Abstract

Introduction

Epidemiological studies provide evidence able to identify risk factors and the importance of the diet in the prevention of non-transmissible chronic diseases (NTCDs)¹1. Barreto ML, Carmo EH. Tendências recentes das doenças crônicas no Brasil. In: Lessa I. O adulto brasileiro e as doenças crônicas da modernidade: epidemiologia das doenças crônicas não-transmissíveis. São Paulo: Hucitec; 1998. p. 15-27.

2. Tomita LY, Cardoso MA. Avaliação da lista de alimentos e porções alimentares de Questionário Quantitativo de Frequência Alimentar em população adulta. Cad Saúde Pública 2002; 18:1747-56. ^{- 3}3. Geraldo JM, Alfenas RCG. Papel da dieta na prevenção e no controle da inflamação crônica - evidências atuais. Arq Bras Endocrinol Metabol 2008; 52(6): 951-67..

The quantification and assessment of the diet are difficult tasks, although vital to the relationship between nutrients and NTCDs and in the monitoring of individual and populational dietetic behavior⁴4. Carroll RJ, Pee D, Freedman LS, Brown CC. Statistical design of calibration studies. Am J Clin Nutr 1997; 65(4 Suppl): 1187S-89S. ^{, 5}5. Willet WC. Nutritional Epidemiology. New York: Oxford University Press; 1998.. Those studies belong to the nutritional epidemiological area⁶6. Sempos CT, Liu K, Ernst ND. Food and nutrient exposures: what to consider when evaluating epidemiologic evidence. Am J Clin Nutr 1999; 69(6): 1330S-38S..

Several types of instruments are used to assess both the present and previous diet: the 24 hour dietary recall (R24h), food-intake record (FR), and the food frequency questionnaire (FFQ). All of those instruments show either advantageous or limiting results⁷7. Slater B, Lima FEL. Validade e reprodutibilidade dos métodos de inquérito alimentar. In: Fisberg RM, Slater B, Marchioni DML, Martini LA. Inquéritos alimentares: métodos e bases científicos. São Paulo: Manole; 2005. p.1-31..

Presently, the FFQ is the most useful instrument in epidemiological studies, intended to assess the relation between diet and disease, since it enables the access to information about the food intake level before the disease and also, from the scientific point of view, a much less expensive method⁵5. Willet WC. Nutritional Epidemiology. New York: Oxford University Press; 1998.. The instrument is structured in awayit can consider the dietary intake recording frequency in time-units and it might be subject to two types of errors: random and systematic ones⁸8. Slater B, Philippi ST, Marchioni DML, Fisberg RM. Validação de questionário de Frequência Alimentar - QFA: considerações metodológicas. Rev Bras Epidemiol 2003; 6: 200-8.. Systematic errors occur when there is a difference between the measured value and the real one, and thus, interfering in the instrument validity⁹9. Matarazzo HCZ, Marchioni DML, Figueiredo RAO, Slater B, Neto JE, Filho VW. Reprodutibilidade e validade do questionário de frequência de consumo alimentar utilizado em estudo caso-controle de câncer oral. Rev Bras Epidemiol 2006; 9(3): 316-24.. In order to minimize such results, more investigative studies on instrument validation are necessary.

A validation study should be conducted in order to improve the FFQ application within a specific population¹⁰10. Salvo VLMA, Gimeno SG. Reprodutibilidade e validade do questionário de frequência de consumo de alimentos. Rev Saude Publica 2002; 36(4): 505-12., and to avoid missinterpretations, improve accuracy and reduce bias levels related to diet-disease occurrences⁷7. Slater B, Lima FEL. Validade e reprodutibilidade dos métodos de inquérito alimentar. In: Fisberg RM, Slater B, Marchioni DML, Martini LA. Inquéritos alimentares: métodos e bases científicos. São Paulo: Manole; 2005. p.1-31. ^{, 11}11. Kaaks R, Riboli E. Validation and calibration of dietary intake measurements in the EPIC project: methodological considerations. European Prospective Investigation into Cancer and Nutrition. Int J Epidemiol 1997; 26 Suppl 1: S15-25. ^{, 12}12. Cade J, Thompson R, Burley V, Warm D. Development, validation and utilisation of food-frequency questionnaires - a review. Public Health Nutr 2002; 5(4): 567-87.. As a result, the instrument's characteristics will be better acknowledged and, whenever necessary, its calibration should be suggested.

The assessment of dietary intake by college students of the health area is justified, considering the important role they will play in relation to people's education in health practices, health promotion and disease prevention. This way, reliable instruments are needed in order to correctly measure the dietary pattern of the university's population. Since no golden-standard is known to exist, related to a dietary intake assessment, this study was intended to verify the relative validity of a quantitative food frequency questionnaire, especially built for college students of the healthcare area.

Methods

Quantitative Food Frequency Questionnaire construction

The study included 104 volunteers from the healthcare area, attending to the Students' Nutrition Ambulatory of the Universidade Federal de São Paulo (UNIFESP). This self-administered quantitative FFQ included 77 items at its initial stage. The food list was based in informed food and preparations for a 3-days dietary report (3dDR)¹³13. Carvalho FS, Van Laer NM, Sachs A, Salvo VLMA, Coelho LC, Santos GMS, et al. Desenvolvimento e pré-teste de um questionário de frequência alimentar para graduandos. Rev Nutr 2010; 23: 847-57.. In addition to that, information about home measurements were changed to grams and mL, with the help of centesimal composition charts¹⁴14. Pinheiro ABV, Lacerda EMA, Benzecry EH, Gomes MCS, Costa VM. Tabela para avaliação de consumo alimentar em medidas caseiras. Rio de Janeiro: Atheneu; 2007.

15. Fisberg RM, Villar BS. Manual de receitas e medidas caseiras para cálculos de inquéritos alimentares. São Paulo: Sigmus; 2002. ^{- 16}16. NEPA-UNICAMP. Tabela brasileira de composição de alimentos (TACO). Campinas: NEPA-UNICAMP; 2004., and calculated by the Nutritional Support software of UNIFESP¹⁷17. Anção MS, Cuppari L, Draibe SA, Sigulem D. Programa de apoio à Nutrição. Versão 1.15. São Paulo: Departamento de Informática em Saúde - SPDM UNIFESP/EPM; 2002.. The charts were then organized according to the percentage contribution for the total energy value informed, and selected and assembled according to their nutritional similarity and also according to the food that responds to the 95% of the energy intake. Based on the distributive percentage related to each weighted portion in grams, it was possible to define as a small portion (S) the one with value equal or lower than 25%; the medium portion (M) at 50% and the large one (L) at 75% or more. Following the development of a pilot project, the instrument was restructured as to include 70 food items. In addition to the food list, the FFQ presents a blank to be filled with the type of food absent from the initial list but pertaining to the dietary habits of the voluntary participant. Thus, their consumption frequency, of at least once a week, was considered valid⁵5. Willet WC. Nutritional Epidemiology. New York: Oxford University Press; 1998..

Study validity

The instrument used to assess the amount of food items consumed was tested in university students of both genders. The inclusion criteria were the absence of any interfering condition related to the usual food intake (such as pregnancy, lactation etc.), no previous nutritional orientation and the student being enrolled in the health care area of UNIFESP.

Convenient samples were attained by invitation, in the classrooms, to participate in the project. All students interested in the project were directed to the Students' Ambulatory of Health Service, where their personal informations were collected. At first, 108 students were recruited, but at the end only 50 of them - from the several health area Graduation courses - comprised the total sample. Participants were asked to sign an Informed Consent form and the investigation was approved by the Ethics and Research Committee from UNIFESP (CEP 1152/09).

The validation study sought to compare results attained by the study-test-method and results obtained by a reference method, considered of superior quality. The 3dDR instrument was used in order to avoid errors co-related to FFQ⁵5. Willet WC. Nutritional Epidemiology. New York: Oxford University Press; 1998. ^{, 7}7. Slater B, Lima FEL. Validade e reprodutibilidade dos métodos de inquérito alimentar. In: Fisberg RM, Slater B, Marchioni DML, Martini LA. Inquéritos alimentares: métodos e bases científicos. São Paulo: Manole; 2005. p.1-31..

The FFQ was filled before the 3dDR in order to estimate the usual diet as suggested by Cade et al. ¹²12. Cade J, Thompson R, Burley V, Warm D. Development, validation and utilisation of food-frequency questionnaires - a review. Public Health Nutr 2002; 5(4): 567-87.. Data collection was performed by trained nutritionists - dietitians attending the Public Health Specialization in Nutrition course at UNIFESP.

Literature suggests that the applied methods should be in consonance with the period of time in which they are to be conducted¹²12. Cade J, Thompson R, Burley V, Warm D. Development, validation and utilisation of food-frequency questionnaires - a review. Public Health Nutr 2002; 5(4): 567-87.. Thus, FFQ data collection was approved for a period of six months. This same period was considered for the reference method (3dDR) data collection. Food items in the 3dDR home measurements were converted to grams /day using the Nutritional Support software of UNIFESP¹⁷17. Anção MS, Cuppari L, Draibe SA, Sigulem D. Programa de apoio à Nutrição. Versão 1.15. São Paulo: Departamento de Informática em Saúde - SPDM UNIFESP/EPM; 2002. and the composition tables for food items¹⁴14. Pinheiro ABV, Lacerda EMA, Benzecry EH, Gomes MCS, Costa VM. Tabela para avaliação de consumo alimentar em medidas caseiras. Rio de Janeiro: Atheneu; 2007.

15. Fisberg RM, Villar BS. Manual de receitas e medidas caseiras para cálculos de inquéritos alimentares. São Paulo: Sigmus; 2002. ^{- 16}16. NEPA-UNICAMP. Tabela brasileira de composição de alimentos (TACO). Campinas: NEPA-UNICAMP; 2004.. As for the FFQ, it was conducted in a previous phase, dealing with the changes occurred in the food intake in grams /day using an Excel(r)¹⁸18. Microsoft Corporation. Microsoft Office Excel (r).[computer program.] Estados Unidos; 2003. spreadsheet, in which frequency was multiplied by the size of the informed portion, while the product was divided by the number of month-days. The following dietetic variables were considered: total energy (kcal), macronutrients (carbohydrates, lipids, and proteins - in grams) and total dietary fiber (in grams).

Statistical analysis

Data were analyzed by the SPSS software, version 15.0, for Windows¹⁹19. SPSS Corporation. SPSS for Windows(r) .[computer program.]. Version 15.0. Chicago; 2006., considering a significance level of 5%. In a previous analysis on the validity of FFQ, the Student's t-test was done in pairs, intending to verify the existence of statistically significant differences. In addition the residual method was performed for energy adjustment previous to intra-class correlation coefficient (ICC) estimate as proposed by Willett and Stampfer²⁰20. Willett WC, Stampfer MJ. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol 1986; 124(1): 17-27.. Relative validity was assessed using ICC for raw nutrients and adjusted by energy²¹21. Fleiss JL. The design and analysis of clinical experiments. New York: John Wiley & Son; 1986..

The Kappa index was used to assemble dietetic variables related to consumption´s thirds of each method. In this case, the assessment was carried out with the total proportion of participants correctly classified in the same consumption category²²22. Marchioni DM, Voci SM, Lima FE, Fisberg RM, Slater B. Reproducibility of a food frequency questionnaire for adolescents. Cad Saúde Pública 2007; 23(9): 2187-96.. Information divergences were examined in a graphic developed by Bland and Altman²³23. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999; 8(2): 135-60., allowing trend verification on sub- or- over data estimation, presented by the test-method.

Results

Most participants were female (74%). The sample consisted of students of medicine (44%), nursing students (22%), nutrition students (18%) and others (16%). From the initial sample, which included 108 participants, 53.7% of them were presented with incomplete documents or desistance and so, only 50 participants were approved for the validation study. Table 1 shows data related to energy and nutrients obtained by both dietary inquiries. As for the percentage of macronutrient contributions, increased values were shown by the FFQ for carbohydrates and proteins when compared to those of 3dDR.

In relation to kappa statistics values when adjusted to energy, little concordance was observed of macronutrients when compared to FFQ and 3dDR data are shown in Table 2. Better results were seen for energy (ICC 0.32; 95%CI 0.1 - 0.55). Table 3 shows the resulting FFQ capacity in classifying individuals within the same thirds, adjacent thirds or opposite thirds, for each consumption variable, after the adjustment by energy.

It was observed that individuals classified in the opposite thirds are represented by the minority (16%), as approximately half of the participants were classified in the same thirds. However, results do not point to a valid agreement since the attained values for kappa were low, mainly for macronutrients, and varying from 0.05 for proteins to 0.14 for carbohydrates and lipids.

It was observed a tendency to overestimation in the FFQ value for energy and nutrients, except for lipids (Table 4). Bland and Altman's²³23. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999; 8(2): 135-60. proposed graphics were built in order to verify the divergences between both methods. Thus analysis of the difference between the nutrients and energy found the same results (Figure 1A and B). Figure 1A shows that higher deviation levels between data on energy obtained by FFQ and 3dDR were seen in intakes of more than 2000 kcal. Figure 1B shows trends of overestimation in the FFQ report, in relation to 3dDR for proteins as well as for other nutrients, whose results were similar.

Discussion

Assessing food intake is not an easy task, due not only by the absence of a golden standard but also by the fact that the existing methods are subject to variations and measurement errors²⁴24. Lopes ACS, Caiaffa WT, Mingoti SA, Lima-Costa, MFF. Ingestão alimentar em estudos epidemiológicos. Rev Bras Epidemiol 2003; 6(3): 209-19..

Both 3dDR and FFQ showed tendency to protein rich diets (Table 1), according to the WHO recommendations²⁵25. World Health Organization. Diet, nutrition and the prevention of chronic diseases. Report of a Joint WHO/ FAO Expert Consultation. Geneva; 2003., which are intended to prevent NTCDs. The contributing percentage of proteins in relation to the total energy value was over 15%. As a consequence, an alteration occurs in the percentage of other participating macronutrients. The low intake of fibers (< 25 g/d)²⁶26. Brasil. Ministério da Saúde. Secretaria de Atenção à Saúde. Coordenação-Geral da Política de Alimentação e Nutrição. Guia alimentar para a população brasileira: promovendo a alimentação saudável. Brasília: Ministério da Saúde; 2005. 236p. - (Série A. Normas e Manuais Técnicos). reinforces the need for nutritional attention and surveillance toward this populational group. The mean levels for energy intake, fibers and macronutrients except lipids, were higher for FFQ in relation to those in 3dDR. Since the comparison between the mean of energy and nutrients between both methods provide limited information concerning the validity⁵5. Willet WC. Nutritional Epidemiology. New York: Oxford University Press; 1998. , other statistical tests were carried out.

The ICC and the kappa index values were low and even furtherly reduced when energy was corrected (Table 2). This adjustment of nutrients using energy is seen as a methodological strategy developed to correctly try to estimate parameters and try to control any eventual confusion effect caused by energy in validation⁸8. Slater B, Philippi ST, Marchioni DML, Fisberg RM. Validação de questionário de Frequência Alimentar - QFA: considerações metodológicas. Rev Bras Epidemiol 2003; 6: 200-8. ^{, 10}10. Salvo VLMA, Gimeno SG. Reprodutibilidade e validade do questionário de frequência de consumo de alimentos. Rev Saude Publica 2002; 36(4): 505-12. studies. The Reduction of coefficient values after the adjustment by energy is useful to other studies²⁷27. Cardoso MA, Kida AA, Tomita LY, Stocco PR. Reproducibility and validity of a food frequency questionnaire among women of Japanese ancestry living in Brazil. Nutr Res 2001; 21: 725-33.

28. Ribeiro AC, Sávio KEO, Rodrigues MLCF, Costa THM, Schmitz BAS. Validação de um questionário de frequência de consumo alimentar para população adulta. Rev Nutr 2006; 19(5): 553-62.

29. Lima FEL, Slater B, Latorre MRDO, Fisberg RM. Validade de um questionário quantitativo de frequência alimentar desenvolvido para população feminina no nordeste do Brasil. Rev Bras Epidemiol 2007; 10(4): 482-90.

30. Galante AP, Colli C. Desenvolvimento e aplicação de um questionário semiquantitativo de frequência alimentar on-line para estimar a ingestão de cálcio e ferro. Rev Bras Epidemiol 2008; 11: 402-10. ^{- 31}31. Zanolla AF, Olinto MTA, Henn RL, Wahrlich V, Anjos LA. dos. Avaliação de reprodutibilidade e validade de um questionário de frequência alimentar em adultos residentes em Porto Alegre, Rio Grande do Sul, Brasil. Cad Saúde Pública 2009; 25: 840-48., which may evidence the presence of systematic errors on sub-or-super account of food consumption⁵5. Willet WC. Nutritional Epidemiology. New York: Oxford University Press; 1998. ^,31.

A large number of validation studies used the Pearson's or the Spearman's correlation coefficient when comparing the methods⁹9. Matarazzo HCZ, Marchioni DML, Figueiredo RAO, Slater B, Neto JE, Filho VW. Reprodutibilidade e validade do questionário de frequência de consumo alimentar utilizado em estudo caso-controle de câncer oral. Rev Bras Epidemiol 2006; 9(3): 316-24. ^{, 27}27. Cardoso MA, Kida AA, Tomita LY, Stocco PR. Reproducibility and validity of a food frequency questionnaire among women of Japanese ancestry living in Brazil. Nutr Res 2001; 21: 725-33.

28. Ribeiro AC, Sávio KEO, Rodrigues MLCF, Costa THM, Schmitz BAS. Validação de um questionário de frequência de consumo alimentar para população adulta. Rev Nutr 2006; 19(5): 553-62.

29. Lima FEL, Slater B, Latorre MRDO, Fisberg RM. Validade de um questionário quantitativo de frequência alimentar desenvolvido para população feminina no nordeste do Brasil. Rev Bras Epidemiol 2007; 10(4): 482-90. ^{- 30}30. Galante AP, Colli C. Desenvolvimento e aplicação de um questionário semiquantitativo de frequência alimentar on-line para estimar a ingestão de cálcio e ferro. Rev Bras Epidemiol 2008; 11: 402-10. ^{, 32}32. Sichieri RE, Everhart J. Validity of a Brazilian food frequency questionnaire against dietary recalls and estimated energy intake. Nutr Res 1998; 18: 1649-59.

33. Chan SG, Ho SC, Kreiger N, Darlington G, Adlaf EM, So KF, et al. Validation of a food frequency questionnaire for assessing dietary soy isoflavone intake among midlife Chinese women in Hong Kong. J Nutr 2008; 138(3): 567-73.

32. Field AE, Peterson KE, Gortmaker SL, Cheung L, Rockett H, Fox MK, et al. Reproducibility and validity of a food frequency questionnaire among fourth to seventh grade inner-city school children: implications of age and day-to-day variation in dietary intake. Public Health Nutr 1999; 2(3): 293-300.

33. Johansson I, Hallmans G, Wikman A, Biessy C, Riboli E, Kaaks R. Validation and calibration of a food-frequency questionnaire measurements in the Northern Sweden Health and Disease cohort. Public Health Nutr 2002; 5(3): 487-96.

34. Imaeda N, Goto C, Tokudome Y, Hirose K, Tajima K, Tokudome S. Reproducibility of a short food frequency questionnaire for Japanese general population. J Epidemiol 2007; 17(3):100-7.

35. Jackson M, Walker S, Cade J, Forrester T, Cruickshank J.K, Wilks R. Reproducibility and validity of a quantitative food-frequency questionnaire among Jamaicans of African origin. Public Health Nutr 2001; 4(5): 971-80.

36. Matthys C, Pynaert I, De Keyzer W, DeHenauw S. Validity and reproducibility of an adolescent web-based food frequency questionnaire. J Am Diet Assoc 2007; 107(4): 605-10.

37. Messerer M, Johansson SE, Wolk A. The validity of questionnaire-based micronutrient intake estimates is increased by including dietary supplement use in Swedish men. J Nutr 2004; 134(7): 1800-5. ^{- 40}38. Parr CL, Veierød MB, Laake P, Lund E, Hjartåker A. Test-retest reproducibility of a food frequency questionnaire (FFQ) and estimated effects on disease risk in the Norwegian Women and Cancer Study. Nutr J 2006; 5: 4. which turned results more difficult to be evaluated. In spite of the coefficient's attained values, the use of ICC was adopted for viewing the several publications which point to over-estimates association when using Pearson's correlation in samples with great inter-subject variability, which as a consequence, could lead to false conclusions concerning the method's validity²³23. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999; 8(2): 135-60. ^{, 41}39. Ribeiro LG, Cardoso LO. Estudos de validação: Qual análise utilizar? Rev Bras Med Esporte 2009; 15: 316.. High values for Pearson's correlation coefficient do not necessarily mean that both methods are consistent, since this measurement is only directed to the strength relation between two variables²³23. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999; 8(2): 135-60..

Related literature shows that correlations between the FFQ diet estimations and the method of reference are found between 0.3 and 0.7⁴²40. Slater B, Marchioni DM, Voci SM. Use of linear regression for correction of dietary data. Rev Saude Publica 2007; 41(2): 190-6., although other studies had been published about similar methodology validation with lower coefficients, than those seen here, for some nutrients¹⁶16. NEPA-UNICAMP. Tabela brasileira de composição de alimentos (TACO). Campinas: NEPA-UNICAMP; 2004. ^{, 31}31. Zanolla AF, Olinto MTA, Henn RL, Wahrlich V, Anjos LA. dos. Avaliação de reprodutibilidade e validade de um questionário de frequência alimentar em adultos residentes em Porto Alegre, Rio Grande do Sul, Brasil. Cad Saúde Pública 2009; 25: 840-48.. Those studies showed a satisfactory reproducibility level and reasonable validity in relation to the studied nutrients.

The FFQ validity evaluation was also tested by comparing the individuals' classification within the consumption thirds for both methods. Results showed to be consistent with other studies in literature, regarding studies with similar methodology⁸8. Slater B, Philippi ST, Marchioni DML, Fisberg RM. Validação de questionário de Frequência Alimentar - QFA: considerações metodológicas. Rev Bras Epidemiol 2003; 6: 200-8. ^{, 9}9. Matarazzo HCZ, Marchioni DML, Figueiredo RAO, Slater B, Neto JE, Filho VW. Reprodutibilidade e validade do questionário de frequência de consumo alimentar utilizado em estudo caso-controle de câncer oral. Rev Bras Epidemiol 2006; 9(3): 316-24. ^{, 43}41. Lee MS, Pan WH, Liu KL, Yu MS. Reproducibility and validity of a Chinese food frequency questionnaire used in Taiwan. Asia Pac J Clin Nutr 2006; 15(2): 161-9. ^{, 44}42. Bilau M, Matthys C, Bellemans M, De Neve M, Willems JL, De Henauw S. Reproducibility and relative validity of a semi-quantitative food frequency questionnaire designed for assessing the intake of dioxin-like contaminants. Environ Res. 2008; 108(3): 327-33., a higher percentage of individuals classified in the same category (43.2%) and a few individuals in opposite thirds (16%). The questionnaire's ability in classifying individuals, according to their intake level, contributed to attain a correct risk estimate³¹31. Zanolla AF, Olinto MTA, Henn RL, Wahrlich V, Anjos LA. dos. Avaliação de reprodutibilidade e validade de um questionário de frequência alimentar em adultos residentes em Porto Alegre, Rio Grande do Sul, Brasil. Cad Saúde Pública 2009; 25: 840-48..

Mc Pherson et al.⁴⁵43. McPherson RS, Hoelscher DM, Alexander M, Scanlon KS, Serdula MK. Dietary assessment methods among school-aged children: validity and reliability. Prev Med 2000; 31: S11-S33. stated that some self-administered dietary inquiries present overestimated results at the first contact with the instrument in relation to the second. In the present study, it could be observed that FFQ overestimated the dietary measurements, except when dealing with lipids. As for proteins and carbohyrates, a higher divergence level between the methods was observed. Other authors³¹31. Zanolla AF, Olinto MTA, Henn RL, Wahrlich V, Anjos LA. dos. Avaliação de reprodutibilidade e validade de um questionário de frequência alimentar em adultos residentes em Porto Alegre, Rio Grande do Sul, Brasil. Cad Saúde Pública 2009; 25: 840-48. ^{, 43}41. Lee MS, Pan WH, Liu KL, Yu MS. Reproducibility and validity of a Chinese food frequency questionnaire used in Taiwan. Asia Pac J Clin Nutr 2006; 15(2): 161-9. found more coherent data during a second application of the FFQ and 3dDR methods and suggested a better participant preparation when answering on the changes of their dietary intake, during the period of the study. Krebs-Smith¹⁰10. Salvo VLMA, Gimeno SG. Reprodutibilidade e validade do questionário de frequência de consumo de alimentos. Rev Saude Publica 2002; 36(4): 505-12. emphasized an over-estimation trend related to dietary intake, which might be the result of a long food list, leading then to the responsive individuals' weariness. The food list was carefully developed and reformulated after the execution and inclusion of 70 food items in the pilot project.

Viewing to improve the results interpretation, the methodology proposed by Bland-Altman²³23. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999; 8(2): 135-60. was used for the graphic construction, based in the consumption data, so then it was possible to observe that in lower energy intake cases, the agreement between methods turns to be more useful. Approximately over 200 kcal data shows to be more distant from the mean levels, even if ICC and higher kappa (0.43 and 0.42, respectively) had been found when compared to macronutrients and fibers.

Among this study's limitations, it is important to mention the higher number of female participants and the losses. Other Brazilian studies that considered the university population have also included larger number of female participants in relation to the number of males⁴⁶44. Fujimori E, Oliveira IMV, Soares MAL, Osso AN. Avaliação nutricional de estudantes universitários do sexo feminino. Rev Esc Enferm USP 1994; 28: 72-82. ^{, 18}18. Microsoft Corporation. Microsoft Office Excel (r).[computer program.] Estados Unidos; 2003., although this is a characteristic of the population of students⁴⁶44. Fujimori E, Oliveira IMV, Soares MAL, Osso AN. Avaliação nutricional de estudantes universitários do sexo feminino. Rev Esc Enferm USP 1994; 28: 72-82.

45. Santos JS, Xavier AAO, Ries EF, Costabeber I, Emanuelli T. Consumo de produtos de origem animal por universitários em Santa Maria - RS. Nutrire: rev Soc Bras Alim Nutr= J Brazilian Soc Food Nutr 2006; 31(2): 13-24. ^{- 48}46. Marcondelli P, Costa THM, Schmitz BAS. Physical activity level and food intake habits of university students from 3 to 5 semester in the health area. Rev Nutr 2008; 21(1): 39-47. enrolled in the Graduation courses of UNIFESP (61.44% females, 2009). In relation to losses, it is possible that this high percentage might have influenced the results concerning the viewing the final sample small size. The reason for such losses, was the poor attendance for nutritional visits, absence of 3dDR, FFQ incorrect record filling, and/or lacking of the Informed Consent.

Literature suggests¹²12. Cade J, Thompson R, Burley V, Warm D. Development, validation and utilisation of food-frequency questionnaires - a review. Public Health Nutr 2002; 5(4): 567-87. that a minimum number (50) of participants may be adequate for a validation study¹²12. Cade J, Thompson R, Burley V, Warm D. Development, validation and utilisation of food-frequency questionnaires - a review. Public Health Nutr 2002; 5(4): 567-87.. Other studies on validation, conducted in Brazil, also presented losses to follow up, even though, with a lower percentage as seen in the present study (53.7%): Assis et al.⁴⁹47. De Assis MA, Kupek E, Guimarães D, Calvo MC, De Andrade DF, Bellisle F. Test-retest reliability and external validity of the previous day food questionnaire for 7-10-year-old school children. Appetite 2008; 51(1): 187-93. and Ribeiro et al.²⁸28. Ribeiro AC, Sávio KEO, Rodrigues MLCF, Costa THM, Schmitz BAS. Validação de um questionário de frequência de consumo alimentar para população adulta. Rev Nutr 2006; 19(5): 553-62. with 14% and Zanolla et al.³¹31. Zanolla AF, Olinto MTA, Henn RL, Wahrlich V, Anjos LA. dos. Avaliação de reprodutibilidade e validade de um questionário de frequência alimentar em adultos residentes em Porto Alegre, Rio Grande do Sul, Brasil. Cad Saúde Pública 2009; 25: 840-48. with 27%. A study by Cardoso et al.²⁷27. Cardoso MA, Kida AA, Tomita LY, Stocco PR. Reproducibility and validity of a food frequency questionnaire among women of Japanese ancestry living in Brazil. Nutr Res 2001; 21: 725-33. showed that from the initial sample of 255 nutritionists, the investigators remained with only 77 individuals for the reproducibility study. For validation, 52 participants were recruited. However, validation studies in Brazil may eventually include less than 50 participants⁹9. Matarazzo HCZ, Marchioni DML, Figueiredo RAO, Slater B, Neto JE, Filho VW. Reprodutibilidade e validade do questionário de frequência de consumo alimentar utilizado em estudo caso-controle de câncer oral. Rev Bras Epidemiol 2006; 9(3): 316-24. ^{, 22}22. Marchioni DM, Voci SM, Lima FE, Fisberg RM, Slater B. Reproducibility of a food frequency questionnaire for adolescents. Cad Saúde Pública 2007; 23(9): 2187-96. ^{, 29}29. Lima FEL, Slater B, Latorre MRDO, Fisberg RM. Validade de um questionário quantitativo de frequência alimentar desenvolvido para população feminina no nordeste do Brasil. Rev Bras Epidemiol 2007; 10(4): 482-90. ^{, 30}30. Galante AP, Colli C. Desenvolvimento e aplicação de um questionário semiquantitativo de frequência alimentar on-line para estimar a ingestão de cálcio e ferro. Rev Bras Epidemiol 2008; 11: 402-10.. In addition to that, it may be possible to emphasize a series of limitations inherent to any FFQ as those related to memory, perception, lacking standartization, structure faults, and also due to chance⁴²40. Slater B, Marchioni DM, Voci SM. Use of linear regression for correction of dietary data. Rev Saude Publica 2007; 41(2): 190-6.. Specifically, in relation to the tested FFQ, a possible biased record should be taken into consideration since this FFQ is self-administered, and subject to false interpretations, even with a previous pilot-study having been conducted and intended to minimize this effect. The low correlation coefficients levels - except those of calories and fibers - for the agreement methods, may occur mainly in the absence of a golden-pattern. Due to this finding, it was opted for the use of 3dDR, since this method independs on the memory and it may lead to less FFQ correlated errors⁵5. Willet WC. Nutritional Epidemiology. New York: Oxford University Press; 1998., and also to its possible application in the studied sample. However, the filling of 3dDR opens the space for possible changes in the present diet and as a consequence, not reflecting the real consumption. At the same time, the role played by the dietary intake variability within the course of time must be emphasized and also, one must not ignore the existence of a different scale in order to estimate the nutrient intake using the FFQ or 3dDR¹⁰10. Salvo VLMA, Gimeno SG. Reprodutibilidade e validade do questionário de frequência de consumo de alimentos. Rev Saude Publica 2002; 36(4): 505-12..

The knowledge of the problems linked to the method enables the development of new studies which seek its correction and calibration. This way, it is recommended that the use of this instrument in university population in studies intended to estimate energy as well as in the classification of individuals in consumption categories. The use of adequate instruments may serve as the basis to develop adequate instruments, able to serve as the basis for strategies, development of new strategies, disease prevention, and also, to improve the quality of life and health promotion.

References

Publication Dates

History