Full text
| |
| Author(s): |
Neuber José Segri
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | São Paulo. |
| Institution: | Universidade de São Paulo (USP). Faculdade de Saúde Pública (FSP/CIR) |
| Defense date: | 2013-02-27 |
| Examining board members: |
Chester Luiz Galvão César;
Maria Cecilia Goi Porto Alves;
Marilisa Berti de Azevedo Barros;
Jair Licio Ferreira Santos;
Jose Maria Pacheco de Souza
|
| Advisor: | Chester Luiz Galvão César |
| Abstract | |
Introduction. Nowadays population-based health surveys employ different strategies in order to obtain data, such as household and telephone interviews with the purpose of monitoring risk factors and evaluate the access and utilization of health services, being essential in planning public health policies. Objective. Study the implementation of some statistical tools used for comparison of estimates obtained by different surveys, creation of post-stratification adjustments and small-area estimation techniques, using data from different population-based health surveys conducted in areas of the state of São Paulo. Methodology. Using the svy commands of Stata (10.0 and 11.0), three studies were carried out. The first (paper 1) compared estimates according to the type of the survey (household survey ISACapital 2008 and telephone survey VIGITEL-SP 2008) using Poisson regression analysis adjusted by age and education. The second (paper 2) compared post-stratification adjustments utilizing different sets of variables and three different strategies (cell weighting technique, rake technique and a third \"combined\" between the rake and cell weighting). The third (paper 3) used five (direct and indirect) small-area estimation techniques for the calculation of prevalence of health characteristics for a smaller area of São Paulo (Butantã Health District). Results. There were no statistically significant differences between the estimates obtained by VIGITEL and ISACapital for the prevalence of mammography in the year prior to the interview. However, estimates for the global results of the Pap smear at least once in life and in the past year as well as, mammography in life, we observed differences, with higher prevalence rates among respondents by telephone (paper 1). Applying the rake post-stratification technique, the largest reductions were observed in bias, especially when taking into account sociodemographic variables associated with each health characteristic analyzed (paper 2). The estimates for small areas obtained by calibration and regression, considering the factors associated with each health characteristic, were the ones most similar to the prevalence considered to be the true population values (paper 3). Conclusions. The results of the comparisons between the two surveys indicate the trend of overestimation in some indicators of prevalence of mammography and Pap smear via telephone survey, indicating a concern for further studies that also contribute to a better understanding of the bias and possible corrections with new post-stratification adjustments (paper 1). Despite the post-stratification adjustments do not completely correct the estimates, it should not be considered an impediment, since the telephone surveys contribute to the direct actions and new health policies in Brazil. (paper 2). The use of smallarea estimation techniques allows extrapolating the utilization of a research that had not been designed for such a purpose, identifying the needs of a particular region, contributing to the implementation of preventive interventions in public health at the local level. (paper 3) (AU) | |
| FAPESP's process: | 10/11463-3 - Social Inequalities in Health in São Paulo city - A population-based study (ISA-CAPITAL 2008) |
| Grantee: | Neuber José Segri |
| Support Opportunities: | Scholarships in Brazil - Doctorate |