Full text
| |
| Author(s): |
Juliana Araujo Teixeira
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | São Paulo. |
| Institution: | Universidade de São Paulo (USP). Faculdade de Saúde Pública (FSP/CIR) |
| Defense date: | 2009-08-31 |
| Examining board members: |
Dirce Maria Lobo Marchioni;
Maria do Rosario Dias de Oliveira Latorre;
Rosely Sichieri
|
| Advisor: | Dirce Maria Lobo Marchioni |
| Abstract | |
Background. The study Natural History of HPV Infection in Men: The HIM Study will be the first prospective cohort to investigate the natural history of HPV infection in men from EUA, Brazil and Mexico. The quantitative food frequency questionnaire (QFFQ) was the method of choice to investigate the role of dietetic factors on process of infection. Objective. To assess the performance and to estimate the calibration factors from the data of QFFQ used in HIM - Brazil study, by a validation, reproducibility and calibration study. Methods. A random sample of 98 individuals from the HIM study, with age range of 18 to 70 years, by an interview, answered three QFFQ and three 24-hour recall (24h-R). The 54-item QFFQ have pre-determinate portions (small, middle, large and extra large). The difference between the estimated values from the two methods and from the two replications of QFFQ was assessed using correlation coefficients, evaluation of categorization agreement among methods and among replications and Bland & Altman plots. The calibration coefficient was calculated by regression of the 24h-R data after deattenuation (dependent variable) on the QFFQ data (independent variable) in linear regression models, with data in natural logarithm, adjusted by energy. Age, body mass index, physical activity, income, education and bias on report from 24h-R were included to adjust the multiple models. Results. Carbohydrate and fiber were well estimated by the QFFQ, with 0.42 and 0.46 in correlation coefficients, respectively, to deattenuated data. These nutrients were estimated by the QFFQ regardless of the magnitude of the intake. The consumption of fats, assessed by the QFFQ, had the worst correlation coefficients and smallest percentage agreement between the questionnaire and the 24h-R. The reproducibility analysis demonstrated that the QFFQ have acceptable precision for most of the nutrients investigated. The calibration coefficients ranged between 0.03 (total fat) and 0.45 (fiber). The power models (R2a) ranged from 0.01 (monounsaturated and polyunsaturated fat) to 0.81 (energy). Conclusion. This study demonstrated that the QFFQ is acceptable to assess nutrient intake from participants of HIM Brazil study. Minimizing bias of measurement is expected correcting the data by calibration. However, to examine the relationship between dietary factors and history of HPV infection in men, data corrected and not corrected by the calibration must be used comparatively (AU) | |