Estimating hidden populations by transferring knowledge from geographically misaligned levels.

The estimation of hidden sub-populations is a hard task that appears in many fields. For example, public health planning in Brazil depends crucially of the number of people who holds a private health insurance plan and hence rarely uses the public services. Different sources of information about these sub-populations may be available at different geographical levels. The available information can be transferred between these different geographic levels to improve the estimation of the hidden population size. In this study, we propose a model that use individual level information to learn about the dependence between the response variable and explanatory variables by proposing a family of link functions with asymptotes that are flexible enough to represent the real aspects of the data and robust to departures from the model. We use the fitted model to estimate the size of the sub-population at any desired level. We illustrate our methodology estimating the sub-population that uses the public health system in each neighborhood of large cities in Brazil.

Where did I get dengue? Detecting spatial clusters of infection risk with social network data.

Typical spatial disease surveillance systems associate a single address to each disease case reported, usually the residence address. Social network data offers a unique opportunity to obtain information on the spatial movements of individuals as well as their disease status as cases or controls. This provides information to identify visit locations with high risk of infection, even in regions where no one lives such as parks and entertainment zones. We develop two probability models to characterize the high-risk regions. We use a large Twitter dataset from Brazilian users to search for spatial clusters through analysis of the tweets' locations and textual content. We apply our models to both real-world and simulated data, demonstrating the advantage of our models as compared to the usual spatial scan statistic for this type of data.

Schistosoma mansoni reinfection: Analysis of risk factors by classification and regression tree (CART) modeling.

Praziquantel (PZQ) is an effective chemotherapy for schistosomiasis mansoni and a mainstay for its control and potential elimination. However, it does not prevent against reinfection, which can occur rapidly in areas with active transmission. A guide to ranking the risk factors for Schistosoma mansoni reinfection would greatly contribute to prioritizing resources and focusing prevention and control measures to prevent rapid reinfection. The objective of the current study was to explore the relationship among the socioeconomic, demographic, and epidemiological factors that can influence reinfection by S. mansoni one year after successful treatment with PZQ in school-aged children in Northeastern Minas Gerais state Brazil. Parasitological, socioeconomic, demographic, and water contact information were surveyed in 506 S. mansoni-infected individuals, aged 6 to 15 years, resident in these endemic areas. Eligible individuals were treated with PZQ until they were determined to be negative by the absence of S. mansoni eggs in the feces on two consecutive days of Kato-Katz fecal thick smear. These individuals were surveyed again 12 months from the date of successful treatment with PZQ. A classification and regression tree modeling (CART) was then used to explore the relationship between socioeconomic, demographic, and epidemiological variables and their reinfection status. The most important risk factor identified for S. mansoni reinfection was their "heavy" infection at baseline. Additional analyses, excluding heavy infection status, showed that lower socioeconomic status and a lower level of education of the household head were also most important risk factors for S. mansoni reinfection. Our results provide an important contribution toward the control and possible elimination of schistosomiasis by identifying three major risk factors that can be used for targeted treatment and monitoring of reinfection. We suggest that control measures that target heavily infected children in the most economically disadvantaged households would be most beneficial to maintain the success of mass chemotherapy campaigns.

Detecção de anomalias nos pagamentos do sistema único de saúde / Anomaly detection in the sistema único de saúde payments / Detección de anomalías en los pagos del sistema único de saúde

OBJETIVOS: este artigo descreve o INFOSAS, um sistema moderno, interativo e automático de detecção de discrepâncias no sistema de pagamento aos prestadores de serviços aos Sistema Único de Saúde (SUS) para posterior auditoria e verificação. MÉTODOS: Algoritmos estatísticos de mineração de dados são aplicados aos dados do SUS cobrindo 269 grupos de procedimentos médicos em 5570 municípios e mais de 23 mil prestadores de saúde, num total de mais de 1.5 milhões de séries temporais. RESULTADOS: Encontramos 6811 prestadores com valores considerados excedentes e discrepantes.Este grupo de prestadores é bastante desigual. O valor excedente concentrado nos 100 prestadores mais críticos é de 210 milhões de reais, ou 51% do total excedente estimado. CONCLUSÃO: O sistema INFOSAS pode ser utilizado no processo de indicação de casos para auditoria, melhorando a sua qualidade e reduzindo a frequência de auditorias desnecessárias.

AIMS: this paper describes INFOSAS, a modern, interactive and automatic outlier detection in the payment system to the Sistema Único de Saúde (SUS) services providers for subsequent audit and verification. METHODS: Weapply statistical data mining algorithms to SUS data covering 269 groups of medical procedures in 5570 municipalitie sand more than 23,000 health care providers, summing up more than 1.5 million time series. RESULTS: We found 6811 providers with amounts considered excessive and discrepant. This group of providers is quite un even. The surplus value concentrated in the 100 most critical providers is 210 million of reais, or 51% of the total surplus estimated. CONCLUSION:The INFOSAS system can be used to point out to cases for auditing process, improving their quality and reducing the frequency of unnecessary audits.

A critical look at prospective surveillance using a scan statistic.

The scan statistic is a very popular surveillance technique for purely spatial, purely temporal, and spatial-temporal disease data. It was extended to the prospective surveillance case, and it has been applied quite extensively in this situation. When the usual signal rules, as those implemented in SaTScan(TM) (Boston, MA, USA) software, are used, we show that the scan statistic method is not appropriate for the prospective case. The reason is that it does not adjust properly for the sequential and repeated tests carried out during the surveillance. We demonstrate that the nominal significance level α is not meaningful and there is no relationship between α and the recurrence interval or the average run length (ARL). In some cases, the ARL may be equal to ∞, which makes the method ineffective. This lack of control of the type-I error probability and of the ARL leads us to strongly oppose the use of the scan statistic with the usual signal rules in the prospective context.

Relative risk estimates from spatial and space-time scan statistics: are they biased?

The purely spatial and space-time scan statistics have been successfully used by many scientists to detect and evaluate geographical disease clusters. Although the scan statistic has high power in correctly identifying a cluster, no study has considered the estimates of the cluster relative risk in the detected cluster. In this paper, we evaluate whether there is any bias on these estimated relative risks. Intuitively, one may expect that the estimated relative risks has upward bias, because the scan statistic cherry picks high rate areas to include in the cluster. We show that this intuition is correct for clusters with low statistical power, but with medium to high power, the bias becomes negligible. The same behavior is not observed for the prospective space-time scan statistic, where there is an increasing conservative downward bias of the relative risk as the power to detect the cluster increases.

Bayes plus Brass: estimating total fertility for many small areas from sparse census data.

Estimates of fertility in small areas are valuable for analysing demographic change, and important for local planning and population projection. In countries lacking complete vital registration, however, small-area estimates are possible only from sparse survey or census data that are potentially unreliable. In these circumstances estimation requires new methods for old problems: procedures must be automated if thousands of estimates are required; they must deal with extreme sampling variability in many areas; and they should also incorporate corrections for possible data errors. We present a two-step procedure for estimating total fertility in such circumstances and illustrate it by applying the method to data from the 2000 Brazilian Census for over 5,000 municipalities. Our proposed procedure first smoothes local age-specific rates using Empirical Bayes methods and then applies a new variant of Brass's P/F parity correction procedure that is robust to conditions of rapid fertility decline. Supplementary material at the project website ( http://schmert.net/BayesBrass ) will allow readers to replicate all the authors' results in this paper using their data and programs.

Mapping the Timing, Pace, and Scale of the Fertility Transition in Brazil.

Between 1960 and 2000, fertility fell sharply in Brazil, but this transition was unevenly distributed in space and time. Using Bayesian spatial statistical methods and microdata from five censuses, we develop and apply a procedure for fitting logistic curves to the fertility transitions in more than 500 small regions of Brazil over this 40-year period. Doing so enables us to map the main features of the Brazilian fertility transition in considerable detail. We detect early declines in some regions of the country and document large differences between early and late transitions in regard to both the initial level of fertility and the speed of the transition. We also use our results to test hypotheses regarding changes in the level of development at the onset of the fertility transition and identify a temporary stall in the Brazilian transition that occurred in the late 1990s. A web site with project details is at http://schmert.net/BayesLogistic.

Empirical Bayes estimation of demographic schedules for small areas.

In this article, we analyze empirical Bayes (EB) methods for estimating small-area rate schedules. We develop EB methods that treat schedules as vectors and use adaptive neighborhoods to keep estimates appropriately local. This method estimates demographic rates for local subpopulations by borrowing strength not only from similar individuals elsewhere but also from other groups in the same area and from regularities in schedules across locations. EB is substantially better than standard methods when rates have strong spatial and age patterns. We illustrate this method with estimates of age-specific fertility schedules for over 3,800 Brazilian municipalities.

[Regionalization as a strategy for the definition of homicide-control public policies]. / Regionalização como estratégia para a definição de políticas públicas de controle de homicídios.

This article analyzes the spatial distribution of homicide rates in Minas Gerais State, Brazil, using data from the Ministry of Health Mortality Information System (SIM/MS) from 1996 to 2000. Population data used to calculate rates were also obtained from the Brazilian Ministry of Health. Minas Gerais has a large territory, so the study used a recently developed statistical methodology implemented in the SKATER software to generate homogeneous spatial clusters. The technique obtained 24 spatial clusters, in which municipalities with similar homicide rates were aggregated. Using these results, the authors discuss the possible creation of "public security administration areas" that would allow the implementation of public policies aimed at controlling and diminishing homicides in municipalities with similar characteristics for this type of crime.

Regionalização como estratégia para a definição de políticas públicas de controle de homicídios / Regionalization as a strategy for the definition of homicide-control public policies

Neste artigo, analisamos a distribuição espacial das taxas de homicídios no Estado de Minas Gerais, Brasil, utilizando dados do Sistema de Informações de Mortalidade do Ministério da Saúde (SIM-MS), no período de 1996 a 2000. Os dados populacionais utilizados para o cálculo das taxas também foram obtidos através do MS. Diante da grande extensão territorial do Estado, utilizamos uma metodologia estatística desenvolvida recentemente, implementada no programa SKATER, para a geração de conglomerados espaciais homogêneos. Como resultado, obtivemos 24 conglomerados espaciais, nos quais os municípios semelhantes em relação às taxas de homicídios ficaram agrupados. A partir desses resultados, discutimos a possibilidade da criação de "Núcleos de Gerenciamento em Segurança Pública", que permitiriam a implementação de políticas públicas voltadas para o controle e a diminuição de homicídios em cidades com características semelhantes no que diz respeito a esse tipo de crime.

Multiple cancer sites incidence rates estimation using a multivariate Bayesian model.

BACKGROUND: In Brazil cancer incidence rates have to be estimated from occasional surveys, due to lack of continuous cancer registries. Many estimated rates have very large variances, because only few years of data were collected. When dealing with a single cancer site, it is possible to adopt a Bayesian method which borrows information about the cancer rates from other geographical areas to estimate the cancer rate in a given area. We suggest an additional improvement to this method which explores the correlation between multiple cancer sites rates in a same area and in different areas. METHODS: Our method works with a multivariate vector of different cancer sites rates in several areas and it borrows information from both, across geographical areas and across different cancer sites. We applied our method to data from a survey carried out in 18 Brazilian cities in São Paulo State in 1991. We estimated age and sex indirect standardized incidence rates for the six most common cancers in men and women, and calculated the 95% interval estimation for the incidence rates. RESULTS: The usual indirect standardized incidence rates had very large confidence intervals for many cancers and cities due to small expected number of cases. The use of the multivariate Bayesian method led to more precise estimates. CONCLUSIONS: More precise age-standardized cancer incidence rates can be calculated using data from other cancers. The method is conceptually simple, easy to perform, has low cost, and can improve substantially the estimation of cancer incidence and other vital rates.

[Data comparison on homicide deaths between two information systems, Brazil]. / Comparação de dados sobre homicídios entre dois sistemas de informação, Minas Gerais.

INTRODUCTION: There is much criticism on the quality of criminal data available in Brazil, but there lacks systematic studies to corroborate it. Thus, it was conducted a comparison between homicides registered in Minas Gerais by two public information systems. METHODS: The number of homicide deaths of the Mortality Information System of the Ministry of Health (SIM/MS) and Minas Gerais Military Police (PMMG), from 1979 to 1998, were assessed. The comparison between the two systems was performed using the ratio between absolute numbers, considering the municipalities categorized in five groups according to the population size in 1998, and homicide rates. Two linear regression models were adjusted for rates in time, one for each system. RESULTS: In cities with less than 100,000 inhabitants PMMG homicide data is more complete than SIM/MS one. In cities with more than 100,000 inhabitants, SIM/MS registry is capable of providing homicide deaths data more efficiently, although it can include homicide deaths perpetrated in other municipalities and violent deaths of unknown nature that have not been due to homicide. The under-registration of PMMG in these municipalities can be due to late deaths in hospitals that are not followed up by PMMG. CONCLUSIONS: PMMG and SIM/MS systems represent the lower and upper limits of actual homicide deaths in larger cities. Both data sources should be analyzed and compared when both are available; when only one of them is available, the analysis should take into account the differences indicated in this study.

Comparação de dados sobre homicídios entre dois sistemas Sistema de Informações, Minas Gerais / Data comparison on homicide deaths between two information systems, Brazil

INTRODUÇÃO: Existem muitas críticas à qualidade dos dados criminais existentes no Brasil, mas há pouco estudo sistemático para corroborar essas críticas. Nesse sentido, foi feito estudo com o objetivo de comparar o número de homicídios registrados entre dois sistemas públicos de informaçäo. MÉTODOS: Foram analisados os óbitos registrados pelo Sistema de Informaçöes de Mortalidade do Ministério da Saúde (SIM/MS) e pela Polícia Militar (PMMG), de 1979 a 1998. A comparaçäo entre os dois sistemas foi feita pela razäo entre os números absolutos, categorizando os municípios em cinco grupos de acordo com o tamanho da populaçäo em 1998, e pelas taxas de homicídio. Foram ajustados dois modelos de regressäo linear das taxas no tempo, um para cada sistema. RESULTADOS: Nas cidades de menos de 100.00 habitantes, as informaçöes de homicídios da PMMG säo mais completas do que as do SIM/MS. Nas cidades de mais de 100.000 habitantes, o SIM/MS é capaz de recuperar mais eficientemente os óbitos, embora possam ser incluídos indevidamente óbitos de homicídios cometidos em outros municípios e óbitos por arma branca e de fogo de intençäo indeterminada que näo tenham sido devido a homicídio. O sub-registro da PMMG pode ser devido às mortes hospitalares tardias, que näo säo acompanhadas pela PMMG. CONCLUSOES: Os sistemas do SIM/MS e PMMG representam limites superiores e inferiores do número real de homicídios ocorridos nas grandes cidades. Sugere-se que, quando disponíveis, as duas fontes sejam analisadas e comparadas; quando apenas uma delas estiver disponível, propöe-se que a análise deva considerar as diferenças

A Bayesian space varying parameter model applied to estimating fertility schedules.

We propose a spatial generalized linear model (GLM) to analyse the vital rates for small areas. In each small area, we have a response vector and covariates to explain its variability. The statistical methodology is based on a spatial Bayesian approach and it allows the covariates' parameters of the generalized linear model to vary smoothly on space. Hence, the effect of a covariate on the response varies depending on the random variables measurement location. Our model is an extension of disease mapping models allowing the space-covariate interaction to be modelled in a natural way and giving space a position of intrinsic interest. We introduce the model in the context of fertility curve estimation. In each small area, we have a curve describing the variation of fertility rates by age modelled by Coale's fertility model, which implies a GLM in each area. A simulation shows the advantages of our approach. In addition, the paper applies the procedure to census data used to study the diffusion of low fertility behaviour in Brazil.