This article was downloaded by: [Purdue University] On: 13 April 2015, At: 00:49 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of the American College of Nutrition Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uacn20
Use of a modified food frequency questionnaire during pregnancy. a
a
a
S Greeley , L Storbakken & R Magel a
Department of Food and Nutrition, North Dakota State University, Fargo 58105. Published online: 02 Sep 2013.
To cite this article: S Greeley, L Storbakken & R Magel (1992) Use of a modified food frequency questionnaire during pregnancy., Journal of the American College of Nutrition, 11:6, 728-734, DOI: 10.1080/07315724.1992.10718274 To link to this article: http://dx.doi.org/10.1080/07315724.1992.10718274
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Use of a Modified Food Frequency Questionnaire During Pregnancy Sharon Greeley, PhD, FACN, Lisa Storbakken, MS, LRD, and Rhonda Magel, PhD Department of Food and Nutrition (S.G., L.S.) and Department of Statistics (RM.), North Dakota State University, Fargo
Downloaded by [Purdue University] at 00:49 13 April 2015
Key words: pregnancy, nutrient assessment tools, food frequency questionnaire The purpose of this study was to examine the usefulness of the self-administered modified Willett food frequency questionnaire (mWFFQ) relative to duplicate 24-hour recall interviews as a dietary assessment tool during pregnancy. Fifty healthy pregnant women living in the upper Midwest were recruited for the study. At gestational weeks 16 and 21 (second trimester) and 30 and 35 (third trimester) the women were interviewed at home; they provided oral responses concerning their food and beverage consumption during the previous 24 hours. At weeks 21 and 35, immediately following the interviews, the women were asked to complete an optically scannable food frequency questionnaire, modified to evaluate daily food consumption during a 2-month time period (mWFFQ). Kilocalories and seven nutrients were computed from all dietary data, based on USDA Handbooks 8 and 456. Regression analyses were performed between nutrient intake from the 24-hour recall averages and the nutrient data generated by the mWFFQ, with the trimesters serving as independent time periods. According to our preset criterion, correlation coefficients indicated that the assessment tools were comparable during the third but not the second trimester. During both trials, the mWFFQ provided consistently higher kilocalorie and nutrient values than did the 24-hour recall with the exception of fat intake. Based on our findings we conclude that the mWFFQ is a useful tool for assessing nutrient intake of groups of pregnant women, but that familiarity with the characteristics of the participants and the instrument may be important to data interpretation.
Abbreviations: FFQ = food frequency questionnaire, mWFFQ = modified Willett food frequency questionnaire, WFFQ = Willett food frequency questionnaire
INTRODUCTION
records of written or weighed food items. However, it remains critically important to assess the nutrient status of such groups, as these demographic traits are often associ ated with an increased prevalence of "high risk pregnan cies." Food frequency questionnaires (FFQ) have been shown to be useful epidemiologie tools to demonstrate relation ships between dietary patterns and specific health out comes [5,6]. For example, FFQ indicate that certain chronic eating patterns appear to be related to the devel opment of cardiovascular disease or cancer [6]. Investiga tors in the area of reproduction have long been interested in the relationship between food consumption patterns during pregnancy and intrauterine development [7-10]. A FFQ may be helpful in showing a relationship between food intake patterns during gestation and the delivery of small-for-date infants and/or poor maternal weight gain. Willett and coworkers developed [11] and validated [12] a self-administered FFQ that was designed to assess frequency of commonly consumed foods and beverages
Despite the continuing need for accurate nutrient intake assessment during pregnancy, a tool that effectively com bines accuracy, reliability and ease of administration at a reasonable cost remains unavailable. Currently the method of choice is the 24-hour recall, although nutritionists in volved in data collection concede that food consumption fluctuates during the 40-week gestation so this traditional method may not be well suited to measuring the nutrient intake of pregnant women [1-3]. Rush et al [4] proposed that repeated 24-hour recalls will help to overcome this inherent problem, but highly skilled interviewers are re quired if data are to be consistent and credible. Food records have been used to assess nutrient status during pregnancy [1,4] but the technique is limited by specific demographic factors of the population being eval uated. Those who are under-educated, poor, culturally disadvantaged or subject to severe time limitations (e.g., working mothers) are unlikely target groups for the use of
Address reprint requests to Sharon Greeley, PhD, Department of Food and Nutrition, NDSU, Fargo, ND 58105. Journal of the American College of Nutrition, Vol. 11, No. 6, 728-734 (1992) Published by the American College of Nutrition 728
FFQ and Pregnancy
Downloaded by [Purdue University] at 00:49 13 April 2015
during a 12-month period. A 1-year diet record served as the basis of comparison, and validation suggested that the questionnaire could be used to provide accurate informa tion on the nutrient intake of groups and individuals over an extended period of time. The purpose of our study was to examine the usefulness of the Willett food frequency questionnaire (WFFQ), orig inally developed to weigh disease risk, as a dietary assess ment tool during pregnancy. Given the developmental time constraints of gestation, we chose to use a Willett instrument that was modified to evaluate a 2-month (mWFFQ), rather than the original 12-month period of food consumption. The average intakes of selected nu trients from duplicate 24-hour recalls of food consumption and food frequency intake data were compared. We per formed regression analyses to examine the relationship between the two intake instruments: initially during the second trimester and finally during the third trimester.
sample of pregnant women from an urban community in eastern North Dakota participated in the study. All were 20-35 years of age. No woman was excluded from the study based on ethnicity or parity. Fifty-six women were recruited and 50 completed the study. Six were dismissed because of pregnancy complication (e.g., pregnancy-in duced hypertension or gestational diabetes), miscarriage, or delivery of more than one infant. Approximately 40% were primipara, and 60% were multipara. Nearly 74% were employed outside the home. Consent forms (ap proved by NDSU Institutional Review Board) were given to prospective subjects, and signatures were obtained from all participants at the initial meeting before health, repro ductive or nutritional information was collected. Schedule Last menstrual period recorded by the physician was the date used to estimate gestational age of the individual participants. Each woman was contacted twice in her second trimester and twice in her third trimester. Appoint ments were arranged for gestational weeks 16, 21, 30, and 35. Twenty-four-hour oral recall interviews were scheduled for each of these gestational weeks. Recall days were based primarily on convenience to the participant and we note that weekdays were more frequently (4:1) selected than Saturdays or Sundays. Immediately following oral recalls at weeks 21 and 35, the participants were presented with
MATERIALS AND METHODS Subjects Our objective was to provide a comparison between the data from the mWFFQ and the average of two 24-hour recall interviews. Study participants were recruited directly between June 1988 and February 1989. A convenience
Table 1. Comparison of the 24-Hour Recall and the m WFFQ by Trimester (Daily Intake, Pearson Correlation Coefficient and P Values) Second trimester
Energy (kcal/day) Protein (g/day) Fat (g/day) CHOt (g/day) Fe (mg/day) Ca (mg/day) Vit C (mg/day) Folacin (meg/day)
Mean 24-hour recall*
Mean mWFFQ*
Product moment correlation coefficient (r)
Significance (P)
2009 83 81 243 13 1140 113 224
2157 98 76 279 16 1419 169 341
-0.01 -0.04 0.14 0.01 0.22 0.19 0.48 0.39
0.94 0.78 0.32 0.97 0.12 0.18 0.01 0.01
0.25 0.23 0.28 0.32 0.56 0.48 0.52 0.48
0.08 0.11 0.05 0.02 0.01 0.01 0.01 0.01
Third trimester Energy (kcal/day) Protein (g/day) Fat (g/day) CHOf (g/day) Fe (mg/day) Ca (mg/day) Vit C (mg/day) Folacin (meg/day)
2041 87 78 255 13 1366 118 245
2182 99 74 290 17 1479 170 343
* Based on a sample size of 50. t Carbohydrate.
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION
729
FFQ and Pregnancy F i t — a n d
T r l r r m t i r
and 456. Kilocalories and 16 "nutrients" were computed from the frequency consumption patterns and expressed as nutrient-intake-units/24 hours; we received "hard copy" as well as a computer tape disc of the entire data set from Channing Laboratory. All 24-hour oral recall interviews were conducted with a simple, single-page form that included a checklist for five food/beverage categories, time of day, quantity, and prep aration. Plastic food models and ordinary kitchen utensils assisted the subjects in estimating serving size. Kilocalories and seven nutrients were calculated from food intake portions and expressed as intake-units/24 hours, using a program (Computrition®) based on USDA Handbooks 8 and 456.
Downloaded by [Purdue University] at 00:49 13 April 2015
Statistical Analysis Dietary intakes from both tools included food and beverage data; prenatal vitamin and mineral supplements were excluded. Average nutrient intakes from the 24-hour recalls were computed for weeks 16 and 21 (representing the second trimester) and for weeks 30 and 35 (representing
Fig. 1. Mean intake comparison of fat from the mWFFQ and the 24-hour recall by trimester.
the mWFFQ, provided with verbal instructions and asked to complete the questionnaire. Body weights were meas ured directly at weeks 16 and 30 and were self-reported at weeks 21 and 35. Infant birth weight was reported as "hospital delivery weight" by the participants. Instruments: mWFFQ and 24-Hour Recall WFFQ were purchased directly from Channing Labo ratories in Boston, MA, and were modified by us to include the phrase "average food use over the last 2 months." A similarly modified questionnaire has been used success fully with nonpregnant populations (personal communi cation, Susan Crockett, University of Syracuse). The opti cally scannable questionnaire sheet contains 116 food items with a nine choice "frequency range of "none to six/ day." The participant completes the questionnaire by iden tifying the frequency that corresponds to food and beverage items consumed by her during the past 2 months. After all data were collected, the questionnaires were returned to Channing Laboratory for computerized nutrient analysis. The data base was compiled from USDA Handbooks 8
730
a
m
to
-is
24-hr
so
ΛΜ
ao
an
*
Raoall
Fig. 2. Mean intake comparison of iron from the mWFFQ and the 24-hour recall by trimester.
VOL. 11, NO. 6
Downloaded by [Purdue University] at 00:49 13 April 2015
FFQ and Pregnancy the third trimester). The average 24-hour recall values served as the basis of comparison with the mWFFQ. The trimesters were regarded as independent time periods and their corresponding data were analyzed separately. Regres sion analysis was performed between the 24-hour recall averages (X) and the mWFFQ data (Y) for seven nutrients and the total energy intake. Pearson product moment correlation coefficients were also computed. The dietary variables considered were: protein (g/day), fat (g/day), carbohydrate (g/day), iron (mg/day), calcium (mg/day), ascorbic acid (mg/day), folacin (meg/day) and total energy intake (kcal/day). In keeping with Willett et al [12], we chose to accept the mWFFQ as valid relative to the 24-hour recall if >60% (5 of 8) dietary variables correlated significantly (p < 0.05) with the counterpart from the average 24-hour recall. Because certain data subsets contained outliers, both par ametric and nonparametric tests were applied to the data. The computer package SAS® was used to perform all computations and statistical analyses.
RESULTS Of the original 56 women recruited, 50 (89%) met our health criteria and completed the dietary assessment study. Pregnancies were uncomplicated and vaginal deliveries were near anticipated due dates. All infants were born in good health according to Apgar scores and delivery weights (data not shown). Pearson product moment sample correlation coeffi cients between the m WFFQ and the average 24-hour recall for energy and seven nutrients are presented in Table 1. The corresponding p-values demonstrated a significant correlation (p < 0.05) between instruments for two nu trients during the second trimester: vitamin C and folacin. Results for the third trimester are shown in the lower panel of Table 1. A significant product moment correlation (p < 0.05) was found between the mWFFQ and the average 24hour recall for six nutrients in the third trimester. Only two items for the mWFFQ and the average 24-hour recall were not significantly correlated: energy (p = 0.08) and protein (p = 0.11). Typical scatterplots showing the rela tionship between the mWFFQ and the average 24-hour recall for fat, iron and folacin are given in Figures 1-3. After examining the residual plots between the average 24-hour recall and m WFFQ for each of the eight items during both trimesters, the assumption of normal distri bution of the error term was clearly not appropriate in all cases. Specifically, normality assumption was questionable for energy, fat, iron, vitamin C, and folacin in the second trimester, and folacin in the third trimester. Therefore, Spearman's correlation coefficients were computed for all variables in both trimesters. The results of the Spearman's
Fig. 3. Mean intake comparison of folacin from the mWFFQ and the 24-hour recall by trimester. correlation test and the corresponding p-values are shown in Table 2. The Spearman's correlation coefficient showed significance (p < 0.05) for folacin during the second trimes ter. Interestingly, vitamin C (significant with the Pearson product moment correlation coefficient) gave a p-value of 0.057 when using the Spearman's correlation coefficient. Five of the eight items (>60%) in the third trimester were found to be significant (p < 0.05) with the nonparametric test, compared to 6 of 8 (>70%) with the parametric test. A significant correlation for fat was found only in the third trimester and only with the Pearson correlation coefficient (Table 1). The differences between group means (all subjects) for each of the eight items using the mWFFQ and the average 24-hour recall were compared for both trimesters and are presented in Table 3. During the second trimester, total energy intake, protein, fat, carbohydrate, calcium, and iron values were within 20% of each other according to the two instruments. The group mean differences for vitamin C and folacin were greater still, at 33 and 34%, respectively. Overall, the mWFFQ showed higher group mean values than the average 24-hour recall values for six nutrients and
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION
731
FFQ and Pregnancy Table 2. Comparison of the 24-Hour Recall and the mWFFQ by Trimester (Daily Intake, Spearman's Rank Correlation Coefficient and P Values) Second trimester
Energy (kcal/day) Protein (g/day) Fat (g/day) CHOt (g/day) Fe (mg/day) Ca (mg/day) Vit. C (mg/day) Folacin (meg/day)
Mean 24-hr. recall*
Mean mWFFQ*
2009 83 81 243 13 1140 113 224
2157 98 76 279 16 1419 169 341
Spearman's
Significance (P)
(r) -0.05 -0.03 0.03 0.02 0.20 0.21 0.27 0.46
0.75 0.83 0.82 0.92 0.16 0.15 0.06
Journal of the American College of Nutrition Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uacn20
Use of a modified food frequency questionnaire during pregnancy. a
a
a
S Greeley , L Storbakken & R Magel a
Department of Food and Nutrition, North Dakota State University, Fargo 58105. Published online: 02 Sep 2013.
To cite this article: S Greeley, L Storbakken & R Magel (1992) Use of a modified food frequency questionnaire during pregnancy., Journal of the American College of Nutrition, 11:6, 728-734, DOI: 10.1080/07315724.1992.10718274 To link to this article: http://dx.doi.org/10.1080/07315724.1992.10718274
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Use of a Modified Food Frequency Questionnaire During Pregnancy Sharon Greeley, PhD, FACN, Lisa Storbakken, MS, LRD, and Rhonda Magel, PhD Department of Food and Nutrition (S.G., L.S.) and Department of Statistics (RM.), North Dakota State University, Fargo
Downloaded by [Purdue University] at 00:49 13 April 2015
Key words: pregnancy, nutrient assessment tools, food frequency questionnaire The purpose of this study was to examine the usefulness of the self-administered modified Willett food frequency questionnaire (mWFFQ) relative to duplicate 24-hour recall interviews as a dietary assessment tool during pregnancy. Fifty healthy pregnant women living in the upper Midwest were recruited for the study. At gestational weeks 16 and 21 (second trimester) and 30 and 35 (third trimester) the women were interviewed at home; they provided oral responses concerning their food and beverage consumption during the previous 24 hours. At weeks 21 and 35, immediately following the interviews, the women were asked to complete an optically scannable food frequency questionnaire, modified to evaluate daily food consumption during a 2-month time period (mWFFQ). Kilocalories and seven nutrients were computed from all dietary data, based on USDA Handbooks 8 and 456. Regression analyses were performed between nutrient intake from the 24-hour recall averages and the nutrient data generated by the mWFFQ, with the trimesters serving as independent time periods. According to our preset criterion, correlation coefficients indicated that the assessment tools were comparable during the third but not the second trimester. During both trials, the mWFFQ provided consistently higher kilocalorie and nutrient values than did the 24-hour recall with the exception of fat intake. Based on our findings we conclude that the mWFFQ is a useful tool for assessing nutrient intake of groups of pregnant women, but that familiarity with the characteristics of the participants and the instrument may be important to data interpretation.
Abbreviations: FFQ = food frequency questionnaire, mWFFQ = modified Willett food frequency questionnaire, WFFQ = Willett food frequency questionnaire
INTRODUCTION
records of written or weighed food items. However, it remains critically important to assess the nutrient status of such groups, as these demographic traits are often associ ated with an increased prevalence of "high risk pregnan cies." Food frequency questionnaires (FFQ) have been shown to be useful epidemiologie tools to demonstrate relation ships between dietary patterns and specific health out comes [5,6]. For example, FFQ indicate that certain chronic eating patterns appear to be related to the devel opment of cardiovascular disease or cancer [6]. Investiga tors in the area of reproduction have long been interested in the relationship between food consumption patterns during pregnancy and intrauterine development [7-10]. A FFQ may be helpful in showing a relationship between food intake patterns during gestation and the delivery of small-for-date infants and/or poor maternal weight gain. Willett and coworkers developed [11] and validated [12] a self-administered FFQ that was designed to assess frequency of commonly consumed foods and beverages
Despite the continuing need for accurate nutrient intake assessment during pregnancy, a tool that effectively com bines accuracy, reliability and ease of administration at a reasonable cost remains unavailable. Currently the method of choice is the 24-hour recall, although nutritionists in volved in data collection concede that food consumption fluctuates during the 40-week gestation so this traditional method may not be well suited to measuring the nutrient intake of pregnant women [1-3]. Rush et al [4] proposed that repeated 24-hour recalls will help to overcome this inherent problem, but highly skilled interviewers are re quired if data are to be consistent and credible. Food records have been used to assess nutrient status during pregnancy [1,4] but the technique is limited by specific demographic factors of the population being eval uated. Those who are under-educated, poor, culturally disadvantaged or subject to severe time limitations (e.g., working mothers) are unlikely target groups for the use of
Address reprint requests to Sharon Greeley, PhD, Department of Food and Nutrition, NDSU, Fargo, ND 58105. Journal of the American College of Nutrition, Vol. 11, No. 6, 728-734 (1992) Published by the American College of Nutrition 728
FFQ and Pregnancy
Downloaded by [Purdue University] at 00:49 13 April 2015
during a 12-month period. A 1-year diet record served as the basis of comparison, and validation suggested that the questionnaire could be used to provide accurate informa tion on the nutrient intake of groups and individuals over an extended period of time. The purpose of our study was to examine the usefulness of the Willett food frequency questionnaire (WFFQ), orig inally developed to weigh disease risk, as a dietary assess ment tool during pregnancy. Given the developmental time constraints of gestation, we chose to use a Willett instrument that was modified to evaluate a 2-month (mWFFQ), rather than the original 12-month period of food consumption. The average intakes of selected nu trients from duplicate 24-hour recalls of food consumption and food frequency intake data were compared. We per formed regression analyses to examine the relationship between the two intake instruments: initially during the second trimester and finally during the third trimester.
sample of pregnant women from an urban community in eastern North Dakota participated in the study. All were 20-35 years of age. No woman was excluded from the study based on ethnicity or parity. Fifty-six women were recruited and 50 completed the study. Six were dismissed because of pregnancy complication (e.g., pregnancy-in duced hypertension or gestational diabetes), miscarriage, or delivery of more than one infant. Approximately 40% were primipara, and 60% were multipara. Nearly 74% were employed outside the home. Consent forms (ap proved by NDSU Institutional Review Board) were given to prospective subjects, and signatures were obtained from all participants at the initial meeting before health, repro ductive or nutritional information was collected. Schedule Last menstrual period recorded by the physician was the date used to estimate gestational age of the individual participants. Each woman was contacted twice in her second trimester and twice in her third trimester. Appoint ments were arranged for gestational weeks 16, 21, 30, and 35. Twenty-four-hour oral recall interviews were scheduled for each of these gestational weeks. Recall days were based primarily on convenience to the participant and we note that weekdays were more frequently (4:1) selected than Saturdays or Sundays. Immediately following oral recalls at weeks 21 and 35, the participants were presented with
MATERIALS AND METHODS Subjects Our objective was to provide a comparison between the data from the mWFFQ and the average of two 24-hour recall interviews. Study participants were recruited directly between June 1988 and February 1989. A convenience
Table 1. Comparison of the 24-Hour Recall and the m WFFQ by Trimester (Daily Intake, Pearson Correlation Coefficient and P Values) Second trimester
Energy (kcal/day) Protein (g/day) Fat (g/day) CHOt (g/day) Fe (mg/day) Ca (mg/day) Vit C (mg/day) Folacin (meg/day)
Mean 24-hour recall*
Mean mWFFQ*
Product moment correlation coefficient (r)
Significance (P)
2009 83 81 243 13 1140 113 224
2157 98 76 279 16 1419 169 341
-0.01 -0.04 0.14 0.01 0.22 0.19 0.48 0.39
0.94 0.78 0.32 0.97 0.12 0.18 0.01 0.01
0.25 0.23 0.28 0.32 0.56 0.48 0.52 0.48
0.08 0.11 0.05 0.02 0.01 0.01 0.01 0.01
Third trimester Energy (kcal/day) Protein (g/day) Fat (g/day) CHOf (g/day) Fe (mg/day) Ca (mg/day) Vit C (mg/day) Folacin (meg/day)
2041 87 78 255 13 1366 118 245
2182 99 74 290 17 1479 170 343
* Based on a sample size of 50. t Carbohydrate.
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION
729
FFQ and Pregnancy F i t — a n d
T r l r r m t i r
and 456. Kilocalories and 16 "nutrients" were computed from the frequency consumption patterns and expressed as nutrient-intake-units/24 hours; we received "hard copy" as well as a computer tape disc of the entire data set from Channing Laboratory. All 24-hour oral recall interviews were conducted with a simple, single-page form that included a checklist for five food/beverage categories, time of day, quantity, and prep aration. Plastic food models and ordinary kitchen utensils assisted the subjects in estimating serving size. Kilocalories and seven nutrients were calculated from food intake portions and expressed as intake-units/24 hours, using a program (Computrition®) based on USDA Handbooks 8 and 456.
Downloaded by [Purdue University] at 00:49 13 April 2015
Statistical Analysis Dietary intakes from both tools included food and beverage data; prenatal vitamin and mineral supplements were excluded. Average nutrient intakes from the 24-hour recalls were computed for weeks 16 and 21 (representing the second trimester) and for weeks 30 and 35 (representing
Fig. 1. Mean intake comparison of fat from the mWFFQ and the 24-hour recall by trimester.
the mWFFQ, provided with verbal instructions and asked to complete the questionnaire. Body weights were meas ured directly at weeks 16 and 30 and were self-reported at weeks 21 and 35. Infant birth weight was reported as "hospital delivery weight" by the participants. Instruments: mWFFQ and 24-Hour Recall WFFQ were purchased directly from Channing Labo ratories in Boston, MA, and were modified by us to include the phrase "average food use over the last 2 months." A similarly modified questionnaire has been used success fully with nonpregnant populations (personal communi cation, Susan Crockett, University of Syracuse). The opti cally scannable questionnaire sheet contains 116 food items with a nine choice "frequency range of "none to six/ day." The participant completes the questionnaire by iden tifying the frequency that corresponds to food and beverage items consumed by her during the past 2 months. After all data were collected, the questionnaires were returned to Channing Laboratory for computerized nutrient analysis. The data base was compiled from USDA Handbooks 8
730
a
m
to
-is
24-hr
so
ΛΜ
ao
an
*
Raoall
Fig. 2. Mean intake comparison of iron from the mWFFQ and the 24-hour recall by trimester.
VOL. 11, NO. 6
Downloaded by [Purdue University] at 00:49 13 April 2015
FFQ and Pregnancy the third trimester). The average 24-hour recall values served as the basis of comparison with the mWFFQ. The trimesters were regarded as independent time periods and their corresponding data were analyzed separately. Regres sion analysis was performed between the 24-hour recall averages (X) and the mWFFQ data (Y) for seven nutrients and the total energy intake. Pearson product moment correlation coefficients were also computed. The dietary variables considered were: protein (g/day), fat (g/day), carbohydrate (g/day), iron (mg/day), calcium (mg/day), ascorbic acid (mg/day), folacin (meg/day) and total energy intake (kcal/day). In keeping with Willett et al [12], we chose to accept the mWFFQ as valid relative to the 24-hour recall if >60% (5 of 8) dietary variables correlated significantly (p < 0.05) with the counterpart from the average 24-hour recall. Because certain data subsets contained outliers, both par ametric and nonparametric tests were applied to the data. The computer package SAS® was used to perform all computations and statistical analyses.
RESULTS Of the original 56 women recruited, 50 (89%) met our health criteria and completed the dietary assessment study. Pregnancies were uncomplicated and vaginal deliveries were near anticipated due dates. All infants were born in good health according to Apgar scores and delivery weights (data not shown). Pearson product moment sample correlation coeffi cients between the m WFFQ and the average 24-hour recall for energy and seven nutrients are presented in Table 1. The corresponding p-values demonstrated a significant correlation (p < 0.05) between instruments for two nu trients during the second trimester: vitamin C and folacin. Results for the third trimester are shown in the lower panel of Table 1. A significant product moment correlation (p < 0.05) was found between the mWFFQ and the average 24hour recall for six nutrients in the third trimester. Only two items for the mWFFQ and the average 24-hour recall were not significantly correlated: energy (p = 0.08) and protein (p = 0.11). Typical scatterplots showing the rela tionship between the mWFFQ and the average 24-hour recall for fat, iron and folacin are given in Figures 1-3. After examining the residual plots between the average 24-hour recall and m WFFQ for each of the eight items during both trimesters, the assumption of normal distri bution of the error term was clearly not appropriate in all cases. Specifically, normality assumption was questionable for energy, fat, iron, vitamin C, and folacin in the second trimester, and folacin in the third trimester. Therefore, Spearman's correlation coefficients were computed for all variables in both trimesters. The results of the Spearman's
Fig. 3. Mean intake comparison of folacin from the mWFFQ and the 24-hour recall by trimester. correlation test and the corresponding p-values are shown in Table 2. The Spearman's correlation coefficient showed significance (p < 0.05) for folacin during the second trimes ter. Interestingly, vitamin C (significant with the Pearson product moment correlation coefficient) gave a p-value of 0.057 when using the Spearman's correlation coefficient. Five of the eight items (>60%) in the third trimester were found to be significant (p < 0.05) with the nonparametric test, compared to 6 of 8 (>70%) with the parametric test. A significant correlation for fat was found only in the third trimester and only with the Pearson correlation coefficient (Table 1). The differences between group means (all subjects) for each of the eight items using the mWFFQ and the average 24-hour recall were compared for both trimesters and are presented in Table 3. During the second trimester, total energy intake, protein, fat, carbohydrate, calcium, and iron values were within 20% of each other according to the two instruments. The group mean differences for vitamin C and folacin were greater still, at 33 and 34%, respectively. Overall, the mWFFQ showed higher group mean values than the average 24-hour recall values for six nutrients and
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION
731
FFQ and Pregnancy Table 2. Comparison of the 24-Hour Recall and the mWFFQ by Trimester (Daily Intake, Spearman's Rank Correlation Coefficient and P Values) Second trimester
Energy (kcal/day) Protein (g/day) Fat (g/day) CHOt (g/day) Fe (mg/day) Ca (mg/day) Vit. C (mg/day) Folacin (meg/day)
Mean 24-hr. recall*
Mean mWFFQ*
2009 83 81 243 13 1140 113 224
2157 98 76 279 16 1419 169 341
Spearman's
Significance (P)
(r) -0.05 -0.03 0.03 0.02 0.20 0.21 0.27 0.46
0.75 0.83 0.82 0.92 0.16 0.15 0.06
