Structural equation modeling in epidemiology / Modelagem de equações estruturais em epidemiologia

Structural equation modeling (SEM) is an important statistical tool for evaluating complex relations in several research areas. In epidemiology, the use and discussion of SEM have been limited thus far. This article presents basic principles and concepts in SEM, including an application using epidemiological data analysis from a study on the determinants of cognitive development in young children, considering constructs related to organization of the child's home environment, parenting style, and the child's health status. The relations between the constructs and cognitive development were measured. The results showed a positive association between psychosocial stimulus at home and cognitive development in young children. The article presents the contributions by SEM to epidemiology, highlighting the need for an a priori theoretical model for improving the study of epidemiological questions from a new perspective.

A modelagem de equações estruturais (MEE) é uma ferramenta estatística importante para avaliar relações complexas em várias áreas do conhecimento. Em Epidemiologia sua divulgação e uso são limitados. Este artigo apresenta princípios e conceitos básicos da MEE, com exemplo de aplicação na análise de dados epidemiológicos. A análise de dados é realizada em estudo que investiga determinantes do desenvolvimento cognitivo infantil, sendo definidos construtos relacionados à organização do ambiente da criança, ao seu status de saúde, e às práticas e estilo de vida dos pais. O impacto positivo da qualidade de estimulação psicossocial do ambiente doméstico sobre o índice de desempenho cognitivo (IDC) esclarece que parte do efeito da estimulação sobre o IDC deve-se ao estilo parental de interação com a criança e às características físico-ambientais do contexto familiar. As potencialidades do uso da MEE em Epidemiologia são apresentadas, enfatizando-se a definição do modelo teórico e seu uso para aprofundamento de questões epidemiológicas sob nova perspectiva.

Structural equation modeling in epidemiology.

Structural equation modeling (SEM) is an important statistical tool for evaluating complex relations in several research areas. In epidemiology, the use and discussion of SEM have been limited thus far. This article presents basic principles and concepts in SEM, including an application using epidemiological data analysis from a study on the determinants of cognitive development in young children, considering constructs related to organization of the child's home environment, parenting style, and the child's health status. The relations between the constructs and cognitive development were measured. The results showed a positive association between psychosocial stimulus at home and cognitive development in young children. The article presents the contributions by SEM to epidemiology, highlighting the need for an a priori theoretical model for improving the study of epidemiological questions from a new perspective.

Periodontal therapy for pregnant women and cases of low birthweight: an intervention study.

BACKGROUND: Over the past decade, strong evidence for an association between maternal periodontitis and low birthweight has started to appear. However, few intervention studies have been proposed for investigating this hypothesis. The aim of this study was to evaluate whether periodontal therapy among pregnant women would reduce the incidence of low birthweight. METHODS: A nonrandomized intervention study was performed, with two control groups. The sample comprised 339 pregnant women: 141 in the experimental group (treated for periodontitis), 145 in control group 1 (without periodontitis) and 53 in control group 2 (with untreated periodontitis). The experimental group received periodontal treatment throughout pregnancy, whereas control group 1 was only monitored over the same period. After delivery, birthweight information on the newborns was obtained. The analysis procedures consisted of stratified analysis followed by logistic regression. RESULTS: The frequency of low birthweight among the women with treated periodontitis was 9.22%, while it was 13.10% in the group without periodontal disease. However, the difference was not statistically significant (RR 0.72; 95% CI 0.36-1.45). The occurrence of this outcome in the group with untreated periodontitis (24.53%) was greater than in the other two groups. CONCLUSION: This suggests that periodontal therapy is a protective factor for birthweight.

Contribution of periodontal disease in pregnant women as a risk factor for low birth weight.

UNLABELLED: Starting in the 1990s, several authors sought to investigate the hypothesis that periodontitis during pregnancy may contribute towards the birth of low-weight children. However, this relationship is still not well established. The aim of this study was to evaluate whether this oral infection is associated with this gestational event. METHODS: This was a case-control study among 548 puerperae, of whom 164 were the mothers of low-weight live births (case group) and 384 were the mothers of live births of normal gestational weight (control group). They were selected at two public hospital units in two municipalities in the State of Bahia. From interviews and data gathered using live birth cards or birth certificates, information was obtained regarding age, height, previous diseases, marital status, socioeconomic situation, smoking and alcohol use. Mothers who presented at least four teeth on which one or more sites had a probing depth of greater than or equal to 4 mm, clinical attachment loss of greater than or equal to 3 mm and bleeding on probing, at the same site, were deemed to present periodontal disease. The data were analysed by stratification from logistic regression. RESULTS: Periodontal disease was diagnosed in 42.7% of the case group and 30% of the control group. A statistically significant association was found between periodontal disease and low birth weight (unadjusted OR = 1.74; 95% CI: 1.19-2.54), particularly among mothers with low schooling levels (adjusted OR = 2.30; 95% CI: 1.14-4.6). CONCLUSION: The findings suggest an association between periodontal disease and low birth weight among mothers with low education levels.

Estimating adjusted prevalence ratio in clustered cross-sectional epidemiological data.

BACKGROUND: Many epidemiologic studies report the odds ratio as a measure of association for cross-sectional studies with common outcomes. In such cases, the prevalence ratios may not be inferred from the estimated odds ratios. This paper overviews the most commonly used procedures to obtain adjusted prevalence ratios and extends the discussion to the analysis of clustered cross-sectional studies. METHODS: Prevalence ratios(PR) were estimated using logistic models with random effects. Their 95% confidence intervals were obtained using delta method and clustered bootstrap. The performance of these approaches was evaluated through simulation studies. Using data from two studies with health-related outcomes in children, we discuss the interpretation of the measures of association and their implications. RESULTS: The results from data analysis highlighted major differences between estimated OR and PR. Results from simulation studies indicate an improved performance of delta method compared to bootstrap when there are small number of clusters. CONCLUSION: We recommend the use of logistic model with random effects for analysis of clustered data. The choice of method to estimate confidence intervals for PR (delta or bootstrap method) should be based on study design.

Exposure measurement in the association between periodontal disease and prematurity/low birth weight.

AIM: To compare the use of different definitions for exposure measurement in cases of association between periodontal disease (PD) and prematurity and/or low birth weight (PLBW). MATERIAL AND METHODS: A database from a previous case-control study was used to compare four different definitions for periodontitis: at least one site with probing depth > or =4 mm (1); at least one site with clinical attachment loss (CAL)> or =3 mm (2); at least four teeth with one or more sites presenting probing depth > or =4 mm, with CAL> or =3 mm at the same site (3); and at least four teeth with one or more sites with probing depth > or =4 mm, with CAL> or =3 mm at the same site and presence of bleeding on probing (4). The PD frequency, diagnostic values and adjusted association measurements were calculated. RESULTS: PD frequency ranged from 33.1% to 94.7%. Odds ratio(adjusted) varied slightly according to the exposure measurement used. CONCLUSIONS: The association between PD and PLBW weight was consistent, except for exposure measurement 1, i.e. using at least one site with CAL> or =3 mm for periodontitis diagnosis, while the magnitude of this varied according to the definition established.

O perfil da violência contra crianças e adolescentes, segundo registros de Conselhos Tutelares: vítimas, agressores e manifestações de violência / The profile of violence against children and adolescents according to Child Protection Council records: victims, aggressors and patterns of violence

OBJETIVO: estimar a prevalência das formas de violência contra crianças e adolescentes, registradas nos Conselhos Tutelares, e a associação dessas violências por faixas etárias das vítimas e vínculo com agressores, em 2003-2004. MÉTODO: foram coletados dados dos prontuários e calculadas as prevalências e associação entre variáveis, através da razão de prevalência (RP), com nível crítico de 5 por cento. RESULTADOS: totalizaram 1.293 registros de violência, sendo 1.011 (78,1 por cento) originados no domicílio. As violências mais freqüentes foram a negligência (727), por omissão de cuidados básicos (304) e abandono (259); a violência física (455), por espancamento (392), nas faixas de 2 a 13 anos; a violência psicológica (374), por amedrontamento (219); a violência sexual (68), por abuso (58), principalmente entre adolescentes. A principal forma de denúncia foi anônima, 398 (30,8 por cento); os agressores para negligência foram os pais; para violência física, a madrasta e "outros agressores"; para violência sexual, o padrasto, "outros familiares/ agressores"; a violência psicológica foi prevalente entre todas categorias de agressores. CONCLUSÕES: Os resultados apontam para a necessidade de divulgação do "Disque Denúncia"; a formação de conselheiros, quanto ao registro adequado, assim como a implementação de políticas de prevenção da violência contra crianças e adolescentes.

OBJECTIVE: To estimate the prevalence of violence against children and adolescents through the records of Child Protection Councils, associating this violence with the age of the victims and their links with the aggressor. METHOD: Data were collected from the records and the prevalence rates were calculated through the Prevalence Ratio (PR), together with the associations among variables, with a critical level of 5 percent. RESULTS: Out of 1,293 records of violence, 1,011 (78.1 percent), originated in the home. The most frequent types of violence were: neglect (727), failure to provide basic care (304) and abandonment (259); physical violence (455), beatings (392) between 2 and 13 years old; psychological violence (374) through threats (219); sexual violence (68) through abuse (58), mainly among adolescents. Denunciations were mainly (398) anonymous (30.8 percent); the aggressors through neglect were the parents; through physical violence, the stepmother and "other aggressors"; through sexual violence, the stepfather and "other relatives / aggressors"; psychological violence was prevalent among all aggressor categories. CONCLUSIONS: The findings indicate the need to extend the Anti-Violence Hot-line facilities, training Council Members in terms of keeping proper records, and implementing policies designed to prevent violence against children and adolescents.

The association between postmenopausal osteoporosis and periodontal disease.

BACKGROUND: The clinical importance of systemic bone loss as a contributory factor to alveolar bone loss and the subsequent loss of teeth merits further study, given that osteoporosis and periodontal disease lead to significantly increased morbidity and mortality and higher public expenditure of funds. This case-control study evaluated the association between osteoporosis and periodontal disease. METHODS: The sample consisted of 139 postmenopausal women: 48 in the case group (with periodontal disease) and 91 in the control group (without periodontal disease). The diagnosis of periodontal disease was established following a complete clinical examination using measurements of probing depth, gingival recession and hyperplasia, clinical attachment loss, and bleeding index, and confirmed by panoramic radiography. The diagnosis of osteoporosis was made by reviewing densitometry reports obtained previously. Descriptive, stratified, and logistic regression analyses were applied to the data collected. Comparison of proportions was performed using the chi(2) and Fisher tests. Association measurements (odds ratios [ORs]) with and without adjustment for confounding factors and control for effect modifiers were obtained at a significance level of 5%. RESULTS: The OR(unadjusted) for the principal association was 2.58 (95% confidence interval [CI]: 1.01 to 6.82). In subgroup analyses of the stratified model, the OR(unadjusted) for low education was 6.40 (95% CI: 1.77 to 23.18). When adjusted for smoking habit and age, the OR(adjusted) was 7.05 (95% CI: 1.90 to 26.19), which also was statistically significant. CONCLUSION: Postmenopausal women with osteoporosis and low educational levels have a greater chance of having periodontal disease than do those without osteoporosis.

[The profile of violence against children and adolescents according to Child Protection Council records: victims, aggressors and patterns of violence]. / O perfil da violência contra crianças e adolescentes, segundo registros de Conselhos Tutelares: vítimas, agressores e manifestações de violência.

OBJECTIVE: To estimate the prevalence of violence against children and adolescents through the records of Child Protection Councils, associating this violence with the age of the victims and their links with the aggressor. METHOD: Data were collected from the records and the prevalence rates were calculated through the Prevalence Ratio (PR), together with the associations among variables, with a critical level of 5%. RESULTS: Out of 1,293 records of violence, 1,011 (78.1%), originated in the home. The most frequent types of violence were: neglect (727), failure to provide basic care (304) and abandonment (259); physical violence (455), beatings (392) between 2 and 13 years old; psychological violence (374) through threats (219); sexual violence (68) through abuse (58), mainly among adolescents. Denunciations were mainly (398) anonymous (30.8%); the aggressors through neglect were the parents; through physical violence, the stepmother and "other aggressors"; through sexual violence, the stepfather and "other relatives / aggressors"; psychological violence was prevalent among all aggressor categories. CONCLUSIONS: The findings indicate the need to extend the Anti-Violence Hot-line facilities, training Council Members in terms of keeping proper records, and implementing policies designed to prevent violence against children and adolescents.