Original ArticleValidity of a food frequency questionnaire varied by age and body mass index
Introduction
The food frequency questionnaire (FFQ) has been established as the primary method for estimating long-term food consumption in large epidemiologic studies. Structured dietary data is easy and inexpensive to process compared to data gathered with an open method. Furthermore, no interviewers are needed if the study subjects fill in the questionnaires themselves. However, the FFQ has its limitations. The information is based on the subjects' memory, and because of the predefined food list, some information about the foods actually eaten may be missed. Therefore, it is essential to examine the validity of the FFQs against another dietary assessment method [1], [2], [3], [4], [5], [6], [7], [8], [9], [10].
It has been recommended that in study populations including both genders, validity has to be examined separately for men and women [11]. In studies in which the correlations between the FFQ and food records (the reference method) were analyzed separately by gender, obvious differences were not observed [6], [8], [12]. However, in a Danish validation study, contrary to the researchers' expectations, the correlations were mostly higher in men than in women [13]. In the Danish study the FFQ was slightly shorter than in the three other studies (92 vs. 116–350 foods).
There is evidence that gender, age, and obesity are associated with underreporting energy intake [14], [15], [16], [17]. In a Finnish study among subjects aged 25–64 years, underreporting was most common among women and overweight [body mass index (BMI) >25 kg/m2] subjects and among subjects older than 45 years [14]. Thus far, only a few validation studies for FFQs that measure the intake of a large variety of foods or nutrients with a food record as the reference method and including more than 100 subjects have shown results for subgroups other than by gender. Some validation studies have included only young [18] or elderly [19], [20] subjects or subjects from a specific profession, for example, nurses [1] or tin miners [21]. A validation study that was part of a study on selenium intake in relation to health status in the United States found that the correlations did not notably differ between the subjects with a high school education or less and the subjects with more than a high school education [6]. The study included 138 subjects with a mean age of 49 years. In another study in the United States, the correlations were reported separately for gender and race [22]. The correlations were higher for White subjects, and the correlations of Black men tended to be the lowest. The reference method was a combined 24-hr recall and food record and included food consumption data for 16 days (4 × 24-hr recall and 4 × 3-day food record).
The objective of our study was to assess the validity of a semiquantitative FFQ among an adult Finnish population with a 3-day food record as the reference method. To deepen our knowledge on the possible effect of the characteristics of the subjects on reporting food consumption, we also examined whether there are differences in the validity by age or BMI. The FFQ was developed for the Health 2000 Survey in Finland.
Section snippets
Study design and subjects
This validation study is a part of a nationally representative survey, the Health 2000 Survey (n = 8,028), which was carried out at 160 study locations in Finland from fall 2000 to spring 2001 [23]. The survey consisted of various health interviews, self-administered questionnaires, and a comprehensive health examination. The main aim of the survey was to estimate health and functional capacity and to gather information on major diseases, their causes, and treatment circumstances. Some basic
Results
In general, the mean nutrient intakes were higher measured with the FFQ than with the food records (Table 1). The only nutrient that was underestimated with the FFQ was alcohol. Overestimation was more prominent in women than in men. In women the intake of 17 of the 21 nutrients included in the validation study was over 40% higher measured with the FFQ than with the food records. The most overestimated nutrients were polyunsaturated fatty acids, long-chain n−3 fatty acids (LC n−3 FA),
Discussion
This validation study was a part of the Health 2000 Survey. The FFQ was specifically designed for the survey to assess the subjects' diet during the previous year, and to classify the subjects according to food consumption and nutrient intake. The validation sample included both genders aged 30–79 years. Although the subjects had participated in a number of interviews and examinations in the main survey before the administration of the FFQs and food records, the response rate in the validation
Acknowledgments
The authors thank all those who contributed to the data collection. Special thanks are due to Tommi Korhonen, M.Sc., and Heli Tapanainen, M.Sc., for assistance with the data management.
References (42)
- et al.
Reproducibility and validity of a food frequency questionnaire in a case–control study on breast cancer
J Clin Epidemiol
(1996) - et al.
Validation of a self-administered diet history questionnaire using multiple diet records
J Clin Epidemiol
(1990) - et al.
Precision and estimated accuracy of two short-term food frequency questionnaires compared with recalls and records
J Clin Epidemiol
(1996) - et al.
Underreporting of energy intake in an elderly German population
Nutrition
(2001) - et al.
Behavioral and body size correlates of energy intake underreporting by obese and normal-weight women
J Am Diet Assoc
(1999) - et al.
Validation of a youth/adolescent food frequency questionnaire
Prev Med
(1997) - et al.
A semiquantitative food frequency questionnaire for use in epidemiologic research among the elderly: validation by comparison with dietary history
J Clin Epidemiol
(1995) - et al.
Comparison of estimated nutrient intakes by food frequency and dietary records in adults
J Am Diet Assoc
(1989) - et al.
Reproducibility and validity of food intake measurements from a semiquantitative food frequency questionnaire
J Am Diet Assoc
(1993) - et al.
What are people really eating? The relation between energy intake derived from estimated diet records and intake determined to maintain body weight
Am J Clin Nutr
(1991)
A qualitative study of participants' experiences with dietary assessment
J Am Diet Assoc
Undereating and underrecording of habitual food intake in obese men: selective underreporting of fat intake
Am J Clin Nutr
Food consumption, habitual physical activity, and body fatness in young Dutch adults
Am J Clin Nutr
Human dietary assessment: methods and issues
Prev Med
The apparent validity of diet questionnaires is influenced by number of diet-record days used for comparison
J Am Diet Assoc
Reproducibility and validity of a semiquantitative food frequency questionnaire
Am J Epidemiol
Reproducibility and validity of dietary assessment instruments. I. A self-administered food use questionnaire with a portion size picture booklet
Am J Epidemiol
Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals
Am J Epidemiol
Validity and reproducibility of a food frequency questionnaire for pregnant Finnish women
Am J Epidemiol
The reproducibility and validity of a self-administered semiquantitative food frequency questionnaire in subjects from South Dakota and Wyoming
Epidemiology
Validation of a dietary questionnaire used in a large-scale prospective cohort study on diet and cancer
Eur J Clin Nutr
Cited by (151)
Change in physical activity, food choices and hemoglobin A1c among American Indians and Alaska Natives with type 2 diabetes
2022, Preventive Medicine ReportsCitation Excerpt :Finally, physical activity and food choices were based on self-reported measures. Under-reporting of total food intake, over-reporting of fruit and vegetable intake (Paalanen et al., 2006), and over-reporting of physical activity is common when using self-report methods (Sallis and Saelens, 2000). Additionally, Food Frequency Questionnaires are not exceptionally sensitive to detecting dietary changes (Cade et al., 2002).
Does better education mitigate risky health behavior? A mendelian randomization study
2022, Economics and Human BiologyPopulation group differences in subjective importance of meat in diet and red and processed meat consumption
2022, AppetiteCitation Excerpt :Direct association between wealth and SIM in men with high RPM consumption may derive from traditional masculine values associated with meat (Kildal & Syse, 2017). Strengths of this study include a large nationally representative sample, population-based health examination study data with standardized measurement methods and a wide range of information available on the subjects (Borodulin & Sääksjärvi, 2019), and RPM consumption measured with a validated and updated FFQ (Kaartinen et al., 2012; Männistö et al., 1996; Paalanen et al., 2006). Some methodological considerations, however, should be taken into account.
Physical inactivity from youth to adulthood and adult cardiometabolic risk profile
2021, Preventive MedicineAssociation between legume consumption and the intake of other foods and nutrients in the Finnish adult population
2023, British Journal of Nutrition