Artigos

Conteúdo (article):

Vigilância epidemiológica de doenças tropicais negligenciadas em áreas silenciosas: o caso da esporotricose zoonótica
Epidemiologic surveillance of neglected tropical diseases in silent areas: the case of zoonotic sporotrichosis


Ligia Neves Scuarcialupi, Universidade de São Paulo, Faculdade de Medicina Veterinária e Zootecnia, Departamento de Medicina Veterinária Preventiva e Saúde Animal, São Paulo – SP, Brasil, ligia.scuarcialupi@usp.br, 0000-0001-6552-2772.
Gabriela Chueiri de Moraes, Universidade de São Paulo, Faculdade de Medicina Veterinária e Zootecnia, Departamento de Medicina Veterinária Preventiva e Saúde Animal, São Paulo – SP, Brasil, gabrielamoraes@usp.br, 0009-0000-7660-217X.
Fernando Cortez Pereira, Universidade de São Paulo, Faculdade de Medicina Veterinária e Zootecnia, Departamento de Medicina Veterinária Preventiva e Saúde Animal, São Paulo – SP, Brasil e Centro de Controle de Zoonoses - Secretaria da Saúde do Município de Guarulhos, mv.fernando@gmail.com, 0000-0003-4633-3882.
Yasmin da Silva Alexandre, Universidade de São Paulo, Faculdade de Medicina Veterinária e Zootecnia, Departamento de Medicina Veterinária Preventiva e Saúde Animal, São Paulo – SP, Brasil, yas.ale@usp.br, 0000-0003-0561-2455.
Oswaldo Santos Baquero, Universidade de São Paulo, Faculdade de Medicina Veterinária e Zootecnia, Departamento de Medicina Veterinária Preventiva e Saúde Animal, São Paulo – SP, Brasil, baquero@usp.br, 0000-0003-2695-7946.


RESUMO
Uma prática comum na análise da distribuição espacial das Doenças Tropicais Negligenciadas é pressupor que em áreas silenciosas (sem notificações) não há casos, o que pode reforçar, quando há subnotificação, a negligência de áreas que deveriam ser prioritárias. Como alternativa, é possível predizer o número de casos em áreas silenciosas, a partir de informações epidemiológicas e de dependência espacial. Neste estudo exemplificamos essa abordagem utilizando a aproximação integrada e aninhada de Laplace, em modelos espaciais Bayesianos, que relacionam a vulnerabilidade social e o número de casos notificados de esporotricose felina (zoonótica) em setores censitários (SCs) do município de Guarulhos. Além de predições para os SCs silenciosos, atribuímos um índice de priorização a todos os SCs, resultando em um cenário epidemiológico mais problemático em comparação ao que assume a inexistência de casos nos SCs silenciosos. A fim de validar iterativamente as predições do índice e calibrar o grau de confiança que se atribui a elas, pode-se comparar a distribuição dos índices de priorização dos SCs silenciosos com a distribuição de casos identificados mediante vigilância ativa numa amostra deles.
Palavras-chave: Doenças Tropicais Negligenciadas. Indicador de Risco. Vulnerabilidade Social. Vigilância Epidemiológica. Prioridades em Saúde.

ABSTRACT
A common practice in the analysis of the spatial distribution of Neglected Tropical Diseases is to assume that in silent areas (no reports) there are no cases. However, when the problem is underreporting, it risks reinforcing the neglect of areas that should be a priority. Instead of this assumption, one can predict the number of cases in silent areas using epidemiologic and spatial dependence information. The present study exemplifies this approach, using the integrated nested Laplace approximation in Bayesian spatial models that relate social vulnerability and the number of reported cases of feline (zoonotic) sporotrichosis in census tracts (CTs) of the municipality of Guarulhos. In addition to predictions for silent CTs, we assigned a priority index to all CTs. The results showed a more problematic epidemiologic situation, compared to the scenario in which it is assumed that there are no cases in silent CTs. To iteratively validate the index predictions and calibrate the degree of confidence assigned to the predictions, one can compare the distribution of the priority indices of silent CTs with the distribution of cases identified through active surveillance in a sample of silent CTs.
Keywords: Neglected Diseases. Risk Index. Social Vulnerability. Epidemiologic Surveillance. Health Priorities.


INTRODUCTION
Neglected Tropical Diseases (NTDs) refer to a group of diseases and conditions that share geographic and social contexts1. They predominate in marginalized multispecies collectives in tropical and subtropical areas, affecting more than one billion people and an unknown number of other animals annually in impoverished territories of the African, Asian, and American continents2,3. Faced with the prospect of insufficient profit, the pharmaceutical industry invests little in the development of vaccines and medicines for diseases that primarily affect populations with low purchasing power, a situation that is aggravated by the limited funding of other types of research on NTDs4,5. However, the problem of the neglect of tropical diseases lies mainly in those who are affected by them, since some continue to cause thousands of human and other-than-human deaths despite the existence of vaccines or medicines, as is the case with rabies, which kills approximately 60,000 humans per year6.
In light of this scenario, in 2015, the United Nations established 17 Sustainable Development Goals (SDGs) to be achieved by 2030, with one of the SDG targets, Goal 3 (Good Health and Wellbeing), being to end NTDs epidemics7. In 2020, with the aim of promoting socioeconomic development and reducing health inequities, the World Health Organization (WHO) launched a plan to combat NTDs, whose proposals include an integrated approach across different diseases/disease groups and increased participation on the part of national and local governments, as well as communities, in establishing priorities and strategies to combat these diseases8. The insufficient participation of local populations in these actions is a key failure. Stopping the marginalization of multispecies collectives, which perpetuates disease transmission in the Global South, requires effective consideration of who actually represents the interests of these collectives6,9. Therefore, the need for structural changes must be highlighted, understanding that NTDs contribute to the maintenance of inequalities4.
At a national level, following the pattern of occurrence in other locations, NTDs disproportionately affect impoverished populations in rural areas and urban peripheries, especially those living in the Northeast region, the Amazon region, and indigenous territories3. Among the countries in the Western Hemisphere, Brazil has the largest population affected by NTDs2, reflecting the critical situation of inequality and poverty present in a country with continental dimensions. Among these diseases, sporotrichosis, the most prevalent and widespread subcutaneous mycosis in the world, caused by fungi of the Sporothrix schenckii complex10, has assumed very particular epidemic proportions over the past two decades11,12.
Although all NTDs have a profound impact on public health, fungal NTDs are the most widely ignored13. Mycetoma was the first mycosis included in the WHO list of NTDs, which happened only in 2016, followed by the inclusion of chromoblastomycosis10 and “other deep mycoses”, including sporotrichosis8,13. In March 2023, the WHO held the first global meeting to specifically address NTDs with cutaneous manifestations, which represent half of all NTDs, highlighting the importance of surveillance and mapping in areas of co-endemicity to guide integrated control and management interventions14. The precarious socioeconomic and structural conditions of many families make them the most affected by sporotrichosis15, highlighting the correlation of the disease with social vulnerability16.
Until the 1970s, sporotrichosis was considered an occupational risk, affecting individuals who had contact with soil, since it is caused by saprophytic fungi, which is widespread in nature and commonly found in soil and vegetation15. However, since the 1990s, its epidemiological profile has changed17 and the progressive increase in cases since then has been related to zoonotic transmission, mostly by cats (Felis catus)18. These animals are highly susceptible hosts to this fungus17, and due to their behavioral characteristics, proximity to humans, and the ability to harbor a significant amount of yeast between their claws, cats have played a key role in this scenario19.
When questioning why zoonotic sporotrichosis continues to be neglected in Brazil, Alvarez et al.15 carried out a systematic review and highlighted that factors, such as the progression of the disease, resulting from late diagnosis and treatment (despite having a good prognosis); the poor sanitary conditions in the country; the large population of cats in contact with humans and victims of abandonment; the complex adaptive evolutionary strategies of the fungus; as well as the atypical and more costly clinical manifestations in terms of treatment, have all contributed to the spread of the disease to 25 Brazilian states. Furthermore, the predominant etiological agent in Brazil is S. brasiliensis, the most virulent of the complex12,20, with a sophisticated pathogen-host-environment interaction17 and exclusive to South America19. From 1907, when it was described in Brazil17, to 2020, 10,400 cases of sporotrichosis in humans and 8,538 in animals (more than 90% in cats) have been reported in the national literature19. However, due to the neglected nature11, these numbers are believed to be underestimated.
The genotype of S. brasiliensis strains from Rio de Janeiro, an endemic state for mycosis and where the first reported outbreaks occurred in Brazil, is the same as that found in strains from Paraná, Minas Gerais, and São Paulo, which could suggest, due to proximity, the spread of this species from Rio de Janeiro12. This hypothesis is supported by the study by Carvalho et al.20, who, by analyzing the genetic diversity, population structure, and different genotypes in a large collection of Sporothrix isolates, covering the main endemic areas at national and international levels, identified the state of Rio de Janeiro as the most likely center of origin for the spread of S. brasiliensis throughout the country, both to the states that border it and to northeastern Brazil. In the state of São Paulo, the first case of zoonotic sporotrichosis dates back to the 1950s21, but it was in 2010 that the disease reached epidemic proportions, with reports by the São Paulo Zoonosis Control Center17. The municipality of Guarulhos, adjacent to the state capital, is located at the confluence of highways connecting São Paulo and Rio de Janeiro, and is home to the largest airport in Latin America16, which results in heavy traffic of people and animals, which can favor the transmission of infectious diseases. S. brasiliensis is also the etiological agent involved in sporotrichosis outbreaks in this municipality22, which witnessed a significant increase in cases after the first report in 201116, culminating in the requirement for mandatory reporting of human sporotrichosis cases from 2016 onwards23.
The first case recorded in Guarulhos occurred in a favela (a Brazilian slum or shantytown), where traditionally prescribed preventive health actions are challenging, and there is already evidence that there is a relationship between social vulnerability and a higher prevalence of this disease15–17,19. This municipality has several areas of medium/high social vulnerability, and the study carried out by Scuarcialupi et al.16 found that many areas with high social vulnerability, close to other areas with a large number of reports, had no recorded cases. The absence or occurrence of few cases in supposed risk areas may indicate that the risk is not real, but it may also denote detection failures, which can be improved through surveillance based on the risk that a given geographic area presents for the occurrence of the disease. Surveillance is essential in the control and elimination of NTDs2, and this type of approach allows for a more efficient use of human and material resources24.
In the case of Guarulhos, surveillance of animal sporotrichosis is carried out by the Center for Zoonosis Control (Centro de Controle de Zoonoses – CCZG), which performs multiple tasks and faces the challenge of serving the entire municipality with limited material and personnel resources16. Silent areas may contain important foci of disease dissemination, and the detection of such foci is crucial for the early diagnosis and interruption of the transmission chain. Therefore, active surveillance should prioritize the care of locations with the highest risk, considering the possible existence of silent areas due to detection failures.
In view of the above, the present study offers a method to prioritize areas in order to aid in risk-based epidemiologic surveillance. Rather than assuming that in silent areas the number of cases is zero, using available epidemiologic and contextual information, this number was predicted. The method, applicable to several NTDs, is exemplified with data on social vulnerability and feline sporotrichosis from the municipality of Guarulhos.

METHODS
This is an ecological modeling study carried out in the municipality of Guarulhos in the Metropolitan Region of the state of São Paulo, Brazil. This city, with an estimated population of 1,379,182 people (second largest in the state) in 2019, borders the municipalities of Mairiporã, Nazaré Paulista, Santa Isabel, Arujá, Itaquaquecetuba, and São Paulo, and was territorially divided into 1,748 census tracts (CTs)25. The outcome of the modeling was the number of reported cases of feline sporotrichosis, between 2011 and 2019, in the CTs. The São Paulo Social Vulnerability Index (Índice Paulista de Vulnerabilidade Social – IPVS) and the spatial dependence between the number of cases in neighboring CTs (neighborhood established by the Queen criterion26) were used to predict the number of cases in silent CTs. The spatiotemporal dynamics of feline sporotrichosis in Guarulhos were described in another study16, which provided evidence of the association between social vulnerability and the occurrence of this disease.
The IPVS is an index that ranges from 1 to 7: 1: extremely low, 2: very low, 3: low, 4: medium, 5: high (urban), 6: very high (subnormal urban), and 7: high (rural). The IPVS of Guarulhos ranged between 1 and 6, according to data from the 2010 census27.
The Guarulhos Health Department provided data on positive cases of feline sporotrichosis detected by the CCZG surveillance service between 2011 and 2019. These cases were aggregated by CT, after being geocoded from their respective addresses, using the R software and the Google Maps API. The IPVS was obtained from the Sistema Estadual de Análise de Dados (SEADE) Foundation of the Government of the State of São Paulo27.

Statistical models

Where i (1, …, n) is the indexer of the CTs, the general equation of the models was given as:
η_i=α+ς_i+βx,
In which ηi = log(E(yi)) is the average of an additive linear predictor, yi is the number of observed cases, α is the fixed intercept, β is the effect of IPVS (x), and ςi is the combination of a structured spatial effect υi and a non-structured spatial effect νi. It was presumed that yi follows a negative binomial distribution with an average θiEi, in which θi is the Relative Density of Cases (RDC) of CT i, and Ti is the area of CTi. The density of cases was a measure of incidence and proxy for the risk of occurrence, in such a way that RDC was a proxy for relative risk. IPVS 1 and 2 were the reference categories for estimating β and were aggregated, as only one case was identified in the CTs with IPVS 1.

The definition of the spatial effect was given by Riebler et al.28:
ς_i=1/√τ (√(1-φνi)+√φ υ)
In which τ is the marginal precision, νi follows a normal distribution, and υi is an autoregressive conditional model:
ν_i∼N(0,1/τ(1-φ) )

υ_i∨υ_(-i),τφ∼N(1/η_δi ∑_(j∈δi)▒υ_j ,1/(η_δi τφ))

In the previous equations, ηẟi is the number of neighbors of i, and φ is the proportion of marginal spatial variance explained by υ.
Using the model described above, we estimated the RDCi, the probability of Excess Density EDi = Prob(RDCi > RD), given that RD is the average density of cases, and a Priority Index (PI) 29:

PI=(RDC_i ED_i)/max(RDC_i ED_i ) 100

This PI indicates the priority that should be assigned to each CT, the highest priority of which receives a PI = 100. The other PIs are relative to this 100. Thus, if CT A has a PI = 100 and CT B has a PI = 50, the priority that should be given to B is equivalent to 50% of the priority that should be allotted to A16.


Priors

According to the principle of parsimony, a penalizing complexity of priors30 (PC priors) was used. PC priors favor models with spatial variance = 0 (τ = ∞) and with φ = 1 (with no structured spatial effect). The penalty was based on a constant decay rate of a Gumbel type 2 distribution, specified by probabilistic statements. For τ, the probabilistic statement was Prob((1/√τ)>U)=α and is equivalent to a constant decay rate equal to -log 28,30. As regards φ, the statement used was Prob(φ < U) = α. More specifically, the statements Prob((1/√τ)>0.3/0.31)=0.01 and Prob(φ < 0.5) = 0.7 were applied, which presupposes a residual RDC (τ) of less than 2, with a probability of 0.99, and in which ν explains the majority of variation16.

Substitution of zeros for “NA”
Based on the data obtained, 66.65% (1,165/1,748) of the CTs in Guarulhos were classified as silent for feline sporotrichosis, which may mean an absence of cases in the area or a lack of information. To avoid assuming an absence of cases, predictions were made based on the number of cases in neighboring sectors and the social vulnerability in the area and its neighbors.


R software packages used in this study

The statistics were calculated using the following packages of the R 3.6.3 software: devtools 31, tidyverse 32, lubridate 33, INLA 34, INLAOutputs 35, lwgeom 36, and cowplot 37.

RESULTS
Since the first report in 2011, there has been a progressive increase in the number of cases of feline sporotrichosis in the city of Guarulhos, especially since 2014, when the curve became steeper, indicating an increase in incidence, greater detection, or both. This growth pattern was interrupted in 2019, when the number of reported cases decreased (Figure 1), reaching even lower levels than in 2017. The total number of cases accumulated between 2011 and 2019 was 2,953.
The spatial distribution of cases was not homogeneous over the years and affected different locations in the city, with a low incidence in less urbanized areas (Figure 2). Few CTs had a high incidence of cases, and those silent for the occurrence of the disease during the studied period accounted for 66.65% of the total CTs in the city.
When the absence of information (“NA”) was considered instead of the absence of cases (“zero cases”), the proportion of priority CTs increased (Figure 3). The proportion of CTs with priority values equal to zero decreased from 64.7% (1,131/1,748) to 0.2% (3/1,748) after the change in approach. The 50%, 75%, and 95% quantiles of the prioritization index were equal to 0, 6, and 12.8, respectively, in the models with zero cases in the silent CTs; in the models with “NA” in the silent CTs, these quantiles were equal to 3.3, 18.8, and 27.2. In other words, only 5% of the silent CTs showed a prioritization index higher than 12.8 in the first model, while in the second, the top 5% of the indices were above 27.2%. There was disagreement between the CTs with the highest priority values in the two approaches. Among the ten highest priorities, with the exception of first place, all diverged regarding the position occupied when considering “zero cases” and “NA”. There was also the inclusion of four new higher priority CTs in the approach that assumes the absence of information, excluding those that occupied the fifth, seventh, eighth, and tenth priority positions when the assumption was the absence of cases (Table 1).

DISCUSSION
The proposed methodological approach allows one to predict the number of cases in silent areas, using epidemiologic and spatial dependence information, and was exemplified with data on feline sporotrichosis and social vulnerability in the municipality of Guarulhos. The epidemiologic scenario predicted by the approach differs from that resulting from the assumption of the absence of cases in silent areas, bringing implications for active surveillance and prioritization of prevention and control actions. This difference is in line with expectations, since the predicted value depends on the average density of cases observed in a CT and its neighboring vicinity. When a silent CT enters the analysis with zero cases, the predicted value within it is a function of an average (structured effect of the model) influenced by this zero, which also affects the predictions in its neighboring vicinity. On the contrary, when it enters without a value, this average is not reduced by the effect of an additional zero. It should be noted that in Bayesian modeling all estimates are predictions. Due to the structured spatial effect (the way the predicted mean is calculated), most silent CTs end up having positive predicted values, even when it is assumed that they have zero cases. This elimination of most zeros, more precisely, of the extremes of the observed distribution, occurs with all spatial smoothing models, which are common in epidemiology. The distribution predicted by such models, from an observed distribution of counts, has fewer zeros, and the extreme values are smaller.
The proposed approach assumes that the silent condition is the result of underreporting in CTs with non-silent neighbors. Although it is possible that there are no cases in them, there are reasons to believe that the proposed approach is more convenient, given that a sector without notifications would hardly receive a high prioritization index if it truly had no cases. This is because if a silent sector receives a high index, it means that it is in the middle of sectors with high incidence, a situation in which it is unlikely that it will have no cases. The cost of assigning an incorrect priority in such situations, translated into surveillance actions, would mean greater attention to the neighboring vicinity of epidemic areas, which is not exactly a problem, given that this area without cases, surrounded by epidemic areas, can easily become epidemic as well. Since there is low probability of misclassifying silent areas and errors have little epidemiologic relevance, predictions without assuming the absence of cases improve the efficiency of active surveillance in silent areas. Sampling silent areas, stratified according to the prioritization index, enables the accuracy of predictions to be validated. If there are more cases in the higher priority strata than in the lower priority strata, there is evidence of correspondence between predictions and observations, which can be described using metrics that are consistent with the sampling design and the type of comparison performed. This accuracy may vary between diseases and for the same disease in different places and times, which is why it is advisable to update it each time active surveillance guided by predictions produces new data.
It should be noted that when using area data, the internal heterogeneity of the analysis units is lost. The consequences of this depend on the size of the areas and the dynamics of disease dissemination. Although geostatistical or point pattern approaches do not present this disadvantage, an area model was chosen in order to consider the effects of covariates available in a spatially aggregated format and to be able to assign priority values to defined areas.
The existence of silent areas due to underreporting is present in the surveillance of several neglected zoonoses in Brazil, such as spotted fever38, leishmaniasis39, and hantavirus40. In addition, underreporting also occurs among non-zoonotic diseases that are no longer considered neglected. Acquired Immunodeficiency Syndrome (AIDS) is an example, as pointed out in the study by Carvalho et al.41, who reported an average of 17.7% underreporting of tuberculosis-HIV/AIDS co-infection. Carmo et al.42 also found an underreporting of deaths from HIV/AIDS.
The case of zoonotic sporotrichosis was chosen, as it is a growing public health problem, as highlighted in Technical Note No. 60/2023 of the Secretariat for Health and Environmental Surveillance of the Ministry of Health43. In Brazil, infections caused by S. brasiliensis are associated with epizootics in felines, with a high potential for zoonotic transmission. Despite the mandatory reporting of human sporotrichosis in Guarulhos since 2016, it is believed that underreporting occurs in both humans and felines. Since cats act as sentinels12, establishing mandatory reporting of the disease in them is a strategic surveillance measure, since the prevention of feline sporotrichosis is a way to prevent both the occurrence of cases in humans and underreporting. Some states, such as Amazonas, Paraná, and Rio de Janeiro, in addition to the capital city of São Paulo, have already adopted this measure.
Guarulhos accumulated 2,953 reported cases between 2011 and 2019. A significant increase in notifications has been noted since the first report, especially from 2014 onwards, interrupted in 2019, which saw an 18.48% drop in notifications when compared to the previous year. It is important to note that, after 2016, when the notification of human cases of sporotrichosis became mandatory in Guarulhos, there were no relevant changes in the temporal trend of the number of notifications. By contrast, the cases stemmed from different areas of the municipality, with a variable spatial distribution over time. Since the impact of underreporting is unknown, the observed spatiotemporal dynamics should be interpreted with caution. In fact, the number of cases in non-silent CTs, used to make the prediction, may be underreported, which would lead to an underestimation of the density of cases in silent CTs. Predicting cases in silent areas helps to understand and reduce this impact, given that, if underreporting tends to be higher in those with a higher prioritization index, it is because the magnitude of the risk and underreporting in silent areas have an associated spatial distribution. Therefore, concentrating detection efforts in priority silent areas becomes a more efficient way to reduce the effects of underreporting.
Such factors as incorrect diagnosis and death prior to diagnosis may contribute to underreporting of sporotrichosis; however, these appear to be less relevant. Contrary to what Oliveira et al.38 reported for spotted fever in Brazil, regarding species of rickettsia causing mild signs that could make it difficult to capture records, generating a lack of knowledge regarding the true magnitude of the problem, S. brasiliensis generally causes evident clinical signs, which facilitate diagnosis. Furthermore, sporotrichosis is a disease that typically presents a subacute or chronic course44, facilitating the establishment of a diagnosis prior to death. Underreporting, however, is subject to the size of the population of non-housed cats, which is unknown and likely significant. Therefore, even directing detection efforts to silent areas with a greater potential of risk, the difficulty in diagnosing and treating these animals is one of the main challenges, perhaps the greatest, to effectively preventing the disease. The cat population, which is the most susceptible to infection, has an unknown size in Guarulhos, so it would be impossible to calculate the cumulative incidence or incidence density. Therefore, incidence was measured by case density (number of cases per area), which is not a measure of risk per se, but it is related to it, since the higher the density of the cat population, the greater the probability of contact between infected and susceptible individuals. In terms of epidemiologic surveillance, this measure allowed for prioritization based on the spatial concentration of cases and may be useful in risk estimates for other zoonotic NTDs.
The proposal to optimize the epidemiologic surveillance of NTDs in silent and vulnerable areas aims to reduce the problems associated with underreporting and limitations in operational capacity. However, this alone is not expected to be enough to meet the demands of municipalities. The need to expand surveillance actions, given the limitations linked to health services, determines popular participation as a possibility to be explored. The actions of organized communities, in a complementary manner, have contributed to the prevention of other diseases, serving as a reference for community surveillance of sporotrichosis. In a pilot study carried out in Tanzania, Madon et al.45 observed a statistically significant reduction in the prevalence of diarrhea and schistosomiasis when a program that encouraged community participation was implemented. In addition, they observed an improvement in the awareness of official interventions, which is a favorable point for the control of NTDs.
However, it should be emphasized that community surveillance does not replace the epidemiological surveillance carried out by official health services, especially when it comes to marginalized communities. In the specific case of feline sporotrichosis, a recommended control measure is the isolation of infected animals until clinical and pathological cure is achieved, which can take months or even years. Scuarcialupi et al.16 showed that, in general, the areas most affected by sporotrichosis in Guarulhos are those with greater social vulnerability, where precarious housing is easily found, with inadequate infrastructure for isolating animals. Consequently, the control of sporotrichosis, as well as of other NTDs, will only be truly effective when measures to reduce health inequities are put into practice.
NTDs are associated with data scarcity5 and silent areas are not necessarily free of the disease, especially when the multispecies collectives that reside in them are marginalized and neglected. Therefore, NTD prevention needs to halt marginalization, as only then can their pathological effects be replaced by more-than-human care relationships that are antagonistic to neglect6,9. Prioritizing silent areas based on the risk of NTD occurrence is a way to mitigate the multifaceted neglect that affects some multispecies collectives, and the methodology proposed in this study is a contribution in this sense.

ACKNOWLEDGEMENTS
This work was conducted with the support of the Coordination for the Improvement of Higher Education Personnel - Brazil (CAPES) - Funding Code 33002010123P4.

REFERENCES

1. Boletim Epidemiológico - Número Especial, mar.2021 - Doenças Tropicais Negligenciadas — Ministério da Saúde [Internet]. [cited 2023 Jun 15]. Available from: https://www.gov.br/saude/pt-br/centrais-de-conteudo/publicacoes/boletins/epidemiologicos/especiais/2021/boletim_especial_doencas_negligenciadas.pdf/view
2. Melo GBT, Angulo-Tuesta A, Silva EN da, Santos T da S, Uchimura LYT, Obara MT. Evolution of research funding for neglected tropical diseases in Brazil, 2004–2020. PLoS Negl Trop Dis [Internet]. Public Library of Science; 2023 Mar 16 [cited 2023 Jun 15];17(3):e0011134. Available from: https://journals.plos.org/plosntds/article?id=10.1371/journal.pntd.0011134
3. Lindoso JAL, Lindoso AABP. Neglected tropical diseases in Brazil. Rev Inst Med Trop São Paulo [Internet]. Instituto de Medicina Tropical de São Paulo; 2009 Oct [cited 2023 Jun 15];51:247–253. Available from: https://www.scielo.br/j/rimtsp/a/g5p5PqgVkR99pf6wVT9mDnz/
4. Oliveira RG de. Sentidos das Doenças Negligenciadas na agenda da Saúde Global: o lugar de populações e territórios. Ciênc Saúde Coletiva [Internet]. ABRASCO - Associação Brasileira de Saúde Coletiva; 2018 Jul [cited 2023 Jun 15];23:2291–2302. Available from: https://www.scielo.br/j/csc/a/Zs9vNx7xqcc3XrjsmdSCRFm/
5. Rosário MS, Oliveira ML de, Lima C de A, Vieira MA, Carneiro JA, Costa FM da. Doenças tropicais negligenciadas: caracterização dos indivíduos afetados e sua distribuição espacial. Rev Bras Pesqui Em SaúdeBrazilian J Health Res [Internet]. 2017 Sep 30 [cited 2023 Jun 15];19(3):118–127. Available from: https://periodicos.ufes.br/rbps/article/view/19574
6. Baquero OS. One Health of Peripheries: Biopolitics, Social Determination, and Field of Praxis. Front Public Health [Internet]. 2021 Jun 30 [cited 2023 Jun 15];9:617003. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8278197/ PMCID: PMC8278197
7. Objetivos de Desenvolvimento Sustentável | As Nações Unidas no Brasil [Internet]. [cited 2023 Jun 15]. Available from: https://brasil.un.org/pt-br/sdgs
8. World Health Organization. Ending the neglect to attain the sustainable development goals: a road map for neglected tropical diseases 2021–2030 [Internet]. World Health Organization; 2020 [cited 2023 Jun 15]. Available from: https://apps.who.int/iris/handle/10665/338565
9. Baquero OS, Benavidez Fernández MN, Acero Aguilar M. From Modern Planetary Health to Decolonial Promotion of One Health of Peripheries. Front Public Health [Internet]. 2021 Jun 10 [cited 2023 Jun 22];9:637897. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8222522/ PMCID: PMC8222522
10. Queiroz-Telles F, Fahal AH, Falci DR, Caceres DH, Chiller T, Pasqualotto AC. Neglected endemic mycoses. Lancet Infect Dis [Internet]. Elsevier; 2017 Nov 1 [cited 2023 Jun 15];17(11):e367–e377. Available from: https://www.thelancet.com/journals/laninf/article/PIIS1473-3099(17)30306-7/fulltext PMID: 28774696
11. Gremião IDF, Oliveira MME, Monteiro de Miranda LH, Saraiva Freitas DF, Pereira SA. Geographic Expansion of Sporotrichosis, Brazil. Emerg Infect Dis [Internet]. 2020 Mar [cited 2023 Jun 15];26(3):621–624. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7045854/ PMCID: PMC7045854
12. Rodrigues AM, de Melo Teixeira M, de Hoog GS, Schubach TMP, Pereira SA, Fernandes GF, Bezerra LML, Felipe MS, de Camargo ZP. Phylogenetic analysis reveals a high prevalence of Sporothrix brasiliensis in feline sporotrichosis outbreaks. PLoS Negl Trop Dis. 2013;7(6):e2281. PMCID: PMC3688539
13. Rodrigues ML, Albuquerque PC. Searching for a change: The need for increased support for public health and research on fungal diseases. PLoS Negl Trop Dis [Internet]. Public Library of Science; 2018 Jun 14 [cited 2023 Jun 15];12(6):e0006479. Available from: https://journals.plos.org/plosntds/article?id=10.1371/journal.pntd.0006479
14. WHO’s first global meeting on skin NTDs calls for greater efforts to address their burden [Internet]. [cited 2023 Jun 15]. Available from: https://www.who.int/news/item/31-03-2023-who-first-global-meeting-on-skin-ntds-calls-for-greater-efforts-to-address-their-burden
15. Alvarez CM, Oliveira MME, Pires RH. Sporotrichosis: A Review of a Neglected Disease in the Last 50 Years in Brazil. Microorganisms [Internet]. Multidisciplinary Digital Publishing Institute; 2022 Nov [cited 2023 Jun 15];10(11):2152. Available from: https://www.mdpi.com/2076-2607/10/11/2152
16. Scuarcialupi LN, Pereira FC, Baquero OS. Feline sporotrichosis: social vulnerability and prioritization of geographic areas in Guarulhos, SP, Brazil. Braz J Vet Res Anim Sci [Internet]. 2021 Dec 22 [cited 2023 Jun 15];58:e188291–e188291. Available from: https://www.revistas.usp.br/bjvras/article/view/188291
17. Montenegro H, Rodrigues AM, Dias MAG, da Silva EA, Bernardi F, de Camargo ZP. Feline sporotrichosis due to Sporothrix brasiliensis: an emerging animal infection in São Paulo, Brazil. BMC Vet Res. 2014 Nov 19;10:269. PMCID: PMC4244058
18. Gremião IDF, Miranda LHM, Reis EG, Rodrigues AM, Pereira SA. Zoonotic Epidemic of Sporotrichosis: Cat to Human Transmission. PLOS Pathog [Internet]. Public Library of Science; 2017 Jan 19 [cited 2023 Jun 15];13(1):e1006077. Available from: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1006077
19. Rabello VBS, Almeida MA, Bernardes-Engemann AR, Almeida-Paes R, de Macedo PM, Zancopé-Oliveira RM. The Historical Burden of Sporotrichosis in Brazil: a Systematic Review of Cases Reported from 1907 to 2020. Braz J Microbiol [Internet]. 2022 Mar 1 [cited 2023 Jun 15];53(1):231–244. Available from: https://doi.org/10.1007/s42770-021-00658-1
20. de Carvalho JA, Beale MA, Hagen F, Fisher MC, Kano R, Bonifaz A, Toriello C, Negroni R, Rego RS de M, Gremião IDF, Pereira SA, de Camargo ZP, Rodrigues AM. Trends in the molecular epidemiology and population genetics of emerging Sporothrix species. Stud Mycol [Internet]. 2021 Dec 17 [cited 2023 Oct 30];100:100129. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8693333/ PMCID: PMC8693333
21. Rodrigues AM, Della Terra PP, Gremião ID, Pereira SA, Orofino-Costa R, de Camargo ZP. The threat of emerging and re-emerging pathogenic Sporothrix species. Mycopathologia [Internet]. 2020 Oct 1 [cited 2023 Jun 15];185(5):813–842. Available from: https://doi.org/10.1007/s11046-020-00425-0 PMID: 32052359
22. Gonsales FF. Estudo da esporotricose em gatos domésticos no município de Guarulhos, estado de São Paulo [Internet] [text]. Universidade de São Paulo; 2018 [cited 2023 Jun 15]. Available from: http://www.teses.usp.br/teses/disponiveis/10/10134/tde-02102018-112134/
23. Diário [Internet]. [cited 2023 Jun 15]. Available from: https://diariooficial.guarulhos.sp.gov.br/index.php?mes=7&ano=2016
24. Stärk KD, Regula G, Hernandez J, Knopf L, Fuchs K, Morris RS, Davies P. Concepts for risk-based surveillance in the field of veterinary medicine and veterinary public health: Review of current approaches. BMC Health Serv Res [Internet]. 2006 Dec 1 [cited 2023 Jun 15];6(1):20. Available from: https://doi.org/10.1186/1472-6963-6-20
25. IBGE | Cidades@ | São Paulo | Guarulhos | Panorama [Internet]. [cited 2023 Jun 15]. Available from: https://cidades.ibge.gov.br/brasil/sp/guarulhos/panorama
26. Jargowsky P. A Measure of Spatial Segregation : The Generalized Neighborhood Sorting Index. 2005 [cited 2023 Jun 15]. Available from: https://www.semanticscholar.org/paper/A-Measure-of-Spatial-Segregation-%3A-The-Generalized-Jargowsky/52a99e2c105a95ae06168aaba2fc8ce11b578fcc
27. IPVS - Governo do Estado de São Paulo [Internet]. [cited 2023 Jun 15]. Available from: http://ipvs.seade.gov.br/view/index.php
28. Riebler A, Sørbye SH, Simpson D, Rue H. An intuitive Bayesian spatial model for disease mapping that accounts for scaling. Stat Methods Med Res. 2016 Aug;25(4):1145–1165. PMID: 27566770
29. Baquero OS, Machado G. Spatiotemporal dynamics and risk factors for human Leptospirosis in Brazil. Sci Rep [Internet]. Nature Publishing Group; 2018 Oct 11 [cited 2023 Jun 15];8(1):15170. Available from: https://www.nature.com/articles/s41598-018-33381-3
30. Simpson D, Rue H, Riebler A, Martins TG, Sørbye SH. Penalising model component complexity: a principled, practical approach to constructing priors. Stat Sci [Internet]. Institute of Mathematical Statistics; 2017 Feb [cited 2023 Jun 15];32(1):1–28. Available from: https://research.monash.edu/en/publications/penalising-model-component-complexity-a-principled-practical-appr
31. Wickham H, Hester J, Chang W, Bryan J, RStudio. devtools: Tools to Make Developing R Packages Easier [Internet]. 2022 [cited 2023 Jun 15]. Available from: https://cran.r-project.org/web/packages/devtools/index.html
32. Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, Grolemund G, Hayes A, Henry L, Hester J, Kuhn M, Pedersen TL, Miller E, Bache SM, Müller K, Ooms J, Robinson D, Seidel DP, Spinu V, Takahashi K, Vaughan D, Wilke C, Woo K, Yutani H. Welcome to the Tidyverse. J Open Source Softw [Internet]. 2019 Nov 21 [cited 2023 Jun 15];4(43):1686. Available from: https://joss.theoj.org/papers/10.21105/joss.01686
33. Grolemund G, Wickham H. Dates and Times Made Easy with lubridate. J Stat Softw [Internet]. 2011 Apr 7 [cited 2023 Jun 15];40:1–25. Available from: https://doi.org/10.18637/jss.v040.i03
34. Rue H, Martino S, Chopin N. Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations. J R Stat Soc Ser B Stat Methodol [Internet]. 2009 [cited 2023 Jun 15];71(2):319–392. Available from: https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1467-9868.2008.00700.x
35. oswaldosantos/INLAOutputs: Process selected outputs from the “INLA” package version 1.4.11 from GitHub [Internet]. [cited 2023 Jul 5]. Available from: https://rdrr.io/github/oswaldosantos/INLAOutputs/
36. Pebesma E, Rundel C, Teucher A, developers liblwgeom. lwgeom: Bindings to Selected “liblwgeom” Functions for Simple Features [Internet]. 2023 [cited 2023 Jun 15]. Available from: https://cran.r-project.org/web/packages/lwgeom/index.html
37. Wilke CO. cowplot: Streamlined Plot Theme and Plot Annotations for “ggplot2” [Internet]. 2020 [cited 2023 Jun 15]. Available from: https://cran.r-project.org/web/packages/cowplot/index.html
38. Oliveira SV de, Pereira SVC, Pinna FV, Fonseca LX, Serra-Freire NM da, Cardoso KM, Borsoi ABP, Amorim M, Caldas EP de, Gazeta GS. Vigilância de ambientes da febre maculosa: explorando as áreas silenciosas do Brasil. Rev Pan-Amaz Saúde [Internet]. 2016 Sep [cited 2023 Jun 16];7(3):65–72. Available from: http://scielo.iec.gov.br/scielo.php?script=sci_abstract&pid=S2176-62232016000300065&lng=en&nrm=iso&tlng=pt
39. Maia-Elkhoury ANS, Carmo EH, Sousa-Gomes ML, Mota E. Análise dos registros de leishmaniose visceral pelo método de captura-recaptura. Rev Saúde Pública [Internet]. 2007 Dec 1 [cited 2023 Jun 16];41(6):931–937. Available from: https://www.revistas.usp.br/rsp/article/view/32333
40. Nunes ML, Oliveira SV de, Elkhoury M da R, Fonseca LX, Pereira SVC, Caldas EP de, Guimarães JC do N, Rosa EST da, Bonvicino CR, D’Andrea PS. Evidência de circulação de hantavírus em área silenciosa da Região Amazônica. Rev Pan-Amaz Saúde [Internet]. 2015 [cited 2023 Jun 16];6(4):5–5. Available from: https://ojs.iec.gov.br/index.php/rpas/article/view/240
41. Carvalho CN, Dourado I, Bierrenbach AL. Subnotificação da comorbidade tuberculose e aids: uma aplicação do método de linkage. Rev Saúde Pública [Internet]. Faculdade de Saúde Pública da Universidade de São Paulo; 2011 Jun [cited 2023 Jun 16];45:548–555. Available from: https://www.scielo.br/j/rsp/a/QvJfrM3hWxxYjzPTM7fMSnC/?format=html
42. Carmo RA do, Policena GM, Alencar GP, França EB, Bierrenbach AL. Subnotificação de óbitos por AIDS no Brasil: linkage dos registros hospitalares com dados de declaração de óbito. Ciênc Saúde Coletiva [Internet]. ABRASCO - Associação Brasileira de Saúde Coletiva; 2021 Apr 19 [cited 2023 Jun 16];26:1299–1310. Available from: https://www.scielo.br/j/csc/a/GVcT84zxF95gXL5xkTcWnXk/
43. Nota Técnica no 60/2023-CGZV/DEDT/SVSA/MS — Ministério da Saúde [Internet]. [cited 2023 Jun 22]. Available from: https://www.gov.br/saude/pt-br/centrais-de-conteudo/publicacoes/notas-tecnicas/2023/nota-tecnica-no-60-2023-cgzv-dedt-svsa-ms/view
44. Larsson CE, Gonçalves M de A, Araujo VC, Dagli MLZ, Correa B, Fava Neto C. Esporotricosis felina: aspectos clínicos e zoonóticos. Rev Inst Med Trop São Paulo [Internet]. Instituto de Medicina Tropical de São Paulo; 1989 Oct [cited 2023 Jun 16];31:351–358. Available from: https://www.scielo.br/j/rimtsp/a/HmwXLmmqqC36FThFDRHd4Cy/
45. Madon S, Malecela MN, Mashoto K, Donohue R, Mubyazi G, Michael E. The role of community participation for sustainable integrated neglected tropical diseases and water, sanitation and hygiene intervention programs: A pilot project in Tanzania. Soc Sci Med [Internet]. 2018 Apr 1 [cited 2023 Jun 22];202:28–37. Available from: https://www.sciencedirect.com/science/article/pii/S0277953618300728