Filetype PDF | Posted on 14 Jan 2023 | 2 years ago
Authentication
digital resources
79x Filetype PDF File size 0.11 MB Source: pdfs.semanticscholar.org
Validity and reproducibility of a food frecuency questionnaire ARTÍCULO ORIGINAL
Validity and reproducibility of
a food frequency questionnaire to assess
dietary intake of women living in Mexico City*
(1) (2) (1)
Mauricio Hernández-Avila, MD, DSc., Isabelle Romieu, MD, DSc., Socorro Parra, MSc.,
(1) (3) (4,5)
Juan Hernández-Avila,MSc., Herlinda Madrigal, Lic., Walter Willett, MD, DPh.
Hernández-Avila M, Romieu I, Parra S, Hernández-Avila M, Romieu I, Parra S,
Hernández-Avila J, Madrigal H, Willett W. Hernández-Avila J, Madrigal H, Willett W.
Validity and reproducibility of Validación y reproducibilidad de un cuestionario de
a food frequency questionnaire to assess frecuencia de consumo para estimar la ingesta
dietary intake of women living in Mexico City. alimentaria en mujeres de la Ciudad de México.
Salud Publica Mex 1998;40:133-140. Salud Publica Mex 1998;40:133-140.
Abstract Resumen
Objective. To assess the reproducibility and validity of a Objetivo. Evaluar la reproducibilidad y validez de un cues-
116 item semi-quantitative food frequency questionnaire tionario semicuantitativo de frecuencia de consumo de ali-
(FFQ), designed to assess the relation between dietary mentos, para ser utilizado en estudios epidemiológicos sobre
intake and chronic diseases. Material and methods. To enfermedades crónicas. Material y métodos. Para evaluar
test the reproducibility of the FFQ questionnaire, the FFQ la reproducibilidad del cuestionario de frecuencia de con-
was administered twice to 134 women residing in Mexico sumo se comparó el resultado obtenido al aplicar la encuesta
City at an interval of approximately one year; to assess a 134 mujeres en dos tiempos separados por 12 meses.
the validity we compared results obtained by the FFQs Para evaluar la validez del cuestionario, comparamos los
with those obtained by four 4-day 24-hour recalls at three resultados obtenidos a partir de su aplicación, con el pro-
month intervals. Validity and reproducibility were evaluated medio de 16 recordatorios de 24-horas, los cuales se obtu-
using regression analysis and Pearson and intraclass vieron en el transcurso de 12 meses. Para estimar la validez
correlation coefficients of log-e and calorie-adjusted nutrient utilizamos modelos de regresión y correlación de Pearson.
scores. Results. Mean values for intake of most nutrients Para evaluar la reproducibilidad estimamos la correlacion
assessed by the two food frequency questionnaires were intraclase entre los cuestionarios de frecuencia de consumo.
similar. However, means for the 24-hr recall were significantly Resultados. Las ingestas promedio diarias estimadas por
lower. Intraclass correlation coefficients for nutrient intakes, los cuestionarios de frecuencia de consumo fueron similares.
assessed by questionnaires, administered one year apart, Sin embargo, las medias estimadas mediante los recorda-
ranged from 0.38 for cholesterol to 0.54 for crude fiber. torios de 24 horas fueron significativamente menores. Las
Correlation coefficients between energy-adjusted nutrient correlaciones intraclase entre los cuestionarios de frecuencia
intakes, measured by diet recalls, and the first FFQ ranged de consumo aplicados 12 meses aparte, variaron entre 0.38
from 0.12 for polyunsaturated fatty acids to 0.67 for para colesterol y 0.54 para fibra cruda. Las correlaciones
saturated fatty acids. Regression coefficients between ajustadas por calorías entre el primer cuestionario y la media
24-hr recall and FFQ,s were all significant were significant de los recordatorios variaron entre 0.12 para ácidos grasos
for all nutrients, except for polyunsaturated fat, folic acid, polinsaturados y 0.67 para ácidos grasos saturados. En el
vitamin E and Zinc. Conclusions. These data indicate that análisis de regresión observamos asociaciones estadística-
* This paper was presented in the II International Conference on Nutritional Survey Methods. Boston, Mass. January 22-24, 1995.
(1) Instituto Nacional de Salud Publica, Cuernavaca, Mexico
(2) Centro Panamericano de Ecologia Humana y Salud, Mexico
(3) Instituto Nacional de la Nurtricion Salvador Zubirán, Mexico
(4) Department of Epidemiology and Nutrition, Harvard School of Public Health, Boston MA, USA
(5) Channing Laboratory, Department of Medicine, Brigham and Women’s hospital and Harvard Medical School, Boston MA, USA
Fecha de recibido: 31 de octubre de 1997 • Fecha de aprobado: 7 de abril de 1998
Correspondence should be addressed to Dr. Mauricio Hernandez, Instituto Nacional de Salud Publica, 655 Av. Universidad, Col. Sta Maria Ahuacatitlan,
Cuernavaca, Morelos, Mexico. Tel :52-73-17-53-91; Fax: 52-73-11-11-48. Supported by a Research Grant of the National Institute of Health (Small
grants for epidemiology), by the Ministry of Health of Mexico and the National Center for Environment from the Centers for Disease Control and
Prevention, Atlanta, GA , USA.
salud pública de méxico / vol.39, no.40, enero-febrero de 1998 133
ARTÍCULO ORIGINAL Hernández-Avila M y col.
this semi-quantitative FFQ is reproducible and provides a mente significativas para casi todos los nutrientes, excepto
useful estimate by which to categorize individuals by level para grasas polinsaturadas, ácido folico, vitamina E y zinc.
of past nutrient intake. However, its application outside Conclusiones. Los resultados indican que el cuestionario
Mexico City or in different age and gender populations will es útil para evaluar la ingesta dietética. Sin embargo, su uti-
require additional modifications and validation efforts. lización en poblaciones con diferente estructura de edad y
género o ubicación geográfica requerirá de acciones adicio-
nales para garantizar su validez y reproducibilidad.
Key words: questionnaire/validity; diet; women; epidem- Palabras clave: cuestionario/validez; dieta; mujeres; epide-
iology; Mexico miología; México
Exposure assessment is a critical issue in epidemi- ulating the results of a dietary survey carried out in
ological studies. The value of any research find- 1983 by the National Institute of Nutrition on a ran-
ing depends largely on the validity of the information dom sample of 240 low to medium income families
3
used to measure exposure. For studies relating the in Mexico City. Data collected in this survey includ-
effect of food consumption and chronic disorders, food ed both 24-hour dietary recalls and home visits to
frequency questionnaires (FFQs) have been widely weigh and measure the food items actually consu-
used because they allow the evaluation of individual med. Using, stepwise linear regression, the foods
dietary intake for large populations at low cost. FFQs which were the best predictors of each nutrient of in-
are designed to measure dietary intake over an extend- terest were identified. We then compiled all the foods
ed period of time, such as months or years. The under- identified using this methodology. In a second step
lying assumption of the FFQ approach is that the we identified important foods that may have not been
measurement of long term dietary exposures is more reported by the population who participated in the sur-
relevant to chronic diseases. Therefore, in this context vey carried out by the National Institute of Nutrition.
it is better to rely on less precise estimates of usual di- For this purpose, a group of Mexican dietitians and
etary intake than on rather precise estimates of short- nutritionists were invited to participate and were asked
term intake, which may not adequately reflect to identify relevant food items that may have not been
1
exposure. Consequently, it is expected that the infor- identified with the stepwise linear regression approach.
mation derived from FFQs will not reflect current di- A final format that included 85 food items was then
etary intake as accurately, but instead will be more valid pilot-tested using a convenient sample of women of
to rank individuals by levels of their past nutrient medium to low socioeconomic status, living in Mexico
2
intake. FFQs are based on a list of foods and ques- City. The final instrument was composed of a ma-
tions regarding how often these foods are eaten. There- trix listing 116 food items and 10 frequencies of con-
fore, in order to provide a reasonable estimation of sumption. The consumption frequencies were: 6 or
intake, the food items included in the questionnaire more per day, 4-5 per day, 2-3 per day, 1 per day, 5-6
need to be selected carefully, using methods that would per week, 2-4 per week, 1 per week, 2-3 per month, 1
yield a list of foods that will reflect the food consump- per month or less and never. (Copy of the question-
tion patterns and choices of the population under naire is abailable from author M. Hernandez-Avila).
study.
We therefore conducted a study to assess the va- Estimation of nutrient scores
lidity and reproducibility of an FFQ to be used in fu-
ture epidemiological studies, to assess nutritional risk For each food item, the nutrient content per average
factors of chronic diseases among women who live in unit (specified serving size: slice, glass, or natural unit)
Mexico City. was compiled. Nutritional composition of each food
included in the questionnaire was derived from the US
Material and methods Department of Agriculture (USDA) food composition
4
tables and, when necessary, complemented by the
Development of the food frequency nutrient data base developed by the National Institute
questionnaire of Nutrition.5 Daily dietary nutrient intakes for study
participants were estimated as follows: for each food
The FFQ was developed using the methodology pro- a nutrient score was calculated using the food nutrient
2
posed by Willett et al. We identified food items by tab- content taken from the food table, adjusted to the spec-
134 salud pública de méxico / vol.39, no.40, marzo-abril de 1998
Validity and reproducibility of a food frecuency questionnaire ARTÍCULO ORIGINAL
ified portion size of the questionnaire. We then multi- Statistical analysis
plied this score by the weight corresponding to the fre-
quency of use. The weights used were: 6 for reported All items consumed, by frequency and quantity, in ei-
frequencies of 6 or more per day; 4.5 for 4-5 per day; ther the FFQ or the 24-hour recalls were coded and
2.5 for 2-3 per day; 1 for 1 per day; 0.8 for 5-5 per week; transformed to daily nutrients with the use of a food-
0.43 for 2-4 per week, 0.08 for 2-3 per month and 0.016 composition database, developed to accommodate
for 1 per month or less. Then the product of frequen- the characteristics of the Mexican diet. We used food
cies times the nutrient scores were summed, produc- composition tables provided by the National Institute
5
ing scores for overall food items, to obtain a total of Nutrition and completed these tables with nutri-
4
nutrient score. tion composition tables from the USDA and other
Portion sizes were estimated by “natural” units published sources and personal communications from
or standard quantities where possible. To account for laboratories and manufacturers.
the seasonal availability of some foods, the question- For the calculation of correlations, we used loga-
naire had a section on the frequency of consumption rithmic transformation for all nutrient values and blood
of these foods during the season (e. g. mangos, mam- levels to reduce skewedness. We calculated energy-
mey). adjusted nutrients as the residual of each nutrient re-
The data from the 16 24-hour dietary recalls were gressed on caloric intake.6 We calculated intraclass
coded in a specifically-made data entry program. We correlation coefficients to assess the reproducibility of
calculated the specific nutrient content of the reported the FFQ.7 To assess the validity of nutrient values,
portion size and estimated a daily nutrient intake for derived from the first and second FFQ, we used re-
each day. gression analysis and Pearson correlation coefficients
and compared them with the mean of the 16 24-hour
Study population recalls. To take into account within-person variation,
we used within- and between-person components of
We randomly selected an age-stratified sample of 527 variation for the 24-hour recalls to “deattenuate” the
women residing in Tlalpan, the southern district of Pearson correlation coefficients.8 All analyses were con-
Mexico City. Among them, 211 agreed to participate. ducted using STATA statistical software.9
A final sample of 134 women completed the study. The
main reason for non-participation was the inability to Results
comply with the complete study protocol. All study
participants signed an informed consent and complet- Mean daily intakes of nutrients estimated by the FFQ
ed the study. Over a one-year period, we obtained four and the average of 16 24-hour recalls for the 134 wom-
4-day 24-hour recalls, corresponding to 16 days of diet en included in the validation analysis, are presented
recall in batches of 4 days, representing all seasons in in Table I. The questionnaire measurements of total
Mexico. Interviewers visited each participating wom- energy intake were higher than the intake assessed by
en at her home and recorded the type and quantity of diet recalls. Intake of most nutrients was slightly high-
food (using local cooking utensils) they had consumed er when measured by the FFQs compared with diet
during the previous 24-h. We arranged visitation of recalls, particularly for retinol. However the absolute
each subject on different days of the week (including values of nutrient intake, estimated by the second FFQ,
Saturdays and Sundays) to account for differences in were quite comparable to the estimate using diet re-
daily dietary habits. At baseline, women completed an calls, with the questionnaire estimates showing a high-
FFQ applied by a trained dietitian and provided blood er standard deviation . Eighty one percent of the mean
samples three and nine months later. At the end of the values for macro and micronutrients, apart from vita-
study, the women completed a second FFQ. Results mins, measured by questionnaire were within 15 per-
from the blood sample analyses are described else- cent of the 24-hr recall values. For vitamins, 82% of
where.* mean intake values measured by questionnaire were
within 25 percent of the diet record values.
To obtain a measure of the existing within-subject
variability, we computed intraclass correlations for
unadjusted and energy-adjusted daily nutrients using
* Romieu Y, Parra S, Hernández-Avila J, Madrigal H, Willett W, the average intake, measured by each of the 4-day diet
Hernández-Avila M. Questionnaire assessment of antioxidants recalls. Alpha-carotene was the nutrient with the low-
retinol intakes in Mexican women. Unpublished paper. est correlation coefficient (intraclass r= 0.09) followed
salud pública de méxico / vol.39, no.40, enero-febrero de 1998 135
ARTÍCULO ORIGINAL Hernández-Avila M y col.
Table 1
MEANS AND STANDARD DEVIATIONS OF DAILY NUTRIENT INTAKE ESTIMATED BY SIXTEEN 24 HOUR RECALLS
QUESTIONNAIRES AND FROM TWO FOOD FREQUENCY QUESTIONNAIRES
Nutrient 24-hr recalls SD FFQ1 SD FFQ2 SD
Calorie (Kcal) 1802 393.2 2200 784.3 1960 646.7
Carbohydrates (g) 253.7 60.2 338.9 126.8 304.1 110.7
Protein (g) 60.8 12.9 76.4 24.8 69.3 21.6
Animal protein (g) 33.9 10.5 36.9 15.5 36.9 15.5
Total fat (g) 63.8 18.1 60.9 24.3 70.8 31.4
Animal fat (g) 33.9 10.5 36.9 15.5 32.0 12.58
Saturated fat (g) 18.2 5.3 20.8 9.1 17.2 6.8
Monosaturated fat (g) 19.8 6.00 19.8 8.13 16.5 5.76
Polyunsaturated fat (g) 16.0 6.2 7.5 2.5 8.3 3.0
Cholesterol (mg) 249.5 89.4 357.7 212.2 295.5 165
Fiber crude (g) 4.5 1.4 6.4 2.3 5.9 2.2
Retinol (UI) 664.8 279.2 809.2 445.2 990.0 559.4
Carotenoid (UI) 3122 2611 3626 2320 3260 1784
Alfacarotene (mg) 185.3 291.2 144.4 125.8 156.7 152.8
Betacarotene (mg) 1079 849.5 1257 812.2 1165 724.6
Vitamin B (mg) 1.17 0.25 1.81 0.72 1.61 0.63
1
Vitamin B (mg) 1.17 0.28 1.64 0.61 1.43 0.45
2
Vitamin B (mg) 3.14 1.66 5.13 3.08 4.60 2.83
6
Vitamin B (mg) 196.4 70.0 334.81 128.14 290.5 111.8
12
Vitamin C (mg) 70.7 42.7 113.4 77.9 84.6 57.7
Vitamin D (mg) 136.4 76.13 189.5 106.7 155.5 96.13
Vitamin E (TE) 6.83 2.4 3.59 1.2 4.10 1.6
Calcium (mg) 761.3 215.9 1116 431.6 1004 370.6
Iron (mg) 10.9 2.7 14.3 4.9 13.1 2.8
Magnesium (mg) 265.3 67.6 354.8 122.4 324.3 109.7
Phosphorous (mg) 1172.2 307.1 1584 561.8 1470 499.1
Potassium (mg) 1760 541.3 2358.5 902.5 2020.2 676.5
Zinc (mg) 7.4 1.9 8.2 2.5 7.6 2.1
24-hr recalls= mean of 16 24 hours recalls
SD= standard deviation
FFQ= first food frequency questionnaire
1
FFQ= second food frequency questionnaire
2
by beta-carotene (intraclass r= 0.12), whereas the high- especially for total proteins, cholesterol, Vitamin B1 and
est intraclass correlation was observed for carbohy- polyunsaturated fatty acids (Table II).
drate (r= 0.34) and calcium (r= 0.31). Energy-adjusted To assess the validity of the questionnaire, we
coefficients were similar. The reproducibility of the calculated the unadjusted and energy-adjusted Pear-
questionnaire was assessed by means of estimating son correlation coefficients between the average of the
correlation between nutrient scores, measured with the 16 24-hour recalls and the first and second administra-
same instrument twice (Table II). Pearson correlation tion of the FFQ (Table II). Crude unadjusted correla-
coefficients between unadjusted nutrient intake from tion between the first questionnaire and diet recalls
the FFQ, spaced one-year apart, ranged from 0.43 for ranged from 0.13 for Folic acid to 0.52 for magnesium.
Alpha-carotene to 0.60 for monounsaturated fatty ac- Average correlation coefficients of 0.40 or higher were
ids. Adjustment of the nutrient intake for total energy noted for total caloric intake, carbohydrate, protein,
intake tended to decrease these correlations slightly, total fat, animal fat, saturated, monosaturated, and
136 salud pública de méxico / vol.39, no.40, marzo-abril de 1998
no reviews yet
Please Login to review.
