; Negreiros et al., 2009). Based on the number of calves born (IBGE, 2011), we calculated the total number of cows (sexually mature bovine females), according to Eq. (2). The total number of cows is the number of calves born from brucellosis-free cows plus the number of calves born from infected cows (considering that a proportion of the infected cows abort and a proportion of the infected cows that aborted in the previous year became sterile). The abortion rate for maximum and minimal impact of the brucellosis (Table 2) was considered in this calculation. Moreover, a proportion of 20% of the infected cows that becomes sterile
354
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
Table 2 Variation of parameters associated with losses due to brucellosis used in the simulation models of economic impact. Parameter
Reduction of milk production Reduction of meat production Abortion rate Infertility of aborting cows Perinatal mortality Mortality of aborting cows Increase in replacement requirements
Minimum impact of disease 15% 5% 10% 10% 10% 1% 15%
Maximum impact of disease
References
Bernués et al. (1997) Bernués et al. (1997) Sheperd et al. (1980) Pacheco and Mello (1956) Bernués et al. (1997) Bernués et al. (1997) Bernués et al. (1997)
after abortion (Pacheco and Mello, 1956) was considered, resulting that 25% of the calves born from infected cows were born from cows that aborted in the previous year. The birth rate was calculated as the total number of calves born (IBGE, 2011) divided by the total herd size. Although the birth rate was available only for the year 2006 in each state, it was propagated for the entire duration of the vaccination program. The number of births was calculated as the total herd size multiplied by the birth rate in year zero. Beginning in year 1, the calculation considered the number of calves born from uninfected cows plus the calves born from infected non-sterile cows that did not abort according to Eq. (3). The excess numbers of heifers born in each state were regarded as destined for slaughter. The general mortality rate was calculated as the total number of deaths reported (IBGE, 2011) divided by the total herd size. Although the mortality rate was available only for the year 2006 in each state, it was propagated for the entire duration of the vaccination programs. The number of deaths was calculated as the total herd size multiplied by the mortality rate in year zero (t0 ). Beginning in year 1, the number of deaths was calculated as the overall number of deaths in t−1 minus the number of infected cows who died after abortion plus the number of infected non-sterile cows who died in the perinatal period in t−1 minus the number of infected cows who died after abortion plus the number of infected non-sterile cows who died in the perinatal period in t, according to Eq. (4). The abortion rate and the perinatal mortality varied according to the maximal and minimal impact of the disease (Table 2). The mortality of aborting cows was considered as being 1% and the proportion of infertile cows, 20%.
2.3.4. Economic indicators Both costs and benefits were expressed in present values (PV) according to Eq. (5) (CASA, 2010). A constant inflation rate of 8% was considered. The NPV is defined as the sum of the discounted present values of the benefits minus the sum of the discounted present values of the costs (Eq. (6)). The NPV is considered one of the simplest economic criteria for decision making. When NPV > 0, the strategy was considered advantageous, whereas when NPV < 0, the strategy was not advantageous, and when NPV = 0, the strategy was neutral (CASA, 2010). The payback time takes into account the return on the invested capital. This was the period in which the sum of the future cash flows equals the initial investment (RegoBordeaux et al., 2010).
20–25% 10–15% 35% 30% 30% – 30%
References
Pacheco and Mello (1956) Pacheco and Mello (1956), Faria (1984) Sheperd et al. (1980) Pacheco and Mello (1956) Pacheco and Mello (1956) – Faria (1984)
The evolution of the NPVs over the economic lives in both states between maximal and minimal impacts of the disease was represented in graphics made in R, using the ggplot2 package (Wickham, 2009). 2.4. Parameterization of the model 2.4.1. General model assumptions For all simulations, some assumptions were made: maintenance of the herd size, a calving interval of 12 months (one calf per cow per year) and a constant number of animals sold. General mortality and birth rates were extracted from the Brazilian Agriculture Census (IBGE, 2011). 2.4.2. Selection of the base year The base year was the year in which the benefits and costs are discounted to present values. All prices were obtained from official databases (DIEESE, 2006; Reis and Firpo, 2007; CEPEA, 2012a,b; Beef Point, 2012). In the present work, the base year was the year in which the samples were collected in the last national brucellosis survey: 2001 (for SP) and 2003 (for MT) (Dias et al., 2009; Negreiros et al., 2009). Table 3 shows the parameters used in the economic analysis of the brucellosis control measures applied in both states. 2.4.3. Valuation of costs and benefits The additional benefits of the control of brucellosis were quantified based on the reduction of the following losses: decreased milk and meat production and increased abortion rates, infertility rates in aborting females, perinatal mortality, calving intervals, mortality of cows experiencing abortions and cow replacement. In Brazil, the average price of meat is traditionally given for a weight of 15 kg (arroba), which was obtained from CEPEA (2012a), and is provided for each state (Table 3). Milk production values were obtained from IBGE (2011), and the average milk price (per liter) was obtained from CEPEA (2012a) for the base year for each state (Table 3). The mean price of calves was obtained in CEPEA (2012b), for each state in the base year. Based on data found in the literature we have considered the minimum and maximum values for each type of loss caused by brucellosis (Table 2). These parameters were used to calculate the benefits and costs.
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
355
Table 3 Mean prices (in Brazilian R$) of variables used in the simulation models of economic impact of brucellosis in the States of Sao Paulo and Mato Grosso, Brazil, according to the base years. Variable (base year)
SP (2001)
MT (2003)
Reference
Gasoline (per liter) Minimum wage (per month) Calf 15 kg of meat (arroba) Heifer for replacement Milk (per liter) Vaccination gun (per unit)
1.618 180.000 336.000 46.230 1000.000 0.382 150.000
2.145 240.000 376.860 60.370 1000.000 0.382 150.000
DIEESE (2006) Reis and Firpo (2007) CEPEA (2012b) CEPEA (2012a) Beef Point (2012) Beef Point (2012) Beef Point (2012)
3. Results The results are summarized in Table 4. The NPVs were compared based on shortest vaccination program: 12 years for SP and 22 years for MT. We have observed that maximal brucellosis impact produces higher NPVs in both states, when compared to the NPVs of the minimal impacts. These differences vary from 38% for 90% vaccination rate to 43%, for 70% in SP and from 40% for 90% vaccination rate to 41%, for 70% in MT. Moreover, when the vaccination rates are higher, NPVs are also higher. In SP, the highest NPV (R$351 million) was observed for the 90% vaccination rate under maximal impact of the disease, which was 29% higher than the NPV for 70% vaccination rate under maximal impact of the disease (R$271 million). In MT, the higher NPV (R$4.7 billion) was also for the 90% vaccination rate under maximal impact of the disease, which was only 17% higher than the NPV for 70% vaccination rate under maximal impact of the disease (R$4 billion). The magnitude of NPVs for MT was billions of Brazilian R$, but in SP, of millions of R$. A comparison of the NPVs of each vaccination strategy (70%, 80% and 90%) for each state during the vaccination programs (in years) was represented in Figs. 1 and 2, for SP and MT, respectively. A comparison between both states was represented in Fig. 3.
Fig. 2. Estimation of the variation of the net present values (NPV) of different vaccination rates (70%, 80% and 90%) between maximal and minimal impacts of brucellosis during the economic lives (in years) in the State of Mato Grosso, Brazil.
Fig. 3. Estimation of the variation of the net present values (NPV) of different vaccination rates (70%, 80% and 90%) between maximal and minimal impacts of brucellosis during the economic lives (in years) in the States of Sao Paulo and Mato Grosso, Brazil.
Fig. 1. Estimation of the variation of the net present values (NPV) of different vaccination rates (70%, 80% and 90%) between maximal and minimal impacts of brucellosis during the economic lives (in years) in the State of Sao Paulo, Brazil.
Higher vaccination rates reduce the time to reach 2% of brucellosis prevalence, resulting in shorter vaccination programs and costs. In SP, it would take 12 years to reduce the prevalence of brucellosis from 3.81% to 2% under a 90% vaccination rate and 14 years under 70% vaccination rate. The cost of the first strategy was only 90% of the cost of the
356
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
Table 4 Economic indicators (NPV and payback) of different vaccination rates (70%, 80%, 90%) under maximal and minimal impacts of brucellosis in the States of Sao Paulo and Mato Grosso, Brazil. State
SP
MT
a b c
Vaccination rate
70%
80%
Disease impact
Minimal
Maximal
Minimal
Maximal
Minimal
NPVb (R$1000) Time (years)a Payback Cost (R$1000)a
189,659
271,309
223,040
313,279
254,943
NPVc (R$1000) Time (years)a Payback Cost (R$1000)a
2,851,484
14 7
13 6
7
147,108 4,009,812
3,113,689
6
6
4,374,665
3,349,784
24 2
170,922
3
351,315 6 132,952 4,703,312 22
2 155,349
Maximal 12
139,010
27 3
90%
2
2 145,248
To reach 2% prevalence. 12 years. 22 years.
last. In MT, it would take 22 years to reduce the prevalence of brucellosis from 10.25% to 2% under a 90% vaccination rate and 27 years under 70% vaccination rate. The cost of the first strategy was only 85% of the cost of the last. The payback period is shorter with higher vaccination rates, especially when the impacts of the disease are maximal. In SP, the payback was six years for 90% vaccination rate (for both in maximal and minimal impact of the disease), six years for maximal impact of the disease and seven years, for minimal impact for 80% and 70% vaccination rates. In MT, the payback period was two years for 90% vaccination rate (for both in maximal and minimal impact of the disease), two years for maximal impact of the disease and three years, for minimal impact for 80% and 70% vaccination rates. The payback can also be verified when the cash flows are calculated for each state. The calculation of the cash flow can be found in the Supplementary material. The biggest cost items in the 90% vaccination rate, for example, are the vet visits (58% in SP and 65% in MT), followed by vaccination gun acquisition (38% in SP and 27% in MT). Obviously, the costs with vaccine are higher in MT (7%) than SP (3%). The costs with transport were small, being around 1% in both states. 4. Discussion Brucellosis is a zoonotic disease and has impact on human health, especially professionals who deal with cattle, like veterinarians and slaughterers, creating extra costs. Although it would be interesting to include such information in the simulation models, human brucellosis is not a notifiable disease in Brazil and no official data was available in the time of writing. The choice of two different study areas allowed the comparison of the economic indicators in two different epidemiological situations: SP with almost half of the herd size and almost a third of the prevalence of MT. Moreover, SP had a significantly higher number of farms (twice than MT) and despite its smaller area, has more roads to travel than MT (four-fold). These factors have made the costs of the implementation of any vaccination strategy in SP higher than MT, since more equipment including vaccination guns require purchase and more veterinary visits are required, generating higher NPVs and longer payback periods in SP
than in MT. Veterinary visits were the costliest item in both states and even with a lower proportional participation in SP, the nominal value was bigger than MT. The only item that was more costly in MT, compared to SP was the vaccine vials, since the herd of MT is bigger. This item showed low impact in the overall costs in both states. Annual vaccination efforts corresponding to 70%, 80% and 90% of the heifers aged between 3 and 8 months were chosen to examine an interval around the cut-off of 80%, empirically recommended by many authors as the optimum rate of vaccination for heifers (Paulin and Ferreira Neto, 2003; Dias et al., 2009; Negreiros et al., 2009). Considering the expected decrease in prevalence over time, periods greater than 12 years and 22 years would be necessary to decrease the prevalence below 2% in SP and MT respectively. Thus, under economic lives of 12 years and 22 years, a vaccination rate of 90% is most advantageous economically in SP and MT respectively. Although the results of the present work corroborate empirical recommendations to vaccinate above 80% of the bovine heifer population aged between 3 and 8 months, based on observations made in countries and regions that have implemented vaccination programs (Paulin and Ferreira Neto, 2003), it is noteworthy that different lengths of vaccination programs are essential to prevent economic losses in different prevalence scenarios. Moreover, NPV curve may differ under different initial prevalence scenarios, which were higher for MT, that presented initial brucellosis prevalence significantly higher than SP, as observed in Fig. 3. This conclusion is corroborated by Amaku et al. (2009), who estimated the reduction of the prevalence of brucellosis associated with different vaccination efforts. These authors indicated that vaccination rates between 70% and 90% would reduce the prevalence of brucellosis to 2% in very similar periods of time, but the logistics and resources could vary significantly in each state. Paulin and Ferreira Neto (2003) indicated that prevalence below 2% cannot be reached using vaccination with S19 alone, since this strain does not protect 100% of the vaccinated heifers aged between 3 and 8 months and reintroductions from brucellosis positive wild and domestic animals can occur. Under this scenario, eradication measures such as mass diagnostic screening and culling of positive animals should be implemented along with vaccination strategies using S19
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
or RB51 vaccines in females not vaccinated between 3 and 8 months of age. Animal movements and the incorporation of breeding animals into the herd should only be completed after screening tests. If a vaccination program is conducted along with these measures, the decrease in prevalence may be greater. For that reason, the eradication strategy is expected to be more costly than vaccination strategies, and the slope of NPV curves would be lower over time. In the present work, we have observed that NPV is higher when vaccination rates are higher as well, especially when the impacts of the disease are maximal, in both states. At the time of writing, no recent data about brucellosis driven losses were found, since most of the countries that usually produced such data are now free from the disease. For that reason, it is reasonable to expect that these parameter may fall between the extreme values (minimal and maximal) for the disease impacts shown in Table 2, but these parameter should have been measured prior to implementation of these models in both states. It is also important to note that the magnitude of losses may influence the performance of a vaccination program, and biological parameters such as abortion rates, perinatal mortality, and the infertility or mortality of aborting cows should therefore be assessed prior to the adoption of control measures. An alternative would be to incorporate the statistical distributions of the parameters associated with losses due to brucellosis, allowing probabilistic assessment of such losses. The results of the present work differ from the results found by Santos et al. (2013), who calculated the losses due to brucellosis in a quite similar way, but ignored the costs of vaccination, making a linear regression of actual economic losses by cattle head and brucellosis prevalence to predict the effects of the reduction of prevalence in the reduction of losses. This approach is unrealistic since brucellosis control actions based in vaccination may differ significantly in different states, due to the geographic distribution of farms, road lengths, and heifer populations, addressed in the present work. Moreover, the NPV and payback estimations (not calculated by Santos et al., 2013) make it easy to compare different vaccination efforts, which is a more important issue to implement a vaccination program than knowing the amount of losses with the change of 1% of prevalence. Some assumptions and simplifications may have affected the results of the simulations, but there were no data to support another approach. For example, the calving interval of one year, which is very useful for the simulations, may decrease with the decrease in brucellosis prevalence, since abortions and mortality due to brucellosis are expected to diminish. The mortality rate by female age group was not available by the time of writing. Considering that brucellosis prevention and control measures are different in each age group, that would be a key point to be accessed in future simulation models. Keeping the inflation constant may not be realistic, but the mean inflation rate in Brazil since 2011 was around 8%, the same value of the discount rate (r). The additional costs of nutrient demand, which are expected to increase with higher production following brucellosis control were not considered, even though the main production system of livestock in Brazil is extensive grazing. Finally, we have not considered
357
the price decrease arising from the increased commodities (meat and milk) produced with the reduction of brucellosis prevalence, although we have assumed that the impact of the commodity price from the selected states do not impact the commodities price nationally. This is due to the fact that these products are not consumed locally in each state, with there being competition from other states and countries. The distribution of costs and benefits between public and private sectors is very difficult to measure, even knowing that most of the prevention and control measures are paid by the farmers and the participation of the federal and state governments is only in auditing the actions. The benefits would certainly be shared by both private and public sectors, since an increased tax collection is expected. 5. Conclusion Regions with higher prevalence of brucellosis and greater herds would experience significant economic advantages with the adoption of vaccination programs, whereas in regions with low prevalence and smaller herds, such advantages would be lower, but these measures would nevertheless be worthwhile. Economic analyses should be considered in the planning of control measures for brucellosis, especially in a country as diverse as Brazil. Strategies cannot be standardized for the entire country, which may cause economic losses associated with the adoption of strategies that are sub-optimal locally. For this reason, the National Program for the Control and Eradication of Animal Brucellosis and Tuberculosis (PNCEBT) should consider prevalence and production chain differences to recommend specific control measures. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. prevetmed.2014.12.010. References Acha, P.N., Szyfres, B., 1986. Zoonosis y enfermedades transmissibles communes al hombre y a los animales, 2nd ed. Organización Panamericana de la Salud, Washington. Amaku, M., Dias, R.A., Ferreira Neto, J.S., Ferreira, F., 2009. Modelagem matemática do controle da brucelose bovina por vacinac¸ão. Arq. Bras. Med. Vet. Zoo. 61 (1), 135–141. Assaf Neto, A., 2012. Matemática financeira e suas aplicac¸ões, 12th ed. Atlas, São Paulo. Beef Point, 2012. Cotac¸ões. Beef Point, Piracicaba, Available from: http://www.beefpoint.com.br/ (accessed 20.12.13). Bernués, A., Manrique, E., Maza, M.T., 1997. Economic evaluation of bovine brucellosis and tuberculosis eradication programmes in a mountain area of Spain. Prev. Vet. Med. 30 (2), 137–149. Centro de Estudos Avanc¸ados em Economia Aplicada [CEPEA], 2012a. Fed Cattle Price Index. CEPEA, Piracicaba, Available from: http://www.cepea.esalq.usp.br/english/cattle/ (accessed 20.12.13). Centro de Estudos Avanc¸ados em Economia Aplicada [CEPEA], 2012b. Calf Price Index. CEPEA, Piracicaba, Available from: http://cepea.esalq.usp.br/english/calf/(accessed 20.12.13). Civil Aviation Safety Authority [CASA], 2010. Cost Benefit Analysis Procedures Manual. CASA, Canberra, Available from: http://www.casa.gov.au (accessed 20.12.13). Costa, M.J.R.P., Toledo, L.M., Schmidek, A., 2006. Boas práticas de manejo – vacinac¸ão. FUNEP, Jaboticabal. Departamento Intersindical de Estatística e Estudos Socioeconômicos [DIEESE], 2006. Nota técnica 19. DIEESE, São Paulo.
358
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
Dias, R.A., Gonc¸alves, V.S.P., Figueiredo, V.C.F., Lôbo, J.R., Lima, Z.M.B., Paulin, L.M.S., Gunnewiek, M.F.K., Amaku, M., Ferreira Neto, J.S., Ferreira, F., 2009. Situac¸ão epidemiológica da brucelose bovina no Estado de São Paulo. Arq. Bras. Med. Vet. Zoo. 61 (1), 118–125. Dijkhuizen, A.A., Huirne, R.B.M., Jalvingh, A.W., 1995. Economic analysis of animal diseases and their control. Prev. Vet. Med. 25 (2), 135–150. Faria, J.F., 1984. Situac¸ão da Brucelose no Brasil. In: Comunicac¸ão Científica da Faculdade de Medicina Veterinária e Zootecnia da Universidade de São Paulo 8, pp. 161–175. Instituto Brasileiro de Geografia e Estatística [IBGE], 2010. Censo populacional de 2010. IBGE, Brasília, Available from: http://ibge/home/estatistica/populacao/censo2010/default.shtm (accessed 20.12.13). Instituto Brasileiro de Geografia e Estatística [IBGE], 2011. Sistema IBGE de ecuperac¸ão automática [SIDRA]. IBGE, Brasília, Available from: http://www.sidra.ibge.gov.br (accessed 20.12.13). James, A., 2004. The use of economic analysis to define animal health the policies. In: Compendium of technical items presented to the International Committee or to the regional commission, pp. 7–12, Available at: http://www.oie.int/doc/ged/D677.pdf Ministério da Agricultura, Pecuária e Abastecimento [MAPA], 2006. Programa Nacional de Controle e Erradicac¸ão da Brucelose e Tuberculose. MAPA, Brasília, Available from: http://www.agricultura.gov.br (accessed 20.12.13). Ministério dos Transportes [MT], 2005. Anuário estatístico dos transportes terrestres. MT, Brasília. Negreiros, R.L., Dias, R.A., Ferreira, F., Ferreira Neto, J.S., Gonc¸alves, V.S.P., Silva, M.C.P., Figueiredo, V.C.F., Lôbo, J.R., Freitas, J., Amaku, M., 2009. Situac¸ão epidemiológica da brucelose bovina no Estado de Mato Grosso. Arq. Bras. Med. Vet. Zoo. 61 (1), 56–65.
Otte, M.J., Chilonda, P., 2000. Animal Health Economies: An Introduction. Animal Production and Healthy Division (AGA), FAO, Rome. Pacheco, G., Mello, M.T., 1956. Brucelose. Distribuidora Livraria Ateneu, Rio de Janeiro. Paulin, L.M., Ferreira Neto, J.S., 2003. O Combate à brucelose bovina: situac¸ão brasileira. FUNEP, Jaboticabal. Poester, F.P., Gonc¸alves, V.S.P., Lage, A.P., 2002. Brucellosis in Brazil. Vet. Microbiol. 90, 55–62. Rego-Bordeaux, R., Paulo, G.P., Spritzer, I.M.P.A., Zotes, L.P., 2010. Viabilidade econômica financeira de projetos, 3rd ed. Fundac¸ão Getulio Vargas, Rio de Janeiro. Reis, M.C., Firpo, S., 2007. O salário mínimo e a queda recente da desigualdade no Brasil. In: Barros, R.P., Foguel, M.N., Ulyssea, G. (Eds.), Desigualdade de renda no Brasil: uma análise da queda recente, vol. 2. Instituto de Pesquisa Econômica Aplicada, Brasília, Available from: http://www.ipea.gov.br/portal/images/stories/PDFs/livros/Cap33.pdf (accessed 20.12.13). Santos, R.L., Martins, T.M., Borges, A.M., Paixão, T.A., 2013. Economic losses due to bovine brucellosis in Brazil. Pesq. Vet. Bras. 33 (6), 759–764. Sheperd, A.A., Simpsom, B.H., Davidson, R.M., 1980. An economic evaluation of the New Zealand bovine brucellosis eradication. In: Proceedings of the Second International Symposium on Veterinary Epidemiology and Economics, pp. 443–447. Vanni, T., Luz, P.M., Ribeiro, R.A., Novaes, H.M.D., Polanczyk, C.A., 2009. Avaliac¸ão econômica em Saúde: aplicac¸ões em doenc¸as infecciosas. Cad. Saude Publica 25 (12), 2543–2552. Wickham, H., 2009. ggplot2: Elegant Graphics for Data Analysis. Springer, New York.
Contents lists available at ScienceDirect
Preventive Veterinary Medicine journal homepage: www.elsevier.com/locate/prevetmed
Economic analysis of vaccination to control bovine brucellosis in the States of Sao Paulo and Mato Grosso, Brazil A.J.S. Alves a , F. Rocha a , M. Amaku a , F. Ferreira a , E.O. Telles a , J.H.H. Grisi Filho a , J.S. Ferreira Neto a , D. Zylbersztajn b , R.A. Dias a,∗ a b
Department of Preventive Veterinary Medicine and Animal Health, School of Veterinary Medicine, University of Sao Paulo, Brazil Department of Business, School of Economics, Business and Accounting, University of Sao Paulo, Brazil
a r t i c l e
i n f o
Article history: Received 5 March 2014 Received in revised form 17 November 2014 Accepted 10 December 2014 Keywords: Brucellosis Bovine Economic analysis Vaccination
a b s t r a c t Brucellosis is a zoonotic disease that causes important economic losses in Brazil, and the country has therefore established a national program for its control and eradication. Using data generated in the last national brucellosis survey, we conducted an economic analysis in two Brazilian States with different brucellosis status, Mato Grosso (with high prevalence) and Sao Paulo (with low prevalence). The economic analysis was based on the calculation of the additional benefits and costs of controlling bovine brucellosis through the vaccination of heifers aged between 3 and 8 months with S19 vaccine, considering maximal and minimal impacts of the disease. The analysis showed that vaccinating 90% of the replacement heifers aged 3–8 months of age offers the best economic performance in a vaccination program against bovine brucellosis if compared to vaccination rates of 70% and 80%. Moreover, regions with higher prevalences of bovine brucellosis would experience significant economic advantages when implementing a vaccination strategy to control the disease. This economic analysis will allow decision makers to plan more economically effective vaccination programs. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The bacterium Brucella abortus, which is responsible for brucellosis in cattle, is transmitted through abortion products and vaginal discharge. The main symptom of the disease is abortion. Brucella melitensis can also cause brucellosis in cattle, although has not been isolated in Brazil (Poester et al., 2002). Brucellosis is a zoonotic disease, so it can be transmitted from animals to man (Acha and Szyfres, 1986). The negative impacts of brucellosis in livestock include reduced milk production, reduced feed
∗ Corresponding author. Tel.: +55 11 3091 7700; fax: +55 11 3091 7928. E-mail address: [email protected] (R.A. Dias). http://dx.doi.org/10.1016/j.prevetmed.2014.12.010 0167-5877/© 2014 Elsevier B.V. All rights reserved.
conversion, abortion, infertility and mortality in aborting females, perinatal mortality, increased calving intervals and an increased need for animal replacement (Pacheco and Mello, 1956; Sheperd et al., 1980; Faria, 1984; Bernués et al., 1997). Although national programs against brucellosis were established worldwide since 1896, just a few countries have reached the elimination of the circulation of the B. abortus in their herds. With the exception of the Western European countries and Canada, most of them are islands (Paulin and Ferreira Neto, 2003). The National Program for the Control and Eradication of Animal Brucellosis and Tuberculosis (PNCEBT) was established in 2001 by the Brazilian Ministry of Agriculture, Livestock and Food Supply (MAPA), with the aim of reducing the negative impacts of this
352
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
disease on human health and promoting the competitiveness of the national livestock industry (MAPA, 2006). The PNCEBT introduced compulsory vaccination against brucellosis in bovine and buffalo females aged between 3 and 8 months with the S19 vaccine throughout the country and implemented a strategy for certifying brucellosis-free farms (MAPA, 2006). An epidemiological study was conducted in 15 Brazilian States based on sampling farms displaying reproductive activity from 2001 and 2004. Results of that study were published as a special issue of the Brazilian Journal of Veterinary and Animal Sciences (61(1) (2009) 1–141). Most of the control measures are paid for by farmers, such as vaccination of heifers, diagnostic tests and veterinary services. The official veterinary service is only responsible for auditing these activities and for the certification of brucellosis-free farms. Although the vaccination of heifers aged between 3 and 8 months is compulsory in Brazil (except in the State of Santa Catarina), the vaccination coverage is below 100%, and few farms have been certified as being brucellosis free thus, making the demand for official auditing variable and difficult to plan. In decision making on animal health problems, economic methods can be of great support. In this area, also referred as “Economics of Animal Health”, quantification of the economic effects of a disease, optimization of decisions to be made when the disease is present and determination of costs and benefits when preventive measures are being implemented represent important decision-making tools in disease control and eradication programs (Dijkhuizen et al., 1995; Otte and Chilonda, 2000; James, 2004; Vanni et al., 2009). With the major expansion of the Brazilian beef production, losses caused by infectious agents are growing in importance. Increasing organization of the production sector along with the gradual organization of veterinary services has resulted in higher productivity and increased credibility of the country as a beef exporter. However, infectious diseases still circulate in the Brazilian territory, including bovine brucellosis. For the decision to vaccinate in order to control bovine brucellosis, knowledge on the economic effects is useful. In Brazil, a single paper about economic losses caused by brucellosis was identified by the time of writing (Santos et al., 2013), estimating annual losses of US$448 million and the variation of US$78 million for each change of 1% of the prevalence of the disease. Considering that there are still high prevalences of bovine brucellosis in Brazil and vaccination as an effective control strategy for this disease, the aim of the present work was to conduct an economic evaluation of the adoption of brucellosis control measures from the perspective of the private sector, as government participation is restricted to vaccination and certification auditing. We considered only control measures based on vaccination because at the time of writing, most Brazilian States continued to exhibit a high brucellosis prevalence, which does not justify the application of eradication measures.
2. Materials and methods 2.1. Study area For the economic analysis of brucellosis control measures, we selected two Brazilian States in which various factors, such as the epidemiological status of the disease, the structure and operational logistics of the official veterinary services and production systems differ: Sao Paulo (SP) and Mato Grosso (MT) (Table 1). The choice of such different scenarios was made to evaluate the robustness of the results. SP is the most populous state in Brazil (IBGE, 2010). Despite its smaller herd size, it presents significant dairy production and is the main beef exporter in Brazil (IBGE, 2011). SP receives live animals to be slaughtered from other Brazilian States. Moreover, SP has been shown to have low brucellosis prevalence, estimated in 3.8% of females over two years of age, in 2001 (Dias et al., 2009). In MT, the livestock system is mainly extensive, with intermediate to low technology levels, low stocking densities and properties of large areas. This state exhibits the largest livestock herd (IBGE, 2011) and the highest brucellosis prevalence among the Brazilian States, estimated in 10.2% of females over two years of age, in 2003 (Negreiros et al., 2009). Each of the selected states (SP and MT) was considered as a single production unit, and the outcomes were evaluated in terms of the benefits and costs for the meat and milk production chains. 2.2. General structure of the model The model was constructed to calculate two economic indicators (net present value – NPV – and payback period) for different vaccination strategies against brucellosis in two Brazilian States using S19 vaccine in two distinct epidemiological situations of bovine brucellosis. To achieve that, both costs and benefits of the implementation of vaccination strategies were calculated as present values (PV) for different durations of vaccination programs (Assaf Neto, 2012). The mathematical model proposed by Amaku et al. (2009) was used to re-calculate the decrease in prevalence under vaccination efforts of 70%, 80% and 90% to determine the duration of the vaccination programs in both states. The explanation of this model can be found in the Supplementary material. Calculations were performed in Matlab software, version R2013a. According to Amaku et al. (2009), a threshold of 2% prevalence was defined as the criteria for changing the control strategy for eradication. Therefore, the simulations were run until 2% prevalence was reached. The period of time of the shortest vaccination program was used to compare NPV among the vaccination strategies in both states, since different periods of vaccination programs were expected. This fixed time horizon used for the model were during that time, animals would be vaccinated was called “payback period”. The additional benefits achieved with the adoption of each vaccination strategy were compared with respective additional costs. The herd sizes were kept constant to allow the comparison of the effects of the vaccination strategies
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
353
Table 1 Epidemiological characterization of the States of Sao Paulo and Mato Grosso, Brazil. State
SP
Year/reference
MT
Year/reference
Cattle herd Overall mortality rate Overall birth rate Number of farms with reproductive activity Brucellosis prevalence (95% CI) Total road length (km) Area (km2 ) Human population
13,257,952 2.8% 16.95% 159,999 3.81% (0.72–6.90%) 195,078 248,209 41,262,199
2001 (Dias et al., 2009) 2006 (IBGE, 2011) 2006 (IBGE, 2011) 2001 (Dias et al., 2009) 2001 (Dias et al., 2009) 2005 (MT, 2005) 2010 (IBGE, 2010) 2010 (IBGE, 2010)
24,613,718 1.36% 15.93% 82,474 10.25% (7.44–13.06%) 53,810 903,357 3,035,122
2003 (Negreiros et al., 2009) 2006 (IBGE, 2011) 2006 (IBGE, 2011) 2003 (Negreiros et al., 2009) 2001 (Negreiros et al., 2009) 2005 (MT, 2005) 2010 (IBGE, 2010) 2010 (IBGE, 2010)
between the different vaccination rates and periods of vaccination programs. Finally, the years in which samples were collected in the last national brucellosis survey (2001 for SP and 2003 for MT) were considered the years for which all benefits and costs are expressed and discounted to present values. 2.3. Formal description of the model 2.3.1. Calculation of benefits The benefits of the implementation of brucellosis vaccination per year consist of a surplus in meat, milk and animal sale revenue and a reduction in replacement costs. The additional benefits from surplus meat were calculated as the increased carcass weight associated with the decrease in brucellosis prevalence, multiplied by the average price of meat in the year in which the samples were collected in the last national brucellosis survey (base year): 2001 (for SP) and 2003 (for MT) (Dias et al., 2009; Negreiros et al., 2009). The values were re-calculated for each year of the vaccination program, considering the same number of animals being slaughtered. The income from milk was calculated as the increased milk yield associated with the decreased brucellosis prevalence multiplied by the average price of milk in the base year. The values were re-calculated for each year of the vaccination program. The additional benefits from animal sales were calculated as the surplus of born animals associated with a decrease in prevalence, considering the reduction in abortions and peri-natal mortality multiplied by the mean price of calves in the base year. The reduction in replacement costs was calculated as the surplus of cows associated with a decrease in prevalence, due to decreased mortality and sterilization, both abort-related, multiplied by the mean price of cows in the base year. The values were recalculated for each year of the vaccination program according to the current average meat, milk and calf prices. 2.3.2. Calculation of costs The additional costs associated with the brucellosis vaccination program included the costs of the travel of the veterinarians to each farm, the visit itself, the depreciation of the vaccination gun and the vaccine. We assumed that all farms with reproductive activity would be visited, independent of the vaccination effort. To conduct one visit to all farms with reproductive activity, we assumed that all of the roads within each state would have to be traveled, since the exact location of each
farm was not available. To calculate the costs of veterinary travel, this distance (MT, 2005) (Table 1) was multiplied by 30% of the mean gasoline liter price for the base year in each state (DIEESE, 2006) and then multiplied by 2, considering round trips. This multiplier is based on the mean gasoline consumption per kilometer, car maintenance and car price depreciation. The additional costs of veterinary visits were calculated as the number of farms with reproductive activity (Dias et al., 2009; Negreiros et al., 2009) (Table 1) multiplied by 20% of the official Brazilian minimum wage (Reis and Firpo, 2007) for the base year in each state. This is a common practice of budgeting veterinary visits in Brazil. We assumed a 10% depreciation of the value of the vaccination gun per year (Beef Point, 2012) and a single vaccination gun for each farm in each state in the calculations. The number of heifers to be vaccinated was calculated as the number of heifers born that year multiplied by the vaccination rate (70%, 80% or 90%). The number of heifers born was calculated as the number of uninfected calves (both male and female) plus the infected calves that did not die after birth divided by two (assuming half of the calves are female) according to Eq. (1). In general, one vial of the B-19 strain brucellosis vaccine includes 15 doses in Brazil. We assumed a 3% loss of the vial volume during the vaccination process (Costa et al., 2006). Thus, the cost of the vaccine was calculated as the number of heifers to be vaccinated divided by 15 (one vial has 15 doses) multiplied by the price of a vial of the vaccine, multiplied by 1.03 (representing 3% loss of the vaccine during vaccination). 2.3.3. Maintenance of the herd size constant The total herd size for each state was kept constant throughout the duration of the vaccination programs, considering the base year. The herd sizes were 13,257,952 and 24,613,718 animals for SP and MT, respectively (Dias et al., 2009; Negreiros et al., 2009). Based on the number of calves born (IBGE, 2011), we calculated the total number of cows (sexually mature bovine females), according to Eq. (2). The total number of cows is the number of calves born from brucellosis-free cows plus the number of calves born from infected cows (considering that a proportion of the infected cows abort and a proportion of the infected cows that aborted in the previous year became sterile). The abortion rate for maximum and minimal impact of the brucellosis (Table 2) was considered in this calculation. Moreover, a proportion of 20% of the infected cows that becomes sterile
354
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
Table 2 Variation of parameters associated with losses due to brucellosis used in the simulation models of economic impact. Parameter
Reduction of milk production Reduction of meat production Abortion rate Infertility of aborting cows Perinatal mortality Mortality of aborting cows Increase in replacement requirements
Minimum impact of disease 15% 5% 10% 10% 10% 1% 15%
Maximum impact of disease
References
Bernués et al. (1997) Bernués et al. (1997) Sheperd et al. (1980) Pacheco and Mello (1956) Bernués et al. (1997) Bernués et al. (1997) Bernués et al. (1997)
after abortion (Pacheco and Mello, 1956) was considered, resulting that 25% of the calves born from infected cows were born from cows that aborted in the previous year. The birth rate was calculated as the total number of calves born (IBGE, 2011) divided by the total herd size. Although the birth rate was available only for the year 2006 in each state, it was propagated for the entire duration of the vaccination program. The number of births was calculated as the total herd size multiplied by the birth rate in year zero. Beginning in year 1, the calculation considered the number of calves born from uninfected cows plus the calves born from infected non-sterile cows that did not abort according to Eq. (3). The excess numbers of heifers born in each state were regarded as destined for slaughter. The general mortality rate was calculated as the total number of deaths reported (IBGE, 2011) divided by the total herd size. Although the mortality rate was available only for the year 2006 in each state, it was propagated for the entire duration of the vaccination programs. The number of deaths was calculated as the total herd size multiplied by the mortality rate in year zero (t0 ). Beginning in year 1, the number of deaths was calculated as the overall number of deaths in t−1 minus the number of infected cows who died after abortion plus the number of infected non-sterile cows who died in the perinatal period in t−1 minus the number of infected cows who died after abortion plus the number of infected non-sterile cows who died in the perinatal period in t, according to Eq. (4). The abortion rate and the perinatal mortality varied according to the maximal and minimal impact of the disease (Table 2). The mortality of aborting cows was considered as being 1% and the proportion of infertile cows, 20%.
2.3.4. Economic indicators Both costs and benefits were expressed in present values (PV) according to Eq. (5) (CASA, 2010). A constant inflation rate of 8% was considered. The NPV is defined as the sum of the discounted present values of the benefits minus the sum of the discounted present values of the costs (Eq. (6)). The NPV is considered one of the simplest economic criteria for decision making. When NPV > 0, the strategy was considered advantageous, whereas when NPV < 0, the strategy was not advantageous, and when NPV = 0, the strategy was neutral (CASA, 2010). The payback time takes into account the return on the invested capital. This was the period in which the sum of the future cash flows equals the initial investment (RegoBordeaux et al., 2010).
20–25% 10–15% 35% 30% 30% – 30%
References
Pacheco and Mello (1956) Pacheco and Mello (1956), Faria (1984) Sheperd et al. (1980) Pacheco and Mello (1956) Pacheco and Mello (1956) – Faria (1984)
The evolution of the NPVs over the economic lives in both states between maximal and minimal impacts of the disease was represented in graphics made in R, using the ggplot2 package (Wickham, 2009). 2.4. Parameterization of the model 2.4.1. General model assumptions For all simulations, some assumptions were made: maintenance of the herd size, a calving interval of 12 months (one calf per cow per year) and a constant number of animals sold. General mortality and birth rates were extracted from the Brazilian Agriculture Census (IBGE, 2011). 2.4.2. Selection of the base year The base year was the year in which the benefits and costs are discounted to present values. All prices were obtained from official databases (DIEESE, 2006; Reis and Firpo, 2007; CEPEA, 2012a,b; Beef Point, 2012). In the present work, the base year was the year in which the samples were collected in the last national brucellosis survey: 2001 (for SP) and 2003 (for MT) (Dias et al., 2009; Negreiros et al., 2009). Table 3 shows the parameters used in the economic analysis of the brucellosis control measures applied in both states. 2.4.3. Valuation of costs and benefits The additional benefits of the control of brucellosis were quantified based on the reduction of the following losses: decreased milk and meat production and increased abortion rates, infertility rates in aborting females, perinatal mortality, calving intervals, mortality of cows experiencing abortions and cow replacement. In Brazil, the average price of meat is traditionally given for a weight of 15 kg (arroba), which was obtained from CEPEA (2012a), and is provided for each state (Table 3). Milk production values were obtained from IBGE (2011), and the average milk price (per liter) was obtained from CEPEA (2012a) for the base year for each state (Table 3). The mean price of calves was obtained in CEPEA (2012b), for each state in the base year. Based on data found in the literature we have considered the minimum and maximum values for each type of loss caused by brucellosis (Table 2). These parameters were used to calculate the benefits and costs.
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
355
Table 3 Mean prices (in Brazilian R$) of variables used in the simulation models of economic impact of brucellosis in the States of Sao Paulo and Mato Grosso, Brazil, according to the base years. Variable (base year)
SP (2001)
MT (2003)
Reference
Gasoline (per liter) Minimum wage (per month) Calf 15 kg of meat (arroba) Heifer for replacement Milk (per liter) Vaccination gun (per unit)
1.618 180.000 336.000 46.230 1000.000 0.382 150.000
2.145 240.000 376.860 60.370 1000.000 0.382 150.000
DIEESE (2006) Reis and Firpo (2007) CEPEA (2012b) CEPEA (2012a) Beef Point (2012) Beef Point (2012) Beef Point (2012)
3. Results The results are summarized in Table 4. The NPVs were compared based on shortest vaccination program: 12 years for SP and 22 years for MT. We have observed that maximal brucellosis impact produces higher NPVs in both states, when compared to the NPVs of the minimal impacts. These differences vary from 38% for 90% vaccination rate to 43%, for 70% in SP and from 40% for 90% vaccination rate to 41%, for 70% in MT. Moreover, when the vaccination rates are higher, NPVs are also higher. In SP, the highest NPV (R$351 million) was observed for the 90% vaccination rate under maximal impact of the disease, which was 29% higher than the NPV for 70% vaccination rate under maximal impact of the disease (R$271 million). In MT, the higher NPV (R$4.7 billion) was also for the 90% vaccination rate under maximal impact of the disease, which was only 17% higher than the NPV for 70% vaccination rate under maximal impact of the disease (R$4 billion). The magnitude of NPVs for MT was billions of Brazilian R$, but in SP, of millions of R$. A comparison of the NPVs of each vaccination strategy (70%, 80% and 90%) for each state during the vaccination programs (in years) was represented in Figs. 1 and 2, for SP and MT, respectively. A comparison between both states was represented in Fig. 3.
Fig. 2. Estimation of the variation of the net present values (NPV) of different vaccination rates (70%, 80% and 90%) between maximal and minimal impacts of brucellosis during the economic lives (in years) in the State of Mato Grosso, Brazil.
Fig. 3. Estimation of the variation of the net present values (NPV) of different vaccination rates (70%, 80% and 90%) between maximal and minimal impacts of brucellosis during the economic lives (in years) in the States of Sao Paulo and Mato Grosso, Brazil.
Fig. 1. Estimation of the variation of the net present values (NPV) of different vaccination rates (70%, 80% and 90%) between maximal and minimal impacts of brucellosis during the economic lives (in years) in the State of Sao Paulo, Brazil.
Higher vaccination rates reduce the time to reach 2% of brucellosis prevalence, resulting in shorter vaccination programs and costs. In SP, it would take 12 years to reduce the prevalence of brucellosis from 3.81% to 2% under a 90% vaccination rate and 14 years under 70% vaccination rate. The cost of the first strategy was only 90% of the cost of the
356
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
Table 4 Economic indicators (NPV and payback) of different vaccination rates (70%, 80%, 90%) under maximal and minimal impacts of brucellosis in the States of Sao Paulo and Mato Grosso, Brazil. State
SP
MT
a b c
Vaccination rate
70%
80%
Disease impact
Minimal
Maximal
Minimal
Maximal
Minimal
NPVb (R$1000) Time (years)a Payback Cost (R$1000)a
189,659
271,309
223,040
313,279
254,943
NPVc (R$1000) Time (years)a Payback Cost (R$1000)a
2,851,484
14 7
13 6
7
147,108 4,009,812
3,113,689
6
6
4,374,665
3,349,784
24 2
170,922
3
351,315 6 132,952 4,703,312 22
2 155,349
Maximal 12
139,010
27 3
90%
2
2 145,248
To reach 2% prevalence. 12 years. 22 years.
last. In MT, it would take 22 years to reduce the prevalence of brucellosis from 10.25% to 2% under a 90% vaccination rate and 27 years under 70% vaccination rate. The cost of the first strategy was only 85% of the cost of the last. The payback period is shorter with higher vaccination rates, especially when the impacts of the disease are maximal. In SP, the payback was six years for 90% vaccination rate (for both in maximal and minimal impact of the disease), six years for maximal impact of the disease and seven years, for minimal impact for 80% and 70% vaccination rates. In MT, the payback period was two years for 90% vaccination rate (for both in maximal and minimal impact of the disease), two years for maximal impact of the disease and three years, for minimal impact for 80% and 70% vaccination rates. The payback can also be verified when the cash flows are calculated for each state. The calculation of the cash flow can be found in the Supplementary material. The biggest cost items in the 90% vaccination rate, for example, are the vet visits (58% in SP and 65% in MT), followed by vaccination gun acquisition (38% in SP and 27% in MT). Obviously, the costs with vaccine are higher in MT (7%) than SP (3%). The costs with transport were small, being around 1% in both states. 4. Discussion Brucellosis is a zoonotic disease and has impact on human health, especially professionals who deal with cattle, like veterinarians and slaughterers, creating extra costs. Although it would be interesting to include such information in the simulation models, human brucellosis is not a notifiable disease in Brazil and no official data was available in the time of writing. The choice of two different study areas allowed the comparison of the economic indicators in two different epidemiological situations: SP with almost half of the herd size and almost a third of the prevalence of MT. Moreover, SP had a significantly higher number of farms (twice than MT) and despite its smaller area, has more roads to travel than MT (four-fold). These factors have made the costs of the implementation of any vaccination strategy in SP higher than MT, since more equipment including vaccination guns require purchase and more veterinary visits are required, generating higher NPVs and longer payback periods in SP
than in MT. Veterinary visits were the costliest item in both states and even with a lower proportional participation in SP, the nominal value was bigger than MT. The only item that was more costly in MT, compared to SP was the vaccine vials, since the herd of MT is bigger. This item showed low impact in the overall costs in both states. Annual vaccination efforts corresponding to 70%, 80% and 90% of the heifers aged between 3 and 8 months were chosen to examine an interval around the cut-off of 80%, empirically recommended by many authors as the optimum rate of vaccination for heifers (Paulin and Ferreira Neto, 2003; Dias et al., 2009; Negreiros et al., 2009). Considering the expected decrease in prevalence over time, periods greater than 12 years and 22 years would be necessary to decrease the prevalence below 2% in SP and MT respectively. Thus, under economic lives of 12 years and 22 years, a vaccination rate of 90% is most advantageous economically in SP and MT respectively. Although the results of the present work corroborate empirical recommendations to vaccinate above 80% of the bovine heifer population aged between 3 and 8 months, based on observations made in countries and regions that have implemented vaccination programs (Paulin and Ferreira Neto, 2003), it is noteworthy that different lengths of vaccination programs are essential to prevent economic losses in different prevalence scenarios. Moreover, NPV curve may differ under different initial prevalence scenarios, which were higher for MT, that presented initial brucellosis prevalence significantly higher than SP, as observed in Fig. 3. This conclusion is corroborated by Amaku et al. (2009), who estimated the reduction of the prevalence of brucellosis associated with different vaccination efforts. These authors indicated that vaccination rates between 70% and 90% would reduce the prevalence of brucellosis to 2% in very similar periods of time, but the logistics and resources could vary significantly in each state. Paulin and Ferreira Neto (2003) indicated that prevalence below 2% cannot be reached using vaccination with S19 alone, since this strain does not protect 100% of the vaccinated heifers aged between 3 and 8 months and reintroductions from brucellosis positive wild and domestic animals can occur. Under this scenario, eradication measures such as mass diagnostic screening and culling of positive animals should be implemented along with vaccination strategies using S19
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
or RB51 vaccines in females not vaccinated between 3 and 8 months of age. Animal movements and the incorporation of breeding animals into the herd should only be completed after screening tests. If a vaccination program is conducted along with these measures, the decrease in prevalence may be greater. For that reason, the eradication strategy is expected to be more costly than vaccination strategies, and the slope of NPV curves would be lower over time. In the present work, we have observed that NPV is higher when vaccination rates are higher as well, especially when the impacts of the disease are maximal, in both states. At the time of writing, no recent data about brucellosis driven losses were found, since most of the countries that usually produced such data are now free from the disease. For that reason, it is reasonable to expect that these parameter may fall between the extreme values (minimal and maximal) for the disease impacts shown in Table 2, but these parameter should have been measured prior to implementation of these models in both states. It is also important to note that the magnitude of losses may influence the performance of a vaccination program, and biological parameters such as abortion rates, perinatal mortality, and the infertility or mortality of aborting cows should therefore be assessed prior to the adoption of control measures. An alternative would be to incorporate the statistical distributions of the parameters associated with losses due to brucellosis, allowing probabilistic assessment of such losses. The results of the present work differ from the results found by Santos et al. (2013), who calculated the losses due to brucellosis in a quite similar way, but ignored the costs of vaccination, making a linear regression of actual economic losses by cattle head and brucellosis prevalence to predict the effects of the reduction of prevalence in the reduction of losses. This approach is unrealistic since brucellosis control actions based in vaccination may differ significantly in different states, due to the geographic distribution of farms, road lengths, and heifer populations, addressed in the present work. Moreover, the NPV and payback estimations (not calculated by Santos et al., 2013) make it easy to compare different vaccination efforts, which is a more important issue to implement a vaccination program than knowing the amount of losses with the change of 1% of prevalence. Some assumptions and simplifications may have affected the results of the simulations, but there were no data to support another approach. For example, the calving interval of one year, which is very useful for the simulations, may decrease with the decrease in brucellosis prevalence, since abortions and mortality due to brucellosis are expected to diminish. The mortality rate by female age group was not available by the time of writing. Considering that brucellosis prevention and control measures are different in each age group, that would be a key point to be accessed in future simulation models. Keeping the inflation constant may not be realistic, but the mean inflation rate in Brazil since 2011 was around 8%, the same value of the discount rate (r). The additional costs of nutrient demand, which are expected to increase with higher production following brucellosis control were not considered, even though the main production system of livestock in Brazil is extensive grazing. Finally, we have not considered
357
the price decrease arising from the increased commodities (meat and milk) produced with the reduction of brucellosis prevalence, although we have assumed that the impact of the commodity price from the selected states do not impact the commodities price nationally. This is due to the fact that these products are not consumed locally in each state, with there being competition from other states and countries. The distribution of costs and benefits between public and private sectors is very difficult to measure, even knowing that most of the prevention and control measures are paid by the farmers and the participation of the federal and state governments is only in auditing the actions. The benefits would certainly be shared by both private and public sectors, since an increased tax collection is expected. 5. Conclusion Regions with higher prevalence of brucellosis and greater herds would experience significant economic advantages with the adoption of vaccination programs, whereas in regions with low prevalence and smaller herds, such advantages would be lower, but these measures would nevertheless be worthwhile. Economic analyses should be considered in the planning of control measures for brucellosis, especially in a country as diverse as Brazil. Strategies cannot be standardized for the entire country, which may cause economic losses associated with the adoption of strategies that are sub-optimal locally. For this reason, the National Program for the Control and Eradication of Animal Brucellosis and Tuberculosis (PNCEBT) should consider prevalence and production chain differences to recommend specific control measures. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. prevetmed.2014.12.010. References Acha, P.N., Szyfres, B., 1986. Zoonosis y enfermedades transmissibles communes al hombre y a los animales, 2nd ed. Organización Panamericana de la Salud, Washington. Amaku, M., Dias, R.A., Ferreira Neto, J.S., Ferreira, F., 2009. Modelagem matemática do controle da brucelose bovina por vacinac¸ão. Arq. Bras. Med. Vet. Zoo. 61 (1), 135–141. Assaf Neto, A., 2012. Matemática financeira e suas aplicac¸ões, 12th ed. Atlas, São Paulo. Beef Point, 2012. Cotac¸ões. Beef Point, Piracicaba, Available from: http://www.beefpoint.com.br/ (accessed 20.12.13). Bernués, A., Manrique, E., Maza, M.T., 1997. Economic evaluation of bovine brucellosis and tuberculosis eradication programmes in a mountain area of Spain. Prev. Vet. Med. 30 (2), 137–149. Centro de Estudos Avanc¸ados em Economia Aplicada [CEPEA], 2012a. Fed Cattle Price Index. CEPEA, Piracicaba, Available from: http://www.cepea.esalq.usp.br/english/cattle/ (accessed 20.12.13). Centro de Estudos Avanc¸ados em Economia Aplicada [CEPEA], 2012b. Calf Price Index. CEPEA, Piracicaba, Available from: http://cepea.esalq.usp.br/english/calf/(accessed 20.12.13). Civil Aviation Safety Authority [CASA], 2010. Cost Benefit Analysis Procedures Manual. CASA, Canberra, Available from: http://www.casa.gov.au (accessed 20.12.13). Costa, M.J.R.P., Toledo, L.M., Schmidek, A., 2006. Boas práticas de manejo – vacinac¸ão. FUNEP, Jaboticabal. Departamento Intersindical de Estatística e Estudos Socioeconômicos [DIEESE], 2006. Nota técnica 19. DIEESE, São Paulo.
358
A.J.S. Alves et al. / Preventive Veterinary Medicine 118 (2015) 351–358
Dias, R.A., Gonc¸alves, V.S.P., Figueiredo, V.C.F., Lôbo, J.R., Lima, Z.M.B., Paulin, L.M.S., Gunnewiek, M.F.K., Amaku, M., Ferreira Neto, J.S., Ferreira, F., 2009. Situac¸ão epidemiológica da brucelose bovina no Estado de São Paulo. Arq. Bras. Med. Vet. Zoo. 61 (1), 118–125. Dijkhuizen, A.A., Huirne, R.B.M., Jalvingh, A.W., 1995. Economic analysis of animal diseases and their control. Prev. Vet. Med. 25 (2), 135–150. Faria, J.F., 1984. Situac¸ão da Brucelose no Brasil. In: Comunicac¸ão Científica da Faculdade de Medicina Veterinária e Zootecnia da Universidade de São Paulo 8, pp. 161–175. Instituto Brasileiro de Geografia e Estatística [IBGE], 2010. Censo populacional de 2010. IBGE, Brasília, Available from: http://ibge/home/estatistica/populacao/censo2010/default.shtm (accessed 20.12.13). Instituto Brasileiro de Geografia e Estatística [IBGE], 2011. Sistema IBGE de ecuperac¸ão automática [SIDRA]. IBGE, Brasília, Available from: http://www.sidra.ibge.gov.br (accessed 20.12.13). James, A., 2004. The use of economic analysis to define animal health the policies. In: Compendium of technical items presented to the International Committee or to the regional commission, pp. 7–12, Available at: http://www.oie.int/doc/ged/D677.pdf Ministério da Agricultura, Pecuária e Abastecimento [MAPA], 2006. Programa Nacional de Controle e Erradicac¸ão da Brucelose e Tuberculose. MAPA, Brasília, Available from: http://www.agricultura.gov.br (accessed 20.12.13). Ministério dos Transportes [MT], 2005. Anuário estatístico dos transportes terrestres. MT, Brasília. Negreiros, R.L., Dias, R.A., Ferreira, F., Ferreira Neto, J.S., Gonc¸alves, V.S.P., Silva, M.C.P., Figueiredo, V.C.F., Lôbo, J.R., Freitas, J., Amaku, M., 2009. Situac¸ão epidemiológica da brucelose bovina no Estado de Mato Grosso. Arq. Bras. Med. Vet. Zoo. 61 (1), 56–65.
Otte, M.J., Chilonda, P., 2000. Animal Health Economies: An Introduction. Animal Production and Healthy Division (AGA), FAO, Rome. Pacheco, G., Mello, M.T., 1956. Brucelose. Distribuidora Livraria Ateneu, Rio de Janeiro. Paulin, L.M., Ferreira Neto, J.S., 2003. O Combate à brucelose bovina: situac¸ão brasileira. FUNEP, Jaboticabal. Poester, F.P., Gonc¸alves, V.S.P., Lage, A.P., 2002. Brucellosis in Brazil. Vet. Microbiol. 90, 55–62. Rego-Bordeaux, R., Paulo, G.P., Spritzer, I.M.P.A., Zotes, L.P., 2010. Viabilidade econômica financeira de projetos, 3rd ed. Fundac¸ão Getulio Vargas, Rio de Janeiro. Reis, M.C., Firpo, S., 2007. O salário mínimo e a queda recente da desigualdade no Brasil. In: Barros, R.P., Foguel, M.N., Ulyssea, G. (Eds.), Desigualdade de renda no Brasil: uma análise da queda recente, vol. 2. Instituto de Pesquisa Econômica Aplicada, Brasília, Available from: http://www.ipea.gov.br/portal/images/stories/PDFs/livros/Cap33.pdf (accessed 20.12.13). Santos, R.L., Martins, T.M., Borges, A.M., Paixão, T.A., 2013. Economic losses due to bovine brucellosis in Brazil. Pesq. Vet. Bras. 33 (6), 759–764. Sheperd, A.A., Simpsom, B.H., Davidson, R.M., 1980. An economic evaluation of the New Zealand bovine brucellosis eradication. In: Proceedings of the Second International Symposium on Veterinary Epidemiology and Economics, pp. 443–447. Vanni, T., Luz, P.M., Ribeiro, R.A., Novaes, H.M.D., Polanczyk, C.A., 2009. Avaliac¸ão econômica em Saúde: aplicac¸ões em doenc¸as infecciosas. Cad. Saude Publica 25 (12), 2543–2552. Wickham, H., 2009. ggplot2: Elegant Graphics for Data Analysis. Springer, New York.