Development and Testing of Measures to Assess Nutrition Behavior Change in Low Income Adults Participating in the Expanded Food and Nutrition Education Program

• Main Author: Murray, Erin K.
• Language: EN
• Published: Colorado State University 2017
• Subjects: Nutrition
• Online Access: http://pqdtopen.proquest.com/#viewpdf?dispub=10267178

<p> <b>Purpose</b></p><p> The primary purpose of this research was to develop measures of diet quality and nutrition behavior to evaluate the effectiveness of the Expanded Food and Nutrition Education Program (EFNEP). EFNEP is a United States Department of Agriculture nutrition education program that works with low-income families to improve their diet quality and food-related behaviors. The research objective was to develop a reliable and valid short dietary assessment instrument that can be used nationally by EFNEP to assess diet quality behavior change among low-income adults. Short dietary assessment instruments provide cost-effective ways to evaluate federally-funded nutrition education programs and the low-income adults they serve. However, few valid instruments exist for use nationally with low-income populations.</p><p> <b>Methods</b></p><p> This research was part of a multi-state, multi-year Agricultural Experiment Station research project, NC2169: EFNEP-Related Research, Program Evaluation and Outreach. Goals of the multi-state project included developing a new EFNEP national evaluation questionnaire that would include diet quality questions. </p><p> A mixed methods observational study design, including 8 phases, was used to develop and test the dietary assessment questions. Phases 1-5 pertain to question development, and phases 6-8 relate to question testing. The research phases are listed below. 1. <b>Nutrition education curricula content analysis. </b> The content analysis determined current nutrition information taught nationally to EFNEP participants. Curricula used by the majority of EFNEP state programs served as a proxy for nutrition content taught to EFNEP participants. A tested data recording instrument captured nutrition content in the curricula. 2. <b>Dietary Guidelines expert panel (expert panel #1).</b> A panel of 6 national nutrition experts identified and prioritized the nutrition recommendations from the 2010 Dietary Guidelines for Americans deemed most critical to teach low-income adults. The study protocol used a modified Delphi Technique to build consensus among experts through structured feedback. 3. <b>EFNEP expert panel (expert panel #2).</b> Twenty-one EFNEP program administrators and other researchers from 15 states further prioritized the nutrition recommendations identified by the first expert panel to those most critical to evaluate in EFNEP. 4. <b>Literature review and question generation.</b> A literature review of published manuscripts and government and research organization websites identified validated questions that addressed the nutrition recommendations deemed critical to evaluate. Questions and response options were revised, eliminated or created to meet the needs of the EFNEP population and program. 5. <b>Content validity expert panel (expert panel #3).</b> Seven EFNEP program administrators from different US geographic regions reviewed the dietary assessment questions and response options to confirm content validity for use in a national EFNEP questionnaire. Questions and response options were revised as necessary. 6. <b>Cognitive interview testing.</b> Researchers conducted cognitive interviews with EFNEP participants (at program enrollment &ldquo;pre&rdquo; or at program completion &ldquo;post&rdquo;) in 7 states. An iterative process and scripted probing questions were used to determine ease of understanding and face validity. Interviewers in each state were trained via webinar. Interviews were audio-recorded, transcribed and analyzed pre/post to identify themes and dominant trends. Questions and response options were revised as needed. 7. <b>Reliability testing.</b> The test-retest method assessed temporal stability reliability of the questions. A total of 217 low income EFNEP-eligible women from 7 states completed the dietary assessment instrument twice, at a 1-month interval. Paired t-tests (p &le; 0.05), Spearman&rsquo;s rank order correlation coefficients (SCC) and intraclass correlation coefficients (ICC) were used to assess reliability (r > 0.5). Exploratory factor analysis was used to determine whether the questions grouped together (factor loading cut-offs > 0.5). 8. <b>Construct validity testing.</b> A total of 60 EFNEP participants were recruited from 8 states to complete the dietary assessment instrument and 3, 24-hour food recalls pre/post the EFNEP intervention. Wilcoxon signed rank test (p &ge; 0.05), SCC (r > 0.5), and Bland-Altman plots were used to assess construct validity of the questions.</p><p> <b>Results</b></p><p> <b>Phases 1-5.</b> The curricula content analysis findings revealed considerable variability in both the frequency of certain nutrition content and depth of educational instruction provided across curricula used in EFNEP. The Dietary Guidelines expert panel determined 2 overarching and 8 specific nutrition messages critical to teach low income adults.</p><p> <i>Overarching recommendations</i> &bull; Focus on nutrient-dense foods, including vegetables, fruits, whole grains, fat-free or low-fat milk and milk products, seafood and fish, lean meats, poultry, eggs, beans and peas, and nuts and seeds. &bull; Cook and eat more meals at home. The behaviors of portion control and how to prepare lower calorie options can be emphasized within teaching how to prepare meals at home.</p><p> <i>Specific recommendations</i> &bull; Increase vegetable and fruit intake. Eat a variety of vegetables. &bull; Consume at least half of all grains as whole grains. &bull; Increase intake of low-fat dairy or fortified soy products. &bull; Eat a variety of protein foods, including beans, legumes, nuts and seeds, eggs, seafood, and lean meats and poultry. &bull; Develop skills in reading the Nutrition Facts panel to identify portion size and calorie intake for packaged foods. &bull; Prepare, serve, and consume smaller portions of food and beverages. &bull; Reduce intake of foods such as chips and crackers and sweets such as cookies, cakes, pie, muffins, doughnuts, and pastries. &bull; Limit the intake of sugar-sweetened beverages.</p><p> Using the Dietary Guidelines panel findings, the EFNEP expert panel prioritized 6 diet quality content areas to evaluate in EFNEP.</p><p> <i>Nutrition behaviors to evaluate in EFNEP</i> &bull; Cook and eat more meals at home. &bull; Increase fruit intake. &bull; Increase vegetable intake. &bull; Eat a wider variety of vegetables. &bull; Increase intake of dairy of fortified soy products. &bull; Limit the intake of sugar-sweetened beverages.</p><p> The literature review produced an average of 40 questions (range 19-71) per diet quality content area and 46 response scales. Questions were eliminated or revised to yield 22 questions with 1-2 response scales per question. Content validity expert panel feedback resulted in 20 revised questions for cognitive interview testing.</p><p> <b>Phases 6-8.</b> A total of 111 cognitive interviews in 3 rounds of interviews/question revisions were completed with EFNEP participants in 7 states. Cognitive interviews yielded a 14-item dietary assessment instrument covering the 6 nutrition content areas. Reliability testing results showed all questions had at the least moderate correlations (SCC > 0.40) and fair agreement (ICC > 0.41), with at least half the questions indicating strong (SCC > 0.60) and moderate (ICC > 0.61) correlations (<i>P</i> &lt; 0.001). The majority of questions (12/14) grouped together to align with the 6 nutrition content areas to evaluate in EFNEP (factor loadings >0.50).</p><p> For construct validity testing, data were collected from EFNEP participants in 8 states (n = 60 pre, 30 post). Food recall data were collected an average of 14 days after participants completed the instrument (range 6-32 days). Results demonstrated significant differences (Wilcoxon signed rank test p = &lt; 0.05) and low or no correlations both pre (SCC = 0.01 &ndash; 0.44) and post (SCC = 0.01 &ndash; 0.44) between the instrument and 24-hour food recall data. The Bland-Altman method was not pursued due to the lack of significant correlations. The researchers concluded that incompatible methods along with measurement error from multiple sources contributed to the lack of association between the instrument and 24-hour food recall data.</p><p> <b>Conclusions and Implications</b></p><p> This research resulted in nationally-tested reliable and valid dietary assessment instrument for low-income adults that can be used to evaluate EFNEP&rsquo;s effectiveness at improving diet quality among participants. The mixed-methods approach established temporal stability reliability and content, face, and factor validity of the dietary assessment instrument. Questions were developed to meet program objectives and tested with the target EFNEP population in multiple states, which confirmed their appropriateness for evaluating behavior change for this national nutrition education program. This research has implications that extend beyond EFNEP, as other nutrition education programs serving low-income adults may adopt the methods used to develop their own validated evaluation questionnaire. Nutrition education programs or interventions may also adopt the validated instrument to evaluate their programs.</p><p> Future research directions include testing the dietary assessment instrument questions for sensitivity to change and/or with a compatible comparison tool to establish construct or convergent validity. For example, a modified 7-day food record tailored to the needs of low-income participants and the behaviors captured in the instrument may be an appropriate comparison measure to assess convergent validity of the dietary assessment questions.</p><p>