![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Applied Research Program Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, MD 20892-7344; Collaborative Consultants, Bethesda, MD 20817; and Office of Dietary Supplements, National Institutes of Health, Bethesda, MD 20892-7517
1 To whom correspondence should be addressed. E-mail: potischn{at}mail.nih.gov.
 |
ABSTRACT |
|---|
 TOP ABSTRACT RECOMMENDED METHODSLITERATURE CITED |
|---|
KEY WORDS: • National Children's Study • dietary assessment • validation studies • maternal and child health • pregnant and lactating women
In the Children's Health Act of 2000 (PL 106-310), Congress authorized the National Children's Study (NCS), which will examine the effects of environmental influences on the health and development of more than 100,000 children across the United States, following them from before birth until age 21 y [http://nationalchildrensstudy.gov]. Among the most important environmental influences are nutritional factors, which exert wide-ranging influences and contribute to determining outcomes in the study population. The process of selecting dietary assessment methodologies to be recommended for the NCS revealed the complexity involved in assessing the food and dietary supplement intakes of pregnant and lactating women, infants, and children as part of large epidemiologic studies.
Efficient, cost-effective, and valid methods of dietary assessment are critical for epidemiologic studies of hundreds of thousands of subjects. Although considerable research has resulted in accepted methods for adult populations, relatively little has been done to develop and validate dietary methods for large epidemiologic studies of pregnant and lactating women, infants, children, and adolescents. Dietary assessments include the quantity and quality of the diet and supplements to yield data on nutrients, nutrient adequacy, and nonnutritive constituents in foods, as well as food intake patterns. Eating behaviors (meal patterns, timing of meals, eating environment, foods avoided, or unusual foods consumed) are also important because they can positively or negatively affect diet quality and thus health outcomes. Instruments and data collection methodologies still have to be developed and validated to collect information on many of the unique dietary exposures and patterns of pregnant and lactating women and children at various developmental stages.
The following text presents some highlights of the issues that became apparent in the process of identifying the best dietary assessment methodologies to be recommended for the NCS. Identified knowledge gaps and research needs on dietary assessments in large epidemiologic studies are presented. In addition, the reader is directed to various websites for further details on many of these issues.
Identifying assessment methodologies
To develop methodological recommendations for the NCS, we first conducted a literature review of dietary assessment methods utilized in large, epidemiologic studies to assess food and dietary supplement intakes and patterns of use in pregnant or lactating women, infants (0–12 mo), toddlers (13–24 mo), preschoolers (25 mo to 5 y), school age children (6–12 y), and adolescents (13–18 y). A copy of the literature review and tables (1) are available at http://riskfactor.cancer.gov/tools/.
The review examined studies of at least 100 subjects that were conducted mainly among healthy subjects in industrialized, developed countries and were published in English between 1982 and December 2003 for pregnancy, lactation, infants, toddlers, and preschoolers. For school age children, this search built on 2 recent reviews of the literature through 2000 or 2002 (2,3) and focused primarily on the literature published since that time. Studies targeting subjects with eating disorders or specific disease conditions were excluded.
The website contains summary tables of studies in which the relative or criterion validity of one diet or supplement assessment method is evaluated by comparisons with measurements obtained from a reference method. The reference methods vary by study and by target age group and include established dietary assessment methods, biological markers of habitual intake, and techniques such as direct observation of intake. Separate summary tables were prepared, focusing on instruments utilized in large population surveys in each age group.
To enhance the knowledge gained from the literature review, experts were brought together at the National Children's Study Workshop: Dietary Assessment in a Prospective Epidemiologic Study of Pregnant Women and Their Offspring on September 21–22, 2004. Participants included maternal and child health nutritionists, dietary assessment methodologists, nutritional epidemiologists, environmental exposure experts, and infant feeding industry representatives. The workshop's objectives were to identify and review the current state of knowledge about methodologies used to assess dietary intake during pregnancy, lactation, infancy, early childhood, and adolescence and the validity, feasibility, strengths, and limitations associated with these methods during each time period. Delineation of these issues along with summaries of presentations and breakout group discussions are available at http://riskfactor.cancer.gov/tools/.
Recommendations were presented by workshop organizers to NCS regarding the best methodologies to be used for the various age groups within the study population, the timing of data collection across the study period, and additional issues to be considered for potential pilot or substudies. These were based on the literature review, workshop presentations and discussions, and knowledge of the constraints involved in the implementation of the NCS. Some of the methodologies suggested relied on instruments that already exist, whereas others still have to be developed. The recommendations, as presented below, were organized by age group categories.
|
RECOMMENDED METHODS |
|---|
|
TOPABSTRACTRECOMMENDED METHODSLITERATURE CITED |
|---|
It was recommended that 3 FFQ's be administered during the pregnancy and 1 in the postpartum period. For a preconception subcohort, the first FFQ would reflect the previous month, coincident with other data collection in the study. For those first seen at the initial prenatal visit, the first FFQ can be used to collect data on the month before a woman's last menstrual period. At 24–28 wk of pregnancy, coinciding with an oral glucose test, it is suggested that a second FFQ, pregnancy eating behavior questionnaire, and four 24-h recalls (on a subset) be administered. The third FFQ and pregnancy eating behavior questionnaire would be repeated during the third trimester. Finally, 4–8 wk postpartum, the FFQ and a lactation eating behavior questionnaire on the mother's diet would be administered.
Infancy. Assessment of human milk, formula, and infant food intake data would rely on 2 checklists in additional to human milk sampling. The first checklist would focus on milk feeding and would include human milk and different types of formulas (8); another checklist would focus on foods eaten by the child on 3 different days. It may be practical to schedule dietary assessments at 1, 6, and 12 mo when immunizations are given. Human milk sampling would be conducted for estimations of intakes of nutrient and other bioactive food components as well as pesticides/contaminants. Samples can be taken using breast pumps at clinics with vessels that are unlikely to interfere with environmental contaminants hypotheses.
Toddler and preschool. Dietary assessments in the NCS for this age group would include an age and culturally appropriate FFQ, a food behavior questionnaire, an expanded 3-d checklist, and a food security questionnaire. The 3-d checklists would be used to calibrate the FFQ data and serve as a simple method for data collection from daycare providers. The checklists help to address the concern of collecting reliable data about children's diets when much of their food is eaten away from their parents. The FFQ, behavior questionnaire, and food security questions would be administered during checkups at 18 mo, 3 y, and 5 y. Expanded 3-d checklists would have been distributed to parents at least 2 wk before these visits and would be brought to the clinical centers with the child. It was suggested that the checklist be tested in daycare centers and with other child caregivers to determine what is practical and feasible. Many of these instruments and data collection methods used were not developed for this diverse population. All instruments should be pilot tested in major subgroups and refined for subgroups of the cohort.
School-aged children (6–11 y). The methods suggested for this group included an FFQ, checklist, and food security questionnaire. One such instrument available for this age group is the "What are kids eating now?" [www.nutritionquest.com], but others may be available and may be more appropriate depending on the population being studied. Data would be collected once between the ages of 6 and 7 y with the parent as the respondent and once between 9 and 10 y with both the parent and the child as respondents. Developmental changes within this age dictate that additional research is warranted to identify the validity and practicality of various methods. Many instruments being used in this age group were not developed for this population and would likely require refinement for this study.
Adolescents. It was suggested that an FFQ and multiple food recalls be used with adolescents, assuming that automated technology is available. These would be administered once in junior high school and once in high school. Social, emotional, and physical determinants of food intake are large issues in this age group. Further, it is not known how best to address them in the context of large studies. The long time period from the study initiation until the need for an adolescent instrument should allow for completion of development of new methodologies and their validation. Ongoing studies of adolescents, such as the Helena project (http://www.helenastudy.com) will help inform the choice of instruments and relevant issues that require further consideration in adolescents.
Knowledge gaps and research needs
Knowledge gaps in dietary assessment methodologies among this population group were identified in the literature review and at the workshop. They are included in the following broad categories: the development of additional validation studies; understanding the effect of age on the selection of a survey respondent; understanding the effect of social desirability factors on the reporting of dietary intake; the development of assessment methods and estimation aides for self-administered use; and coordination with large international epidemiologic studies.
Validation studies. The literature review and workshop presentations consistently pointed to the relatively limited number of validation studies, especially studies using biomarkers that are highly sensitive and specific to intake (see Summary Table 1.1 at http://riskfactor.cancer.gov/tools/) (9). There is also a need for the development of appropriate dietary assessment data collection and analytic methodologies for the various stages of life cohorts. Among the studies that were conducted, most have been among white, upper-income, and well-educated groups. Appropriate instruments for low-literacy populations and various ethnic groups must be developed and validated across all age groups.
Most of the studies were assessments of relative validity, i.e., they compared a new measurement method with a more established dietary method that is believed to have a greater degree of face validity (10). Validity can better be assessed by comparing a new instrument or method with an independent reference measurement such as a biomarker of intake, preferably one that is not influenced by genetics, metabolism, or environmental factors. The doubly labeled water (DLW) method has been used as such, although this biomarker is limited to validating energy intake. The need remains to identify and study additional biomarkers of nutrient intake and novel methodologies to analyze simultaneously multiple aspects of nutritional status (9,11). Finally, although there were some studies assessing dietary supplement use, mostly with FFQ, additional methods have to be developed and validated.
Age effect on choosing respondents. There are 2 main knowledge gaps that reflect upon the age of a child and the proper selection of a survey respondent. Among toddlers or preschoolers who spend the majority of their day away from their parents or primary caregivers, and thus consume a large portion of their daily intake away from them, there is a need to develop and validate data collection methodologies that can be used to collect dietary intake data from secondary caregivers such as daycare providers or babysitters. Issues that must be considered include the multiple responsibilities (other children) of a daycare provider, a lack of primary motivation as a survey respondent, the numerous types of daycare settings, and the potential need to train daycare providers and parents to transfer information appropriately to one recording instrument.
The second knowledge gap is determination of the appropriate age at which children should report intake for themselves without assistance from their parents. Factors affecting this include the cognitive skills of different age groups to complete a dietary assessment accurately by understanding what is being asked and accurately recalling the foods eaten, times at which they were eaten, and the names of all foods (3,12). The accuracy of recall or reporting may depend on the type of method used or the timing of the assessment. Often parents are asked to assist children in their responses to dietary assessments and the effect of this assistance should be explored.
Social desirability and reporting of dietary intake. In each of the stage-of-life cohorts among this population, there are various social desirability factors that might bias the reporting of dietary intake. Pregnant and lactating women's reports of dietary intake may be affected by how their weight gain or loss compares with their beliefs about appropriate weight gain during pregnancy or loss after pregnancy. Beliefs in what a pregnant or lactating woman is supposed to eat (or not eat) to be able to provide her unborn or nursing child with the best possible nutrition may also bias reports of dietary intake or eating patterns. Parents of young children may let their desire to appear to have good parenting skills (reflected by food intake and eating patterns) influence their reporting of dietary intake and eating behaviors. Similarly, older children and adolescents often use food as a means of self-expression (3). Their reporting may be biased by their current food and body-image–related beliefs (which are apt to change throughout adolescence). There is a need to identify these and other social desirability factors across the ages and to determine how best to lessen their potential bias on data collection.
Self-administered assessment methods. As we continue to work in a technologically oriented environment, it is natural that people will become more comfortable using a self-administered automated data collection system. These systems allow researchers to broaden their selection of respondents without concern about the availability or costs of interviewers, especially with hard-to-reach populations. Several automated applications have already been developed, although few if any have been validated in the NCS groups of interest. Three examples of automated applications of FFQ are the NCI web-based Diet History Questionnaire [http://riskfactor.cancer.gov/DHQ]; the Block Online FFQ [www.nutritionquest.com], and the Fred Hutchinson Cancer Research Center VioFFQ [weiss@viocare.com]. All 3 instruments have been used in research, are being designed as web-based instruments, and will be available in the near future. The University of North Carolina developed a web-based Diet History application [www.diethistory.com] that collects habitual diet over the last 12 mo. Three examples of web-based Food Record applications include BalanceLog [www.healthetech.com], DINE Healthy 4.2 [www.dinesystems.com], and the USDA Interactive Healthy Eating Index [www.usda.gov/cnpp/ihei.html]. All 3 applications are supported by USDA databases. Examples of 24-h recall instruments that are electronically based include the USDA Automated Multiple Pass Method; the Baylor College of Medicine Food Intake Recording Software System, and the Colorado State University Bi-Lingual Interactive Multi-Media Computerized Food Recall. These instruments are not yet publicly available. The NCI is examining the use of a checklist instrument to calibrate food frequency information obtained from a FFQ. After several rounds of cognitive testing, the NCI has developed a machine-scannable checklist instrument to be coupled with its Diet History Questionnaire [http://riskfactor.cancer.gov/diet/screeners/daily.html].
Research is warranted to determine the validity of these and similar instruments across the different age and ethnic groups in this study population. Research also can help determine which method or combination of methods is best used by different subgroups of the population, the biases that the use of the instrument may introduce (or exclude), and how they can successfully be made appealing to the population.
Comparisons with international studies. Coordination of U.S.-based large epidemiologic studies with other large international studies could generate comparable data across studies allowing for combined analyses that would yield greater power and exposure variations, as well as make it possible to replicate findings. This is particularly relevant to rare diseases, which would be present in only a small number of children in any one study but could be pooled from various cohorts for research purposes. Two existing comparable studies are the Danish National Birth Cohort (http://www.ssi.dk/sw9314.asp) and the Norwegian Mother and Child Cohort (http://www.fhi.no/artikler/?id=51488) with dietary instruments for many of the groups relevant to the NCS.
In summary, the National Children's Study will involve a large cross section of the U.S. population, with geographic and ethnic diversity incorporated into the design. A limited number of instruments have been validated in the population groups of interest to the National Children's Study. Knowledge gaps in dietary assessment methodologies were identified and it is anticipated that work will be undertaken to develop new instruments, validate existing instruments, understand the effect of social and developmental factors on the reporting of dietary intake, and pursue coordination with other large studies of pregnant women and their offspring.
Manuscript received 10 October 2005.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTRECOMMENDED METHODSLITERATURE CITED |
|---|
1. National Institute on Child Health and Development. National Children's Study literature review on the assessment of food and dietary supplement intakes and patterns of use in women and young children. National Children's Study. Final report; 2004.
2. McPherson RS, Hoelscher DM, Alexander M, Scanlon KS, Serdula MK. Dietary assessment methods among school-aged children: validity and reliability. Prev Med. 2000;31:S11–33.
3. Livingstone MB, Robson PJ. Measurement of dietary intake in children. Proc Nutr Soc. 2000;59:279–93.[Medline]
4. Siega-Riz AM, Herrmann TS, Savitz DA, Thorp JM. Frequency of eating during pregnancy and its effect on preterm delivery. Am J Epidemiol. 2001;153:647–52.
5. Bayley TM, Dye L, Jones S, DeBono M, Hill AJ. Food cravings and aversions during pregnancy: relationships with nausea and vomiting. Appetite. 2002;38:45–51.[Medline]
6. Pope JF, Skinner JD, Carruth BR. Cravings and aversions of pregnant adolescents. J Am Diet Assoc. 1992;92:1479–82.[Medline]
7. Nord M, Bickel G. Measuring children's food security in U.S. households, 1995–99. U.S. Department of Agriculture. Food Assistance and Nutrition Research Report No. 25; 2002.
8. Ryan AS, Wenjun Z, Acosta A. Breastfeeding continues to increase into the new millennium. Pediatrics. 2002;110:1103–9.
9. Potischman N. Biologic and methodologic issues for nutritional biomarkers. J Nutr. 2003;133 (suppl. 3):875S–80.
10. Block G, Hartman AM. Issues in reproducibility and validity of dietary studies. Am J Clin Nutr. 1989;50:1133–8; discussion 1231–5
11. Daures JP, Gerber M, Scali J, Astre C, Bonifacj C, Kaaks R. Validation of a food-frequency questionnaire using multiple-day records and biochemical markers: application of the triads method. J Epidemiol Biostat. 200;5:109–15.
12. Baranowski T, Domel SB. A cognitive model of children's reporting of food intake. Am J Clin Nutr. 1994;59 (suppl. 1):212S–7
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
