ABSTRACT
OBJECTIVE: to evaluate the application of a deterministic routine for identifying multiple pregnancies on the Brazilian Live Birth Information System (SINASC). METHODS: SINASC data deduplication and linkage with the mortality database (fetal deaths) for Rio de Janeiro state for the period 2007-2008; we used a deterministic routine, using a key based on SINASC maternal and birth information, complemented by manual review. RESULTS: of the 433,874 SINASC records, 9,036 (2.1%) were classified as multiple pregnancy newborns; after implementing the routine, we reclassified 385 records as twins, and 286 as singletons; accuracy of multiple pregnancy information on the SINASC database was high (sensitivity=95.8%; specificity=99.9%); applying the routine without the manual review process increased sensitivity by 4.2%, with no significant change of specificity. CONCLUSION: despite the accuracy of information regarding multiple pregnancy held on SINASC, we suggest the use of this routine as an option for improving classification of twins.
Subject(s)
Information Systems/statistics & numerical data , Live Birth , Pregnancy, Multiple/statistics & numerical data , Pregnancy, Twin/statistics & numerical data , Brazil , Data Accuracy , Databases, Factual/standards , Databases, Factual/statistics & numerical data , Female , Humans , Information Systems/standards , PregnancyABSTRACT
INTRODUCTION: Road traffic crashes (RTC) are an important public health problem, accounting for 1.2 million deaths per year worldwide. In Brazil, approximately 40,000 deaths caused by RTC occur every year, with different trends in the Federal Units. However, these figures may be even greater if health databases are linked to police records. In addition, the linkage procedure would make it possible to qualify information from the health and police databases, improving the quality of the data regarding underlying cause of death, cause of injury in hospital records, and injury severity. OBJECTIVE: This study linked different data sources to measure the numbers of deaths and serious injuries and to estimate the percentage of corrections regarding the underlying cause of death, cause of injury, and the severity injury in victims in matched pairs from record linkage in five representative state capitals of the five macro-regions of Brazil. METHODS: This cross-sectional, population-based study used data from the Hospital Information System (HIS), Mortality Information System (MIS), and Police Road Traffic database of Belo Horizonte, Campo Grande, Curitiba, Palmas, and Teresina, for the year 2013 for Teresina, and 2012 for the other capitals. RecLink III was used to perform probabilistic record linkage by identifying matched pairs to calculate the global correction percentage of the underlying cause of death, the circumstance that caused the road traffic injury, and the injury severity of the victims in the police database. RESULTS: There was a change in the cause of injury in the HIS, with an overall percentage of correction estimated at 24.4% for Belo Horizonte, 96.9% for Campo Grande, 100.0% for Palmas, and 33.2% for Teresina. The overall percentages of correction of the underlying cause of death in the MIS were 29.9%, 11.9%, 4.2%, and 33.5% for Belo Horizonte, Campo Grande, Curitiba, and Teresina, respectively. The correction of the classification of injury severity in police database were 100.0% for Belo Horizonte and Teresina, 48.0% for Campo Grande, and 51.4% for Palmas after linkage with hospital database. The linkage between mortality and police database found a percentage of correction of 29.5%, 52.3%, 4.4%, 74.3 and 72.9% for Belo Horizonte, Campo Grande, Palmas, Curitiba and Teresina, respectively in the police records. CONCLUSIONS: The results showed the importance of linking records of the health and police databases for estimating the quality of data on road traffic injuries and the victims in the five capital cities studied. The true causes of death and degrees of severity of the injuries caused by RTC are underestimated in the absence of integration of health and police databases. Thus, it is necessary to define national rules and standards of integration between health and traffic databases in national and state levels in Brazil.
Subject(s)
Accidents, Traffic/mortality , Wounds and Injuries/epidemiology , Brazil/epidemiology , Cause of Death , Cross-Sectional Studies , Databases, Factual , Female , Hospital Information Systems , Humans , Male , Police/statistics & numerical data , Population Surveillance , Urban PopulationABSTRACT
ResumoA qualidade da informação é fundamental no monitoramento e na avaliação das ações de controle dos agravos, como a tuberculose (TB). O objetivo deste trabalho foi analisar a concordância entre o encerramento do Sistema de Informação de Agravos de Notificação (SINAN) e as causas de morte no Sistema de Informação sobre Mortalidade (SIM). Realizou-se um linkage probabilístico entre o SINAN de 2006 e o SIM de 2006 a 2008. A confiabilidade do encerramento foi analisada por meio do índice kappa. Dos 417 casos encerrados por óbito no SINAN, 88,7% foram encontrados no SIM. Dos 82 casos encerrados como óbito por outra causa, 42,7% apresentaram a TB como causa básica ou associada no SIM, enquanto 41,5% não tinham menção à TB. O coeficiente PABAK (Prevalance and Bias Adjusted Kappa) revelou concordância excelente entre o desfecho óbito no campo encerramento do SINAN e a presença ou não do óbito de TB no SIM. Uma recomendação para os Estados e municípios que utilizam o relacionamento entre o SINAN e o SIM para aumentar a completude e a consistência do SINAN-TB é a investigação no SIM não apenas dos casos notificados sem encerramento, mas também dos casos encerrados por abandono e por transferência.
The quality of information is crucial in the monitoring and evaluation of control actions of diseases such as tuberculosis (TB). However, studies have pointed out deficiencies in the quality of data obtained by the TB surveillance system. Thus the aim of this study was to analyze the agreement between treatment outcome of the Notifiable Diseases Information System (SINAN) and information on the underlying and associated cause of death in the Mortality Information System (SIM). We conducted a probabilistic record linkage with the databases of the SINAN 2006 and the SIM 2006-2008. The accuracy of treatment outcome was analyzed by kappa index. Of the 417 cases closed as death in the SINAN, 88.7% were found in the SIM. Of the 82 cases closed as death from other causes, 42.7% presented TB as the underlying or associated cause in the SIM, while 41.5% did not mention TB. PABAK coefficients revealed excellent agreement between the treatment outcome of death of the SINAN and the presence or absence of TB death in the SIM. We recommend that the states and municipalities that use the linkage between the SINAN and SIM to increase the completeness and consistency of the SINAN-TB search on the SIM, not only for cases without treatment outcome, but also for those discontinued and transferred.
ABSTRACT
The objective of this study was to analyze the multiple causes of death in a cohort of patients with tuberculosis (TB) and to introduce an investigation proposal death for TB from a list of presumable causes. We performed a probabilistic record linkage with the databases of the Information System for Notifiable Diseases (SINAN) 2006 and the Mortality Information System (SIM) 2006-2008. There were 825 deaths, of which 23% for death for TB, deaths due to TB with 16% and 61% without mention of TB. Two hundred and fifteen (42.7%) deaths occurred within the period of treatment, whose profile differed from the pattern of causes when TB was an associated cause, with high frequency of respiratory diseases, AIDS and ill-defined causes. We elaborated a proposal for correction of associated causes of death and an investigation proposal death for TB from a list of presumable causes. According to the proposal, 26 deaths could have modified the underlying cause. This study highlights the importance of record linkage to TB surveillance and improvement of information the SIM and SINAN.
Subject(s)
Information Systems/statistics & numerical data , Tuberculosis/mortality , Acquired Immunodeficiency Syndrome/mortality , Adolescent , Adult , Age Factors , Brazil/epidemiology , Cause of Death , Cohort Effect , Disease Notification , Female , Humans , Longitudinal Studies , Male , Middle Aged , Sex Factors , Survival Rate , Young AdultABSTRACT
O objetivo deste trabalho foi analisar as causas múltiplas de morte de uma coorte de pacientes notificados com tuberculose (TB) e apresentar uma proposta de investigação de causas presumíveis. Realizou-se linkage probabilístico entre o Sistema de Informação de Agravos de Notificação (SINAN) 2006 e o Sistema de Informação sobre Mortalidade (SIM), 2006-2008. Ocorreram 825 mortes, das quais 23% por TB, 16% com TB e 61% sem menção da TB. Duzentos e quinze (42,7%) óbitos ocorreram antes do término do esquema básico de tratamento e não tinham menção da TB, cujo perfil foi distinto do padrão quando a TB era uma das causas associadas. A elevada frequência de doenças do aparelho respiratório, AIDS e causas mal definidas sugerem falha na qualidade da informação. Elaborou-se proposta de correção das causas associadas no SIM e de investigação de óbito com base na relação de causas presumíveis. De acordo com a proposta, 26 óbitos poderiam ter a causa básica modificada. Este estudo destaca a gravidade do quadro da TB e a importância do linkage para a vigilância da TB e melhoria das informações do SIM e do SINAN.
The objective of this study was to analyze the multiple causes of death in a cohort of patients with tuberculosis (TB) and to introduce an investigation proposal death for TB from a list of presumable causes. We performed a probabilistic record linkage with the databases of the Information System for Notifiable Diseases (SINAN) 2006 and the Mortality Information System (SIM) 2006-2008. There were 825 deaths, of which 23% for death for TB, deaths due to TB with 16% and 61% without mention of TB. Two hundred and fifteen (42.7%) deaths occurred within the period of treatment, whose profile differed from the pattern of causes when TB was an associated cause, with high frequency of respiratory diseases, AIDS and ill-defined causes. We elaborated a proposal for correction of associated causes of death and an investigation proposal death for TB from a list of presumable causes. According to the proposal, 26 deaths could have modified the underlying cause. This study highlights the importance of record linkage to TB surveillance and improvement of information the SIM and SINAN.
El objetivo fue analizar las múltiples causas de muerte en una cohorte de pacientes con tuberculosis (TB) y presentar una propuesta de investigación de las causas presumibles. Hicieron linkage probabilística entre el Sistema de Información de Enfermedades de Declaración Obligatoria (SINAN) 2006 y el Sistema de Información sobre Mortalidad (SIM) 2006-2008. 825 muertes, de las cuales 23% para TB, 16% con TB y 61% sin mención de TB. Doscientos quince (42,7%) muertes ocurrieron antes del final de el tratamiento primario y tenía ninguna mención de TB, cuyo perfil era distinto del patrón observado cuando la TB era una causa asociada. La alta frecuencia de enfermedades respiratorias, SIDA y causas mal definidas sugieren insuficiencia en la calidad de la información. Hemos preparado propuestas de corrección de la causa asociada en el SIM y de investigación de muerte basado en las causas presumibles. De acuerdo con la propuesta, 26 muertes podrían haber modificado la causa subyacente. Destacase la gravedad de la TB y la importancia de el linkage para la vigilancia y la mejora de la información en el SIM y SINAN-TB.
Subject(s)
Animals , Dogs , Female , Humans , Male , Magnetic Resonance Imaging/methods , Wakefulness/physiology , Echo-Planar Imaging , Reproducibility of Results , Restraint, Physical , RewardABSTRACT
Poisoning is one of the three main means used in suicide and suicides attempts. In order to improve the quality of such information, a cross-sectional descriptive study was conducted on case records in the State of Rio de Janeiro, Brazil, from 2006 to 2008, in the databases of the Information System on Diseases of Notification (SINAN), the Mortality Information System (SIM), and the Poison Control Center in Niterói (CCIn-Niterói). Probabilistic linkage was used to develop a model for monitoring suicides and suicide attempts by poisoning. Although 68.8% of the 948 records at CCIn had been reported by health professionals, only 2.6% were included in the other databases. Linking CCIn and SIM showed 61.3% underreporting. Information is important for surveillance, public policy-making, and decision-making. Probabilistic linkage allowed identifying problems in each system and provided better data quality and a more accurate diagnosis of the real situation in a complex and serious problem like suicidal behavior.
Subject(s)
Health Information Systems , Poisoning/epidemiology , Suicide, Attempted/statistics & numerical data , Brazil/epidemiology , Cross-Sectional Studies , Female , Humans , Male , Suicide/statistics & numerical dataABSTRACT
A intoxicação exógena é um dos três principais meios utilizados nas tentativas e suicídios. Com o objetivo de melhorar a qualidade destas informações foi realizado um estudo seccional descritivo dos registros sobre casos do Estado do Rio de Janeiro, Brasil, período 2006-2008, presentes nos bancos de dados do Sistema de Informações sobre Agravos de Notificação (SINAN), Sistema de Informações sobre Mortalidade (SIM) e Centro de Controle de Intoxicação de Niterói (CCIn-Niterói). Através do método de relacionamento probabilístico desenvolveu-se um modelo de monitoramento dos casos de tentativas/suicídios devido intoxicação. Embora 68,8% dos 948 registros do CCIn tenham sido reportados por profissionais de saúde, apenas 2,6% constavam nos demais sistemas. O pareamento CCIn e SIM apresentou uma subnotificação de 61,3%. Sabe-se da importância da informação para a vigilância, formulação de políticas públicas, além de tomadas de decisão. O método de relacionamento de bancos de dados permitiu a identificação dos problemas existentes em cada sistema, proporcionou melhor qualidade das informações e maior proximidade com a situação real de agravos complexos e graves como o comportamento suicida.
Poisoning is one of the three main means used in suicide and suicides attempts. In order to improve the quality of such information, a cross-sectional descriptive study was conducted on case records in the State of Rio de Janeiro, Brazil, from 2006 to 2008, in the databases of the Information System on Diseases of Notification (SINAN), the Mortality Information System (SIM), and the Poison Control Center in Niterói (CCIn-Niterói). Probabilistic linkage was used to develop a model for monitoring suicides and suicide attempts by poisoning. Although 68.8% of the 948 records at CCIn had been reported by health professionals, only 2.6% were included in the other databases. Linking CCIn and SIM showed 61.3% underreporting. Information is important for surveillance, public policy-making, and decision-making. Probabilistic linkage allowed identifying problems in each system and provided better data quality and a more accurate diagnosis of the real situation in a complex and serious problem like suicidal behavior.
La intoxicación exógena es uno de los tres métodos principales utilizados en los intentos de suicido y suicidios. Con el objetivo de mejorar la calidad de esta información, se realizó un estudio seccional descriptivo de los registros de casos en el estado de Río de Janeiro, Brasil, 2006-2008, presentes en las bases de datos del Sistema de Información sobre Dolencias Notificables (SINAN por sus siglas en portugués), Sistema de Información sobre Mortalidad (SIM) y Centro de Control de Intoxicación de Niterói (CCIn-Niterói). A través del método probabilístico, se desarrolló un modelo para el seguimiento de los casos de intentos/suicidios por envenenamiento. Aunque un 68,8% de la CCIn (948 discos) ha sido reportado por profesionales de salud, sólo un 2,6% de ese tipo contenido está presente en otros sistemas. El emparejamiento CCIn y SIM presentó un subregistro de un 61,3%. Conocemos la importancia de esta información para la vigilancia, formulación de políticas públicas y la toma de decisiones. El método de relación probabilística de bancos de datos permitió la identificación de los problemas existentes en cada sistema, proporcionó una mejor calidad de la información y proximidad a la situación real de las enfermedades graves y complejas como el comportamiento suicida.
Subject(s)
Female , Humans , Male , Health Information Systems , Poisoning/epidemiology , Suicide, Attempted/statistics & numerical data , Brazil/epidemiology , Cross-Sectional Studies , Suicide/statistics & numerical dataABSTRACT
OBJETIVO: Analisar a qualidade do campo município de residência no Sistema de Informações Hospitalares do Sistema Único de Saúde (SIH-SUS). METODOLOGIA: Foram selecionadas as internações de mulheres com diagnóstico principal de câncer de mama e de colo de útero e que morreram durante a internação em hospital do estado do Rio de Janeiro, 2001-2002. Estes registros foram comparados manualmente com os de óbito de residentes no estado do Rio de Janeiro, 2001-2002, considerados a referência para a comparação. RESULTADOS: Dos 808 registros de óbitos hospitalares: 16 (1,9%) não foram encontrados na base do Sistema de Informação sobre Mortalidade (SIM); em 670 (82,9%) os municípios de residência coincidiram; 78 (9,6%) indicavam município de residência Rio de Janeiro diferente do encontrado no SIM; e, destes, 53 (74,6%) eram municípios da Região Metropolitana. O coeficiente kappa foi de 0,80 (IC95% 0,76-0,84). CONCLUSÃO: Observou-se elevada concordância, menor quando a mulher não residia na capital, mas no seu entorno.
OBJECTIVE: To analyze the quality of the field municipality of residence in the Hospital Information System of the Unified Health System (SIH-SUS). METHODS: We selected hospitalizations of women diagnosed with primary breast cancer and cervical cancer and who died during hospital admission in the state of Rio de Janeiro, 2001-2002. These records were compared manually with the death records of residents in the state of Rio de Janeiro, 2001-2002, considered the reference for comparison. RESULTS: Of 808 hospital deaths records: 16 (1.9%) were not found in the Mortality Information System, in 670 (82.9%) municipalities of residence coincided, 78 (9.6%) indicated municipality of residence Rio de Janeiro, different from that found in the SIM, and, of these, 53 (74.6%) were municipalities in the metropolitan region. The kappa coefficient was 0.80 (95%CI 0.76-0.84). CONCLUSION: There was high overall agreement, lower when the woman did not live in the capital proper, but in its surroundings.
ABSTRACT
O objetivo deste trabalho foi analisar a relação entre características dos indivíduos e uso de serviços odontológicos, e se estas relações variaram de acordo com o estado de residência das pessoas. Foram analisados os dados de uso de serviços odontológicos da Pesquisa Nacional de Amostras por Domicílios de 1998, segundo sexo, idade, quintil de renda e unidade da federação, a partir de um modelo de regressão logística hierárquico de coeficientes aleatórios. Foi observado que o uso cresceu com a idade, até a primeira metade da adolescência (10 a 14 anos), decrescendo acentuadamente a partir da faixa etária de 50 a 64 anos. As chances de uso para as mulheres foram 33 por cento maiores do que para os homens e foi observado gradiente acentuado com a renda, tendo sido as chances de uso mais de 5 vezes maiores para os 20 por cento mais ricos, em comparação com os 20 por cento mais pobres. Foi observado que o efeito das características individuais no uso de serviços odontológicos é modificado pelo estado de residência.
