5

Moving Forward

The broad range of research areas laid out in the Child and Adult Care Food Program (CACFP) report (IOM, 2011), from program participation to program quality, calls for an equally broad range of methodologies for designing a nationally representative study to assess dietary intake and program participation in child day care. As elaborated throughout this summary, many lessons have been learned from past experience about what data to collect and how to collect and interpret those data. Still, there are many unanswered questions about how to move forward. For example, what kind of baseline data should be collected, before CACFP meal pattern recommendations are implemented? How should the broad range of research areas be prioritized? Which data are the most important to collect as part of an ongoing program evaluation of CACFP? What are the values of a longitudinal survey versus a set of cross-sectional surveys? How can state-level data be used? Given that there is no single child care system, rather a variety of types of child care systems (e.g., center versus home-based settings), will researchers need to employ different methodologies for those different settings? Finally, are there any additional research areas not covered during the workshop that should be included in a nationally representative study? These questions were explored further in the final session of the workshop, an open discussion among all participants moderated by Suzanne Murphy (see Box 5-1 for the actual list of questions that was used to guide the discussion). This chapter summarizes that discussion.

Some questions were prepared by committee members of the CACFP study (IOM, 2011) and some by the U.S. Department of Agriculture (USDA) Food and Nutrition Service (FNS). As previously explained, the legislative



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 65
5 Moving Forward T he broad range of research areas laid out in the Child and Adult Care Food Program (CACFP) report (IOM, 2011), from program participation to program quality, calls for an equally broad range of methodologies for designing a nationally representative study to assess dietary intake and program participation in child day care. As elaborated throughout this summary, many lessons have been learned from past experi- ence about what data to collect and how to collect and interpret those data. Still, there are many unanswered questions about how to move forward. For example, what kind of baseline data should be collected, before CACFP meal pattern recommendations are implemented? How should the broad range of research areas be prioritized? Which data are the most important to collect as part of an ongoing program evaluation of CACFP? What are the values of a longitudinal survey versus a set of cross-sectional surveys? How can state-level data be used? Given that there is no single child care system, rather a variety of types of child care systems (e.g., center versus home-based settings), will researchers need to employ different methodolo- gies for those different settings? Finally, are there any additional research areas not covered during the workshop that should be included in a nation- ally representative study? These questions were explored further in the final session of the workshop, an open discussion among all participants moder- ated by Suzanne Murphy (see Box 5-1 for the actual list of questions that was used to guide the discussion). This chapter summarizes that discussion. Some questions were prepared by committee members of the CACFP study (IOM, 2011) and some by the U.S. Department of Agriculture (USDA) Food and Nutrition Service (FNS). As previously explained, the legislative 65

OCR for page 65
66 RESEARCH METHODS TO ASSESS DIETARY INTAKE BOX 5-1 Questions Addressed During the Closing Open Discussion 1. What baseline measures are particularly needed to track the impact of change to CACFP, such as those that were recommended in the recent Institute of Medicine report? 2. Since funding will be unavailable to do all of the proposed research at one time, what is a reasonable ordered approach for assessing these topics? Pri- oritize the research needs, including the importance of data collection at the individual (child) level, the provider level, and the state or national level. 3. What are the best survey designs to gather nationally representative data, and to allow an evaluation of trends? 4. Are data collected by states able to be used to compile nationally representa- tive data, replacing the need to conduct a survey? What types? 5. Are there different evaluation measures that should be considered for CACFP family day care homes versus child care centers? 6. Are there measures that have not been discussed today that should be con- sidered in evaluating CACFP (other than physical activity measures)? language in the Healthy, Hunger-Free Kids Act of 2010 very closely aligns with some of the research recommendations laid out in the CACFP report. The act authorizes funding for a study of the nutrition and wellness quality in all child care settings, including but not limited to CACFP programs, and provides USDA with $5 million for conducting such research. According to Jay Hirschman, USDA will give precedence to research proposals that address items in the act on which FNS must report back to Congress. It is discretionary as to how far and in depth they will delve into each of the areas mandated by the legislation and how far and in depth they will go into other areas not explicitly included in the legislative directive. PUBLIC COMMENTS Before the open discussion began, Jennifer Weber of Nemours, a foundation that operates a children’s health system, offered some public comments based on the population-based prevention component of their mission. (As part of its prevention mission, Nemours worked with CACFP

OCR for page 65
67 MOVING FORWARD in Delaware to adopt new state best practice standards and policies.) First, maintaining and investing in regular assessments of the population is im- portant so that population-level changes can be monitored. In the case of CACFP, ongoing assessment will be critical to evaluating the role that CACFP participation plays in children’s dietary intake. According to We- ber, Nemours encourages the collection of baseline data before the recom- mended CACFP meal standards are implemented. Second, certainly there is a need for data on the food and nutrient content of meals and snacks served in CACFP homes and centers, as well as participants’ overall food and nutrient intakes and their alignment with current dietary guidance. In addition, Nemours also encourages the collection of information on provider characteristics and other factors that might influence the impact of CACFP on children’s nutritional well-being. As examples, Weber listed caregivers’ knowledge, awareness, attitudes, beliefs about changes to the nutrition standards, and beliefs about nutrition standards overall. Third, Nemours encourages the collection of data that can assess systems-level changes (i.e., policy, program, and practice changes) that, in turn, could lead to changes in population health outcomes. As examples of this type of data, Weber listed the number of child care providers who implement new standards and the speed at which they do so. QUESTION 1 What baseline measures are particularly needed to track the impact of change to CACFP, such as those that were recommended in the recent Institute of Medicine (IOM) report? The CACFP report (IOM, 2011) recommended substantial changes to CACFP meal requirements in an effort to bring the requirements into align- ment with the best available dietary guidance and to improve consistency with the requirements of other USDA food assistance programs. The report also recommended steps for ongoing evaluation and periodic reassessment to determine the impact of the meal requirement changes on participants’ nutritional well-being. As part of that process, the report recommended that “USDA should take appropriate actions to establish the current baselines prior to implementation of the new Meal Requirements for comparison purposes” (Program Evaluation Recommendation 2, IOM, 2011). Thus, the question was raised: Which of the measures discussed throughout the course of the workshop would be especially informative to collect before the meal requirement changes are implemented?

OCR for page 65
68 RESEARCH METHODS TO ASSESS DIETARY INTAKE General Observations Participants made some general observations and suggestions for mov- ing forward on gathering baseline information. One participant suggested that one way to “organize thoughts about these issues” is, first, to cat- egorize the issues as pertaining to either provider characteristics, provider measures (i.e., what providers are offering to the children), children’s ex- periences (i.e., what the children actually eat), or parents’ expectations, and then to decide the minimum amount of data needed for each category in order to assess the impact of the meal requirement changes. Following discussion, others seemed to be of the same mind on the need for provider data (both provider characteristics and measures) and children’s experi- ence data (i.e., what children are actually eating) in particular, not just at baseline but also as part of any ongoing assessment of CACFP. Another participant added that another category of data to consider is that at the state level (e.g., data from state agencies involved with CACFP), especially with respect to the cost of the changes (e.g., how many additional staff and resources are redirected toward CACFP activities). Gina Adams touched on what emerged as a major overarching theme of the workshop discussion: the importance of keeping the end in mind and collecting data based on desired outcomes. According to Adams, research- ers have conducted good qualitative work in the past on many relevant issues. Adams reminded the workshop audience of this work and suggested studying it to get a better sense of the types of effects to expect (i.e., after the changes are implemented) and the type of data to collect in anticipation of those effects. Measuring Cost Fred Glantz emphasized that, at a minimum, baseline data should be collected on foods being served to children. Next would be data on what children are actually eating and whether consumption is in care or outside of care. Next would be data on cost and participation. On cost, he said, “Providers have long argued that it costs more to serve high-quality meals. There is no information on that one way or the other. And if costs go up, and if a school district or child care provider has to raise rates, what does that do to participation?” Glantz’s remarks led to a more in-depth discus- sion of cost (not just to the provider but also to the family and the state) and how to measure the cost of implementing the recommended changes, with several participants suggesting that cost is an important measure to track over time but that it is not an easy measure to track. Suzanne Murphy identified fluctuating food prices as one challenge to measuring cost. Throughout any evaluation, food prices would need

OCR for page 65
69 MOVING FORWARD to be adjusted back to baseline. An unidentified audience member men- tioned aggregated data as another challenge. When new nutrition standards went into effect in Delaware, aggregated state-level data from participat- ing providers indicated that overall cost went up. However, even then, the estimate was based on “rough numbers” from the providers, with no child-level data available. Yet another challenge is “hidden” costs, such as donated space and services. Glantz mentioned that the National Day Care study conducted in the mid-1970s found that donated services and, more importantly, donated space and how that space was valued were huge is- sues. Many church-based centers used their basements and therefore had far more space available for child care services. How does one value that space? If the centers had to purchase the donated space, how much would it cost? The same is true of donated labor. How does one value it? Another audience member remarked that donated labor will be particularly impor- tant to track when the CACFP meal requirement changes go into effect, given that implementing the recommended changes will probably require additional labor. The recommended CACFP meal requirement changes are expected to impact not just the cost to providers but also the cost to states in the form of training and other activities that will need to be implemented in order to oversee the requested changes. One audience member urged, therefore, that baseline data also be collected on all costs. With respect to existing methodologies for measuring cost, Murphy referred to the Monsivais and Johnson (2012) study, which Lorrene Ritchie mentioned during her presentation, where investigators calculated food expenditures by matching food receipts with menus. Murphy asked, is it possible to do what Monsivais and Johnson (2012) did on a national level? Are there other methods? Monica Rohacek mentioned that the Study of Cost, Quality and Child Outcomes in Child Care Centers, a cost-quality child outcome study from the mid-1990s, collected and compiled cost data (Helburn, 1995). QUESTION 2 Since funding will be unavailable to do all of the proposed research at one time, what is a reasonable ordered approach for assessing these topics? Prioritize the research needs, including the importance of data collection at the individual (child) level, the provider level, and the state or national level. During its deliberations, the expert committee that prepared the CACFP report identified several significant gaps in knowledge about CACFP (IOM, 2011). Thus, they recommended steps for targeted research along with an assessment of the impact of the recommended meal requirement changes.

OCR for page 65
70 RESEARCH METHODS TO ASSESS DIETARY INTAKE Together, the research and evaluation recommendations make for a very ambitious agenda. Yet, the Healthy, Hunger-Free Kids Act of 2010 au- thorizes only $5 million for FNS-supported research on child care. This raises the question: How should the very ambitious research agenda laid out in the CACFP report and explored in detail during this workshop be prioritized? Suzanne Murphy questioned whether data should be collected only during day care (the “provider day”) or throughout the day (“full day”). Her interpretation of the legislative language is that the Healthy, Hunger- Free Kids Act of 2010 calls for an evaluation of the quality of foods being served, possibly consumed, during the provider day. However, the CACFP report is more expansive in its scope (IOM, 2011). According to Murphy, one of the greater goals of the recommended changes in the report is to create a healthier food environment during the provider day that would carry over into the full day. That is, children would become accustomed to eating fruits, vegetables, and whole grains and would continue to eat those foods after leaving child care. She said, “I would be sad not to see that somehow measured.” But with respect to using the $5 million in research funds being provided by Congress as part of the Healthy, Hunger-Free Kids Act of 2010, Murphy said that collecting data on the full day is “probably not the top priority.” Others concurred that gathering child-level data on foods served and/ or consumed both in and outside of care would be ideal but that $5 mil- lion will not cover such an ambitious research agenda. Lynne Oudekerk remarked that, from a state agency perspective, while it would be interest- ing to know whether and how the changes impact what happens outside of CACFP, it will be much more important to know whether and how the changes impact the food environment inside CACFP. Are the changes mak- ing a difference in what is served in CACFP? She asked, “Are we creating a positive food environment so that those 3.4 million children, and the few adults as well, are actually being exposed to healthier foods?” Another audience member predicted that it would be too challenging to correlate changes in CACFP meal requirements with such a distal outcome (i.e., creation of a healthier eating environment for the full day) and agreed that a better use of funds would be to focus on more proximal outcomes (e.g., creation of healthier eating environment during the provider day). Yet another decision that needs to be made is whether data are to be collected on foods served, foods consumed, or both. Workshop participants expressed varying opinions. Oudekerk remarked that while it would be interesting to know what is consumed, the more relevant data from a state agency perspective would be foods served. Beth Dixon agreed that assess- ing what is served is an important component of evaluating the impact of the proposed changes on CACFP itself, but asserted that consumption data

OCR for page 65
71 MOVING FORWARD are critical if the goal is to assess not just the health of the program but also the health of the children in that program. Based on a study of 110 child care centers in New York City, she and colleagues have shown that children eat only about two-thirds of what is served to them, and even less than two-thirds of the “difficult foods” (i.e., whole grains and vegetables). Dixon noted that, from a cost perspective, even an expensive method like direct observation could be used at marginal cost to collect additional consumption data if observers are already in the setting collecting data on foods served. While most of the discussion was focused on measuring foods either served or consumed by children already participating in CACFP, Gina Adams reflected on the value of also measuring participation in CACFP itself. If the ultimate public health goal is to improve the overall nutrition of low-income children, then participation matters. Regardless of the type of data collected (e.g., provider day versus full day, foods served versus foods consumed, and participation), Fred Glantz suggested that one of the first decisions to be made is whether the research will be a descriptive study of CACFP programs or a comparative study between CACFP and eligible but nonparticipating child care programs. QUESTION 3 What are the best study designs to gather nationally representative data, and to allow an evaluation of trends? The Healthy, Hunger-Free Kids Act of 2010 calls for a “nationally rep- resentative study of child care centers and family or group day care homes.” But what type of study? Can researchers use multiple cross-sectional sur- veys to measure trends, or can (should) they track the same set of children over time via one large longitudinal study? Fred Glantz remarked that it was not clear how one could track the same set of children over time, given that children eat differently as they age. He suggested multiple cross-sectional studies across similar age distri- butions. Rupa Datta agreed that a longitudinal methodology would not be helpful for evaluating child-level changes but might be helpful for assessing provider-level changes. For example, which providers are implementing the new standards? How quickly are they implementing them? How many providers leave or enter the program after changes are put into place? In addition to discussing the advantages and disadvantages of a longi- tudinal versus cross-sectional survey design, participants also discussed the practicality of nationally representative data. With respect to provider-level data, Datta wondered if a nationally representative sample would be too limiting. Specifically, she questioned whether there might be a way to focus

OCR for page 65
72 RESEARCH METHODS TO ASSESS DIETARY INTAKE on providers that serve low-income children (or providers located in low- income areas) rather than on providers that serve the general population at large. Not using a nationally representative sample of the entire general population would, in her words, “really free up money” for collecting data on segments of the population that would be more “analytically useful.” When asked to clarify what she meant by “analytically useful,” Datta explained that National Survey of Early Childhood Care and Education (NSECE) is much more interested in children from low-income households than children from households at other income levels. Yet, when conduct- ing their nationally representative study, they are collecting data on about two and a half times as many non-low-income children as low-income chil- dren. The money spent on data collection among so many non-low-income children prohibits doing additional analyses on low-income children (e.g., examining differences between rural and urban populations, between two household earners and households with only one earner, or between blacks and Latinos). When Jay Hirschman expressed concern that CACFP is an entitlement program for any child in the United States who participates in a CACFP site and that the ultimate goal is to improve the diets of children in general, Datta replied that sampling from a nationally representative pool will present a challenge to getting enough CACFP providers and partici- pants for a meaningful analysis. QUESTION 4 Are data collected by states able to be used to compile nationally representative data, replacing the need to conduct a survey? What types? A recurrent theme throughout the day was the availability of state-level data on CACFP. For example, during the second session on barriers and facilitators to providing meals and snacks that align with current dietary guidelines, it was suggested that some of the states currently implementing new nutrition guidelines are also collecting preliminary data on barriers and facilitators that could help frame questions for a nationwide study. The idea that state-level data could inform a nationally representative study of CACFP was revisited here. Lynne Oudekerk suggested that USDA query states to find out what data are being collected and what analyses have been conducted. She noted that state agencies accumulate a great deal of unpublished data, some of which might be of merit. As one example, New York has been collecting breast-feeding data that might be helpful when developing survey questions aimed at evaluating CACFP practices around breast feeding. But could state-level data be used to do more than help guide design of a national study? Could they actually replace nationally representative data?

OCR for page 65
73 MOVING FORWARD Those who spoke expressed a similar view that state data cannot replace nationally representative data. However, state data could supple- ment national data. Joanne Guthrie commented that state-level CACFP data might enable in-depth investigations not possible with national data because of the way data are aggregated after they are submitted to USDA (e.g., rural versus urban trends). Another audience member suggested lever- aging non-CACFP state databases. For example, there might be opportuni- ties to link CACFP data with Supplemental Nutrition Program for Women, Infants, and Children (WIC) data. Virginia Stallings wondered whether any states collect data on service delivery or good business practices. Oudekerk replied that yes, some states collect those types of data. For example, New York researchers are collecting data that will help them address reimburse- ment bottlenecks (e.g., payment processing). She said that different states are probably approaching the research in different ways. Several participants commented on the fact that, regardless of the type of data, a key challenge with state data is that each state collects, stores, and analyzes data differently. Guthrie said, accessing state data is “not as easy as it sounds.” This is especially true of data on eligible nonparticipants. Glantz remarked that, while state-level data may be obtainable on CACFP participants, data on nonparticipants is not. Finally, Oudekerk wondered whether a nationally representative study of CACFP would prompt USDA to collect more raw data from states in the future. She opined that, with advances in information technology, there is no reason for states to continue aggregating data before submission. QUESTION 5 Are there different evaluation measures that should be considered for CACFP family day care homes versus child care centers? As Virginia Stallings pointed out in her keynote address, a major chal- lenge to gathering and analyzing CACFP data is variation in setting (e.g., child care center versus family day care home). Variation in setting raises the question: Do different settings require different evaluation measures? Several participants voiced a similar opinion that different settings do require different evaluation measures. The goal should be the same (e.g., assessing the impact of the meal requirement changes), but reaching that goal requires different strategies in different settings. With some issues, variation in setting requires asking different types of questions. For example, one workshop participant commented that the cost (and labor) associated with purchasing meals from a third-party source, which is often how Head Start and other child care centers get their meals, is “more straightforward” than the cost (and labor) of making meals in

OCR for page 65
74 RESEARCH METHODS TO ASSESS DIETARY INTAKE house, which is what family day homes typically do. Understanding the latter requires asking different questions. As another example, with respect to participation trends, Monica Rohacek pointed out that different factors affect both the desire to participate and the nature of participation. For example, family day care homes and child care centers may have different training mechanisms or administrative structures that impact participa- tion. Those “different contextual factors” need to be accounted for in a questionnaire. Gina Adams observed that the respondent pool is different in the differ- ent settings and that the questions need to be “oriented” accordingly. For example, a child care center may have one person responsible for money, another person responsible for nutrition, and so on, whereas in a family day care home usually only one person (i.e., the provider) is responsible for everything. In addition to asking different types of questions, many questions need to be worded differently. One audience member commented on the greater difficulty associated with gathering information from family day care homes operated by non-English-speaking providers with low English literacy skills. The provider may understand “play,” but not “active play” or “struc- tured play.” During development of a California statewide survey, Lorrene Ritchie and colleagues hired a child care consultant to reword some of the questions so that their survey tool, which had been developed for use in child care centers, could be used in family day care homes. The redesigned questionnaire was much longer than the original because it included more definitions and examples. The investigators piloted the survey in day care homes to make sure that it had been reworded appropriately. Finally, Angela Odoms-Young pointed out that with some measures, such as barriers and facilitators to implementation, there have not even been enough data collected yet to know whether measures used to assess barriers to implementation in one setting are relevant to another setting. She encouraged more in-depth case studies of family day care homes, per- haps using state-level data, as a way to gather some of those data. Variation in setting calls not only for different types of survey tools, but also for different comparison groups. One workshop participant opined that the comparison group for use in a survey of child care centers (i.e., eligible but nonparticipating child care centers) would be different than a comparison group for use in a survey of family day care homes (i.e., eligible but nonparticipating family day care homes). The latter would probably contain more unlicensed providers, creating more challenges. Lor- rene Ritchie cautioned that, with both settings, the comparison group will probably need to be oversampled. Her experience has shown that response rates are very similar for CACFP child care centers and family day care

OCR for page 65
75 MOVING FORWARD homes and that the lowest response rates were among nonparticipating centers and homes. QUESTION 6 Are there measures that have not been discussed today that should be considered in evaluating CACFP (other than physical activity measures)? Participants listed four additional measures: (1) food environment, such as the physical structure of the kitchen (e.g., What is in the kitchen? What is the water availability?), staff behavior (e.g., Are the staff eating with the children? Are they encouraging vegetables?), and the appropriate- ness of the facilities for children (e.g., Are the table and utensils appropri- ate sizes for children?); (2) purchasing of food (e.g., Are providers buying prepared food? Where are they shopping?); (3) Cost to sponsors and states (i.e., What is the cost of implementing the recommended meal requirement changes?); and (4) providers’ experiences with implementation (e.g., How much paperwork is required? What other “burdens” does implementation impose?). FINAL REMARKS This 1-day workshop was a continuation of the review of the CACFP meal requirements carried out by an IOM expert committee, which re- leased the consensus report Child and Adult Care Food Program: Aligning Dietary Guidance for All in October 2010 (IOM, 2011). The workshop agenda was based on the research recommendations in that report. Specifi- cally, workshop participants addressed how to assess (1) the alignment of young children’s dietary intake with current dietary guidance, (2) barriers and facilitators to providing foods that align with dietary guidance, and (3) program access and participation trends. Workshop participants covered a very broad range of topics within each of these categories. Drawing on lessons learned from past research experience, participants explored both what types of data to collect and how to collect those data. Overarching Themes of the Workshop Discussion Workshop participants revisited several major, cross-cutting themes over the course of the day’s various discussions: • Gaps in knowledge. As Virginia Stallings made clear in her keynote address and as elaborated in the CACFP report (IOM 2011), there is a considerable lack of up-to-date data on food and nutrient

OCR for page 65
76 RESEARCH METHODS TO ASSESS DIETARY INTAKE intake among children attending CACFP sites, how intake from foods and meals served in CACFP settings contributes to overall dietary intake, and how both in-care and overall dietary intakes align with current dietary guidance. As Lorrene Ritchie remarked, even less is known about the barriers and facilitators to providing meals and snacks that align with current dietary guidance. Equally important are gaps in knowledge about participation in CACFP and incentives (and disincentives) for providers to become involved in the program (e.g., the cost of participation). Indeed, a major reason for holding this workshop was to explore these many gaps in knowledge and identify which gaps need to be filled first. • The need to be very clear about the key research question(s) be- ing asked and the type of answer(s) being sought. Regardless of whether the issue is dietary intake, barriers and facilitators to pro- viding healthy foods, or program participation, Beth Dixon, along with speakers from all three sessions, emphasized clearly that iden- tifying the question(s) being asked and the desired outcome(s) is key to adapting existing methodologies for use in studying CACFP. For example, with respect to dietary intake, is the goal to assess whether the implemented changes are having the desired effect (i.e., aligning foods served with current dietary guidance)? Or, is the goal to examine the contribution of CACFP meals and snacks to overall nutrient intake? Or, is the goal to compare nutrient in- take among CACFP participants versus non-CACFP participants? Different questions require different types of data and different methodologies for collecting and analyzing those data. As Sara Benjamin Neelon and others explained, while direct observation may be the preferred method for collecting data during child care (because it provides the most specific and accurate information), the 24-hour dietary recall is widely considered the “gold standard” for collecting data outside of child care. • Existing methodologies used in past studies may be relevant, al- though the methodologies need to be adapted for CACFP setting(s). Again, all speakers in all three sessions touched on this major theme. For example, Mary Kay Fox discussed the potential rel- evancy of dietary intake methodologies used in two previous large national studies, the School Nutrition Dietary Assessment and Feeding Infants and Toddlers Studies. The tools employed in both of those studies could be used in a study of CACFP, but they would need to be “translated.” Dianne Ward explored methodologies described in various published studies for assessing foods served in child care settings. She concluded that there is no single best meth- odology and suggested that protocols be tested before wide-scale

OCR for page 65
77 MOVING FORWARD implementation. In their exploration of factors that shape provider participation, Gina Adams and Monica Rohacek drew on lessons learned by Urban Institute researchers about what kind of data to collect and how to collect it (e.g., the value of qualitative versus quantitative experimental designs). As a final example, Rupa Datta considered how methodologies employed in the NSECE might be useful in a nationally representative study of CACFP. • There is no single child care system. As Jay Hirschman pointed out, CACFP serves various types of child care centers and family day care homes. Stallings identified variation in setting as a major challenge to collecting and analyzing CACFP data. Differences between child care centers and family day care homes, not to men- tion differences between licensed and unlicensed family day care homes, cut across the three main areas of research addressed by the workshop (i.e., food and nutrient intake, barriers and facilitators to providing healthy meals and snacks, program access and participa- tion trends). Based on the discussion that took place in response to Question 5 (see previous section), the differences impact not just the type of questions that need to be asked, but also how those questions are asked and of whom they are asked. Angela Odoms- Young remarked that with some issues, for example, barriers and facilitators to implementing programmatic change, there is such a lack of data for family day care homes in particular that it is not clear whether and how the questions need to be adjusted for the different settings. She encouraged more in-depth studies of family day care homes. • Existing databases are a “gold mine” of information. Workshop participants laid out what amounts to a very ambitious research agenda. Rupa Datta suggested that one way to tackle the agenda is to exploit as much as possible existing administrative and house- hold data before deciding what type of original data to collect. For example, Fred Glantz elaborated on the wealth of administrative CACFP data that is collected at the provider level but becomes inaccessible to researchers because of the way those data are ag- gregated after providers send in their monthly reimbursement claim forms. If those data could be accessed, they would be very infor- mative. As another example, Susan Jekielek discussed the potential relevance of other early childhood program administrative data collected by the Administration for Children and Families (e.g., Head Start and Child Care Subsidy Program data). Plus, there were many calls throughout the workshop for a greater consideration of state data (e.g., see Question 4 above). State data cannot replace nationally representative data, but they might inform design of a

OCR for page 65
78 RESEARCH METHODS TO ASSESS DIETARY INTAKE nationally representative survey (e.g., how to frame questions) and could supplement such a survey. Next Steps for the FNS: How Can We Do a Better Job Feeding Our Nation’s Children? USDA will be issuing a Request for Proposals (RFP) in order to carry out the research commissioned by the Healthy, Hunger-Free Kids Act of 2010. The agency will use the information presented during this workshop to guide its development of the RFP and decision making about which questions to address first and which methodologies might be helpful for addressing those questions. Hirschman encouraged workshop participants and other experts in the field to become involved, if not by way of proposal submission then perhaps by helping contractors prepare their proposals, serving on an advisory panel to the contractor (FNS studies have advisory panels that provide advice to the researchers throughout the course of the study), or serving as a reviewer of interim deliverables (e.g., reviewing survey questionnaires before they are distributed). Improving child nutrition, while also combating overweight and obesity, poor nutritional habits, and the high levels of food insecurity and hunger in America, is, in Hirschman’s words, “something that we all have to work on.” REFERENCES IOM (Institute of Medicine). 2011. Child and Adult Care Food Program: Aligning dietary guidance for all. Washington, DC: The National Academies Press. Helburn, S. W. 1995. Cost, quality and child outcomes in child care centers. Denver: University of Colorado. Monsivais, P., and D. B. Johnson. 2012. Improving nutrition in home child care: Are food costs a barrier? Public Health Nutrition 15(2):370-376.