Abstract
Background:
American Indian and Alaska Native (AI/AN) people have high rates of diabetes and limited access to nutrition education. The “What Can I Eat?” (WCIE) diabetes nutrition education program was culturally adapted for AI/AN adults.
Objective:
This analysis was designed to evaluate the reliability and validity of items developed to measure diabetes nutrition self-efficacy (i.e., confidence one can engage in specific behaviors) and diabetes nutrition behavior among participants in the WCIE program for AI/AN adults.
Design:
This study was a secondary analysis of data from a randomized controlled trial designed to evaluate the WCIE program for AI/AN adults. Baseline data were used to assess the reliability and validity of the self-efficacy and behavior items, which were collected via survey. Due to COVID-19 safety protocols, the intervention was conducted via Zoom, and both survey and clinical data were collected at home by participants.
Participants/setting:
The study was conducted from January to December 2021 with five AI/AN-serving health care programs in Oklahoma, Illinois, North Carolina, California, and New York. AI/AN adults with type 2 diabetes who spoke English and had Internet access were eligible. Sixty participated.
Main outcome measures:
Analyses examined validity and reliability of diabetes nutrition self-efficacy and behavior items.
Statistical analysis performed:
To test reliability, internal consistency and factor structures of the scales were examined. To evaluate convergent validity, Pearson correlations were computed to examine the association of the self-efficacy and behavior measures with each other and with clinical indicators (i.e., Body Mass Index, blood pressure, hemoglobin A1C).
Results:
Two self-efficacy factors were identified. Each showed strong internal consistency (Cronbach alphas ≥ 0.85; McDonald omegas ≥ 0.88) and was directly associated with diabetes nutrition behavior (Ps < 0.001). The factor assessing Confidence in Using the Diabetes Plate was inversely associated with A1C (Pearson Correlation = −0.32, P = 0.0243). The behavior measure capturing Healthy Nutrition Behavior showed strong internal consistency (alpha = 0.89; omega = 0.92) and was inversely associated with A1C (Pearson Correlation = − 0.38, P = 0.0057).
Conclusions:
Diabetes nutrition self-efficacy and behavior items developed for the WCIE program for AI/AN adults are valid and reliable. These items can facilitate rigorous and consistent evaluation of the AI/AN WCIE program.
Keywords: American Indians and Alaska Natives, Type 2 Diabetes Mellitus, Nutrition Education, Diabetes Self-Management Education and Support, Validation Study
American Indian and Alaska Native (AI/AN) people develop type 2 diabetes (T2D) at a higher rate than other racial and ethnic groups in the U.S. In 2019-2021, 13.6% of AI/AN adults were diagnosed with diabetes, compared to 6.9% of non-Hispanic White adults.1 Diabetes complications – including amputations, hypertension, heart disease, and kidney failure – are also common in AI/AN individuals.1,2
Although good nutrition is critical to maintaining healthy blood sugar levels,3,4 AI/AN populations often have poor access to healthy food. Traditional AI/AN food systems have been devastated by systemic inequities and colonialism.5-7 Because of these root causes, poverty, food insecurity, limited access to retail stores selling fresh fruits and vegetables, and unhealthy food environments persist for many AI/AN people.8-12 Likewise, although evidence suggests that nutrition education is associated with improved behavioral and clinical outcomes in diabetes,13-15 especially if programming is tailored to the cultural group of interest,16-20 AI/AN individuals have limited access to nutrition counseling and education,21 and few programs are designed for AI/AN audiences.22-24
To facilitate diabetes nutrition education in communities disproportionately impacted by T2D, the American Diabetes Association (ADA) developed the “What Can I Eat? Healthy Choices for People with Type 2 Diabetes” (WCIE) curriculum.25 WCIE is a classroom-based program taught by registered dietitian nutritionists (RDNs) who have been master-trained by RDNs from the ADA. WCIE includes five 90-minute sessions, each following an outline that includes didactic content, interactive learning activities, a physical activity, a mindful eating activity, and goal setting. Topics covered in the classes include the use of the Diabetes Plate26 to manage blood sugar, healthful shopping and cooking strategies, and healthy eating away from home. The curriculum was pilot tested in 2017-2019, showing beneficial effects on participants’ confidence that they can make healthy dietary choices and on their shopping and food preparation behaviors.25
In 2018, the Shakopee Mdewakanton Sioux Community of Minnesota partnered with the ADA to lead and fund the development of a culturally adapted version of WCIE, called “What Can I Eat? Healthy Choices for American Indians and Alaska Natives with Type 2 Diabetes” (AI/AN WCIE). Cultural adaptation of the curriculum was informed by extensive qualitative work, including focus groups with AI/AN adults with T2D and their family members as well as interviews with community leaders and clinicians in four AI/AN communities and with national experts on diabetes nutrition and AI/AN health.27-29 The AI/AN WCIE uses the same format and outline as the general WCIE program (e.g., five 90-minute classes) but incorporates content addressing the unique assets and challenges of AI/AN communities. For instance, the curriculum includes content on the valuable role that traditional Indigenous foods and food acquisition practices can play in a healthy diabetes diet. The adapted curriculum also includes original photography of AI/AN people, culturally relevant arts-based learning activities, and ample opportunities for peer-to-peer storytelling and sharing.
Given the vast diversity of Native communities across the country, the AI/AN WCIE curriculum was designed to allow local communities to tailor the program to their own needs. The curriculum includes “placeholders,” where class facilitators are asked to insert regionally relevant foods, food acquisition and preparation techniques, physical activities, or community-based resources. Likewise, the curriculum offers alternative activities and lessons from which facilitators can choose the content that best serves their communities. Finally, the program provides opportunities for participants themselves to select the nutrition topics they wish to discuss in peer-to-peer learning activities. These unique features of the AI/AN WCIE are designed to ensure that the curriculum can fit the needs of a wide range of AI/AN communities and settings.
The adapted curriculum was tested as part of a randomized controlled trial conducted in five diverse AI/AN communities.30 The trial examined the impact of the AI/AN WCIE program on participants’ confidence in their ability to make healthy nutrition choices (i.e., “self-efficacy”), their actual dietary behaviors, and their clinical outcomes.30 The objective of the secondary analysis reported here was to evaluate the reliability and validity of the diabetes nutrition self-efficacy and behavior items developed to evaluate the impact of the AI/AN WCIE program. The research team hypothesized that the self-efficacy and behavior items, respectively, would demonstrate internal consistency reliability. Informed by the Expanded Health Belief Model,31 the research team expected that participants with greater diabetes nutrition self-efficacy would engage in healthy dietary behaviors more frequently and have better clinical indicators than participants with lower levels of self-efficacy. The team also hypothesized that participants who engaged in healthy nutrition behaviors more frequently would have better clinical indicators.
Materials and Methods
Overview
To evaluate the reliability and validity of the self-efficacy and behavior measures, the research team conducted secondary analyses using baseline data from the randomized controlled trial entitled “Diabetes Nutrition Education for American Indian and Alaska Native Communities.” The trial was designed to evaluate the effectiveness of the AI/AN WCIE curriculum and was registered on ClinicalTrials.gov (Trial Registration Number NCT04254653). To assess reliability, analyses examining the internal consistency and factor structures of the self-efficacy and behavior items were conducted. To evaluate convergent validity, analyses were conducted to examine the association of the self-efficacy and behavior measures with each other and with clinical indicators with which they were expected to be associated (i.e., Body Mass Index [BMI], blood pressure, and hemoglobin A1C).
The Indian Health Care System
As a result of treaties between the U.S. government and tribes, congressional legislation, and court cases, the U.S. federal government has a trust responsibility, or legal obligation, that includes the provision of health care services to members of federally recognized tribes. Since 1955, the Indian Health Service (IHS), an agency within the Department of Health and Human Services, has had responsibility for providing those services, with the goal of improving the health status of AI/AN individuals.32 IHS provides direct patient care through a network of hospitals and clinics that are located on or near tribal land. As a result of congressional legislation intended to facilitate self-determination, tribes have assumed responsibility for the operation of many health care facilities and health care programs (with IHS funding). Urban Indian Health Organizations are clinical programs authorized through Title V of the Indian Health Care Improvement Act (Public Law 94-437) to serve the health care needs of AI/AN individuals who live in cities.33 Overall, the AI/AN health care system represents a complex network of facilities operated by IHS as well as tribal and urban organizations, collectively known as the I/T/U health care system.
Study Setting
The clinical trial that was the source of the data for this secondary analysis was conducted in partnership with five AI/AN-serving health care programs across the U.S. The research team sought to include sites that were geographically diverse and represented different components of the Indian health care system. Participating sites included three tribally run clinical programs in rural, reservation communities in California, New York, and North Carolina. These sites provide health care services mainly to members of the federally recognized tribes that call these reservations home. Two Urban Indian Health Organizations located in Oklahoma and Illinois, which serve the health care needs of Native people from a wide variety of tribal backgrounds, also participated. Although the team initially included an IHS health care facility in the Southwest, the recruitment and retention of an RDN to teach the classes as well as the devastating impact of COVID-19 on reservation communities34 disrupted the ability of this site to participate in the clinical trial.
Participant Eligibility & Recruitment
Study coordinators at the participating sites identified and recruited participants using their local diabetes registries. Eligibility criteria included the following: (1) self-reported AI/AN heritage, (2) age ≥ 18 years, (3) fluent in English, (4) current diagnosis of T2D, and (5) access to the Internet via a smartphone or computer. Potential participants were excluded if they planned to move out of the area during the study period or if they had medical conditions that could interfere with participation (e.g., cancer not in remission, receiving dialysis for end-stage renal disease, active substance use disorder).
Because the study was conducted during the COVID-19 pandemic (January to December 2021), the recruitment and consent processes were conducted remotely. Participants were screened for eligibility by phone and participated in a virtual informed consent visit, during which the local study coordinator described study procedures and reviewed the consent form. Interested individuals mailed a signed copy of the consent form to the local study coordinator. Participants provided written informed consent and HIPAA authorization before participation.
Across the five participating clinical sites, 83 people were screened for eligibility. Because study coordinators at the participating clinical sites only sought to enroll individuals whom they knew were eligible and interested in participation, there were no screen failures and all participants who were screened consented to participate. Twenty-three participants were lost to follow-up. Ultimately, 60 participants contributed baseline data to this analysis.
Study Design
This study was a secondary analysis of data from a randomized controlled trial that utilized a waitlist design.30,35 After providing informed consent, participants were randomized to the immediate intervention group or the waitlist group using a 1-to-1 randomization procedure implemented using the REDCap (Research Electronic Data Capture) randomization module.36 The immediate group started the intervention immediately following enrollment and randomization, completing Classes 1-4 weekly for the next four weeks and completing Class 5 three months after enrollment. The waitlist group started the intervention three months after enrollment and randomization, completing Classes 1-4 weekly thereafter and Class 5 six months after enrollment. To ensure the safety of study participants and staff during the pandemic, the intervention was conducted remotely using synchronous Zoom sessions to which participants connected via phone or computer. Classes were taught by an RDN from the local clinical site who had been trained by the research team’s lead RDN.
Study Approval
Approval for the study was obtained from the research review board of a participating tribe, the Indian Health Service (IHS) National Institutional Review Board, and the Colorado Multiple Institutional Review Board. Because of the pandemic, the research team was also required to secure pandemic-specific safety approvals from all relevant Institutional Review Boards and each clinical organization.
Data Collection
After enrollment, a “class kit” was mailed to participants’ homes. The kit included educational resources related to the intervention classes (e.g., class activity handouts, items for interactive arts-based learning and physical activities) as well as data collection tools and forms (i.e., body weight scale, home blood pressure monitor, home hemoglobin A1C test kits, printed copies of the AI/AN WCIE Impact Survey, and stamped return envelopes). Because of the pandemic, all data collection procedures had to be conducted remotely. Hence, the participants themselves collected all data using the measurement tools and paper surveys provided in the class kit.
The analyses reported here used baseline data. For each treatment arm, baseline data were collected just before participants began the AI/AN WCIE classes. For participants in the intervention group, the baseline time point coincided with the consent visit. For participants randomized to the waitlist group, baseline data were collected three months after enrollment, just before the beginning of AI/AN WCIE classes.
To collect the baseline data, the local study coordinator conducted virtual visits with each participant to collect clinical data, including height, weight, blood pressure, and A1C, as reported by participants. If the medical record included weight, blood pressure, and/or A1C values collected in the 30 days before the baseline visit, these were used as the baseline values instead. Likewise, when available, the most recent height measurement in the past 12 months was obtained from the medical record; if unavailable, the participant’s self-reported height was used. At baseline, participants also completed the AI/AN WCIE Impact Survey, which included items assessing sociodemographic characteristics, diabetes nutrition self-efficacy, and diabetes nutrition behavior. The survey was completed on paper, submitted by mail to the local study coordinator, and entered into a REDCap database by the study’s data manager.
Measures
Sociodemographic Characteristics.
The AI/AN WCIE Impact Survey included questions assessing sociodemographic characteristics, including age, gender (male, female, other), highest grade completed, total household income before taxes in the prior year, and employment status.
Diabetes Nutrition Self-efficacy.
The survey included 12 items assessing diabetes nutrition self-efficacy. Items were newly developed by the research team to assess participants’ confidence in their ability to engage in healthy diabetes nutrition behaviors. The research team identified key behaviors that the AI/AN WCIE curriculum is designed to encourage, developed items assessing confidence in these behaviors, and finalized items through an iterative review and revision process involving the research team and the study coordinators at the five participating sites. Participants were asked: “How confident are you that you can… [engage in specific activity]?” Eleven activities focused on diabetes nutrition behaviors, while one focused on physical activity. Although physical activity is not a dietary behavior, the AI/AN WCIE program highlights the role of physical activity in combination with healthy nutrition in the management of blood sugar. The self-efficacy items used a four-point scale: not at all confident (0), not very confident (1), somewhat confident (2), and very confident (3). Items did not undergo pilot testing prior to implementation in the randomized controlled trial.
Diabetes Nutrition Behavior.
The survey included 24 items measuring diabetes nutrition behaviors. Behavior items were newly developed by the research team to assess engagement in the healthy and unhealthy behaviors addressed by the AI/AN WCIE program. The team identified 24 specific behaviors that the AI/AN WCIE curriculum addresses, developed items assessing the frequency with which participants engaged in each of these behaviors, and finalized items through an iterative review and revision process involving the research team and study coordinators from the participating sites. Participants were asked: “During a typical week in the last month, how often did you… [specific behavior]?” Participants responded to each item using a six-point scale: more than once a day (6), once a day (5), 5-6 days a week (4), 3-4 days a week (3), 1-2 days a week (2), and never (1). A “don’t know” option was included. Responses of “don’t know” were set to missing before analysis. Items did not undergo pilot testing prior to implementation in the randomized controlled trial.
Clinical Indicators.
For each participant, BMI was computed and blood pressure and A1C were measured. BMI was calculated as weight in kilograms divided by height in meters squared. Systolic and diastolic blood pressure were analyzed separately.
Data Analysis
Analyses examined the reliability and validity of the diabetes nutrition self-efficacy and diabetes nutrition behavior items. Analyses used data from the baseline time point, which occurred just before the beginning of AI/AN WCIE classes for each treatment arm.
Descriptive Statistics
Descriptive analyses were conducted to examine participant characteristics and baseline responses for all measures. Means and standard deviations (SDs) were computed for continuous variables and frequencies (percentages) for categorical variables.
Reliability Analyses
To examine internal consistency reliability, the standardized Cronbach alpha and McDonald omega37 were computed for the self-efficacy and behavior scales. A Cronbach alpha and McDonald omega ≥ 0.70 was considered to reflect an acceptable degree of consistency among items.38
An exploratory factor analysis (EFA) was conducted to examine the internal consistency and factor structure of the self-efficacy scale. Following well-established methods, factors that had eigenvalues ≥ 1.0 before rotation and that appeared from the scree plot to be before the inflect point were retained. Varimax rotation was used to achieve a simpler factor structure that maximized high loadings and minimized low loadings. Factor loadings ≥ 0.40 in absolute value were considered to be of meaningful magnitude.39 In addition, the Kaiser-Meyer-Olkin (KMO) measure was estimated to assess sampling adequacy and common variance. An overall KMO value of ≥ 0.70 was considered to reflect a good degree of consistency across items, making factor analysis appropriate. At the item level, KMO values < 0.50 were deemed unacceptable.40
Although the analysis plan called for conducting an EFA for the behavior scale as well, missing data for some items resulted in a very small sample of participants with complete behavior data (n=20). Following the approach used in this team’s prior research on nutrition behavior among AI/AN people living with diabetes, the nutrition behavior items were classified into two categories based on the behavioral recommendations in the AI/AN WCIE curriculum: Healthy Nutrition Behavior (18 items) and Unhealthy Nutrition Behavior (5 items). To explore the internal consistency reliability of the items assessing healthy and unhealthy behaviors, the standardized Cronbach alpha and McDonald omega for each behavioral category were computed. Preliminary correlation analyses showed that one behavior item, which was related to drinking water or unsweetened beverages, was not correlated with any other item. Hence, this item was excluded from the internal consistency reliability analysis for behavior items (and is not included in the table reporting results related to the behavior items).
Convergent Validity Analyses
To assess convergent validity, analyses examined the relationship of diabetes nutrition self-efficacy and behavior with constructs to which they were expected to be related. In the convergent validity analyses, self-efficacy was measured using two factor scores that were computed based on the results of the self-efficacy EFA (see Results section). These scores reflected (and are called) “Confidence in Making Healthy Nutrition Choices” and “Confidence in Using the Diabetes Plate,” respectively, and were computed as the mean of the items that loaded highly on the relevant factor. Internal consistency reliability analyses of the behavior items suggested that the healthy nutrition items had strong internal consistency, but the unhealthy nutrition items did not (see Results section). Hence, in the convergent analyses, a single behavior measure was used, which captured the average frequency with which participants engaged in “Healthy Nutrition Behavior.” Clinical indicators examined were BMI, systolic and diastolic blood pressure, and A1C.
The association of the self-efficacy measures, the behavior measure, and the clinical indicators were estimated using Pearson correlation. Analyses were performed using SAS version 9.4.41 P values < 0.05 were considered statistically significant.
Results
Sample Characteristics
The final sample included 60 AI/AN adults with T2D. Table 1 provides descriptive information about the sample. At baseline, participants were 58.6 years of age on average (range: 36-77) and most were women (71.7%). Nearly one-quarter of the sample (24.1%) had a 2- or 4-year college degree, with 11.1% having a graduate or professional degree. More than one-quarter of participants (26.0%) reported having a household income < $20,000 in the prior year. The same percentage of participants had incomes between $20,000 and $39,999 (26.0%) and 34.0% reported incomes of $40,000 or more. While 27.8% of participants reported working full or part time, 29.6% were disabled, 13.0% unemployed, and 25.9% retired. The average BMI across participants was 33.9 (range: 22.3-63.1). On average, participants had an A1C of 7.7% (range: 5.0-14.0%). Average systolic and diastolic blood pressure values were 130.0 mmHg and 78.5 mmHg, respectively.
Table 1.
Baseline Characteristics of 60 American Indian and Alaska Native Adults with Type 2 Diabetes Participating in a Randomized Controlled Trial Examining the Impact of the “What Can I Eat?” Diabetes Nutrition Education Curriculum Adapted for Native People
| Participant Characteristics | Mean (Standard Deviation) or % |
|---|---|
| Age (Years) | 58.6 (11.1) |
| Gender | |
| Female | 71.7% |
| Male | 28.3% |
| Other | 0.0% |
| Highest Level of Education Completed | |
| < High school graduate | 9.3% |
| High school graduate/General Educational Development (GED) Test | 16.7% |
| Some vocational or vocational certificate | 11.1% |
| Some college (no degree) | 27.8% |
| 2-year college degree | 16.7% |
| 4-year college degree | 7.4% |
| Graduate/professional degree | 11.1% |
| Total Annual Household Income | |
| < $10,000 | 6.0% |
| $10,000 to < $20,000 | 20.0% |
| $20,000 to < $40,000 | 26.0% |
| ≥ $40,000 | 34.0% |
| Missing | 14.0% |
| Employment Status | |
| Working full or part time | 27.8% |
| Disabled | 29.6% |
| Retired | 25.9% |
| Unemployed | 13.0% |
| Other (in school, seasonal job, never worked for pay) | 3.7% |
| Clinical Indicators | |
| Body Mass Index (kg/m2) | 33.9 (7.1) |
| Hemoglobin A1C (%) | 7.7 (1.9) |
| Systolic blood pressure (mmHg) | 130.0 (17.8) |
| Diastolic blood pressure (mmHg) | 78.5 (9.7) |
Internal Consistency Reliability
Diabetes Nutrition Self-efficacy Items
Table 2 presents participant responses to the self-efficacy items at baseline. Participants were most confident in their ability to drink water or unsweetened drinks. The average score of 2.72 for this item (on a scale from 0 to 3) suggested that participants were “somewhat confident” to “very confident” that they could successfully engage in this behavior. For 6 items, the average score hovered around 2.0 (range: 1.94-2.13), suggesting that participants were “somewhat confident” in their ability to engage in these activities (e.g., knowing which foods are carb foods). Participants showed the least confidence in response to the five items related to using the Diabetes Plate, knowing which foods have healthy and unhealthy fat, and getting 30 minutes of physical activity every day. Average scores for these items ranged from 1.56 to 1.81, suggesting that participants were “not very confident” to “somewhat confident” in their ability to engage in these activities.
Table 2.
Reliability of Diabetes Nutrition Self-efficacy Items Developed to Evaluate the “What Can I Eat?” Diabetes Nutrition Education Curriculum Adapted for American Indian and Alaska Native Adults with Type 2 Diabetes
| How confident are you that you can: | N | Mean (Standard Deviation) a |
Kaiser-Meyer- Olkin Measure |
Factor Loadings b | |
|---|---|---|---|---|---|
| Overall = 0.74 | Factor 1 Make Healthy Choices |
Factor 2 Use Diabetes Plate |
|||
| Know which foods are carb foods | 54 | 1.94 (0.86) | 0.83 | 0.30 | 0.67 |
| Use the Diabetes Plate to eat a healthy amount of carb foods | 54 | 1.67 (0.99) | 0.87 | 0.12 | 0.82 |
| Use the Diabetes Plate to fit sweet or sugary foods into your meals | 52 | 1.56 (1.00) | 0.80 | 0.15 | 0.89 |
| Choose to drink water or unsweetened drinks | 53 | 2.72 (0.57) | 0.47 | −0.05 | 0.20 |
| Use the Diabetes Plate to fit healthy protein foods into your meals | 53 | 1.79 (0.99) | 0.77 | 0.31 | 0.86 |
| Plan healthy meals to cook at home | 54 | 2.02 (0.86) | 0.65 | 0.70 | 0.16 |
| Know which foods have healthy and unhealthy fat | 49 | 1.73 (0.91) | 0.84 | 0.71 | 0.45 c |
| Know how to limit salt in your diet | 54 | 2.04 (0.99) | 0.56 | 0.39 d | 0.14 |
| Know how to eat healthy meals outside of your home | 53 | 1.94 (0.79) | 0.76 | 0.77 | 0.23 |
| Get 30 minutes of physical activity every day | 54 | 1.81 (1.01) | 0.38 | 0.36 | −0.19 |
| Make thoughtful or mindful nutrition choices | 50 | 2.02 (0.80) | 0.67 | 0.84 | 0.12 |
| Set goals to eat healthy foods | 54 | 2.13 (0.67) | 0.74 | 0.84 | −0.01 |
| Standardized Cronbach Alpha | 39 | --- | --- | 0.86 | 0.85 |
| McDonald Omega | 39 | 0.89 | 0.88 | ||
| Variance Explained | --- | --- | 45.1% | 39.0% | |
| Cumulative Variance Explained | 45.1% | 84.1% | |||
The self-efficacy items used a four-point scale: Not at all confident (0), not very confident (1), somewhat confident (2), and very confident (3).
Factor loadings are presented for the two factors that had eigenvalues > 1.0. Factor loadings ≥ 0.40 are formatted in bold and identify items deemed to load well on the relevant factor.
Item loaded on both factors. Because its factor loading was much stronger for Factor 1 than Factor 2, we included this item in the factor score for Factor 1.
The factor loading for this item (i.e., 0.39372) did not meet the threshold of 0.40. However, given that the factor loading was quite close to the threshold and that the item is highly relevant to Factor 1, we chose to include this item in the Factor 1 factor score.
Although missing data for individual items reduced the sample size for the self-efficacy EFA from 60 to 39, the analysis provided strong support for the reliability of the items. The self-efficacy EFA resulted in two factors with eigenvalues ≥ 1.0 (Table 2). Together, these factors accounted for 84.1% of the total variance, with Factors 1 and 2 explaining 45.1% and 39.0% of the variance, respectively. Two items did not load well on either factor (loadings < 0.40) and had unacceptable KMO values (< 0.50). Five items had strong loadings on Factor 1 (range: 0.70-0.84), which reflected Confidence in Making Healthy Nutrition Choices. One item just failed to reach the threshold of 0.40 for identifying acceptable fit with Factor 1 and had a fairly low KMO score of 0.56 (i.e., “How confident are you that you can: …know how to limit salt in your diet?”). Given the relevance of this item to Factor 1, that the item was only 0.006 below the factor loading threshold of 0.40, and that the item exceeded the 0.50 acceptability cutoff for the KMO statistic, the team chose to retain this item and identify this item as loading on Factor 1. Five items loaded on Factor 2 (range: 0.45-0.89), which mainly captured Confidence in Using the Diabetes Plate. One item loaded on both factors (i.e., “How confident are you that you can: …know which foods have healthy and unhealthy fat?”). Because the item loaded much more highly on Factor 1 than Factor 2 (0.71 vs. 0.45, respectively), it was included in Factor 1.
Factor scores for each of the two self-efficacy factors were computed. At baseline, participants were somewhat confident that they could successfully engage in the behaviors identified in the items that loaded on the Confidence in Making Healthy Nutrition Choices factor (x̄ = 1.99, SD = 0.64). The lower mean score for the Confidence in Using the Diabetes Plate factor suggested that participants were somewhat less confident in their ability to use the Diabetes Plate (x̄ = 1.74, SD = 0.80). Both factors had high internal consistency (standardized Cronbach alphas = 0.86 and 0.85, and McDonald omegas = 0.89 and 0.88, respectively). The overall KMO measure had a value of 0.74, indicating an acceptable proportion of common variance was shared across all self-efficacy items.
Diabetes Nutrition Behavior Items
Table 3 presents participant responses to the diabetes nutrition behavior items. Participants reported greatest success drinking healthy beverages. On average, they reported drinking water or unsweetened beverages once a day or more (mean = 5.41 on a scale from 1 to 6). (Because this item was not correlated with the other behavior items, it was excluded from the Healthy Nutrition Behavior category and omitted from Table 3.) Participants engaged in several other beneficial behaviors most days of the week (means ranged from 3.56-4.29; e.g., filling most of your plate with vegetables) and engaged in several other healthy behaviors about three or four days a week (range: 2.50-3.37; e.g., setting nutrition goals). Participants engaged in some healthy behaviors more rarely (range: 2.02-2.45, about one or two days a week; e.g., planning meals for the week). Participants reported engaging in unhealthy behaviors three or four days a week, or less (range: 2:08-2:82; e.g., eating processed foods).
Table 3.
Reliability of Diabetes Nutrition Behavior Items Developed to Evaluate the “What Can I Eat?” Diabetes Nutrition Education Curriculum Adapted for American Indian and Alaska Native Adults with Type 2 Diabetes
| During a typical week in the last month, how often did you: | Item # | N | Mean (Standard Deviation) a |
|---|---|---|---|
| Healthy Nutrition Behavior | |||
| Fill only 1 small part of your plate with carb foods | 1 | 46 | 3.87 (1.95) |
| Use the Diabetes Plate to plan the carb, protein, and veggie parts of your meals | 2 | 50 | 2.50 (1.76) |
| Fill most of your plate with vegetables (not including corn, peas, squash, or potatoes) | 3 | 52 | 3.62 (1.69) |
| Choose plant-based protein foods, like nuts, nut butters, beans, or legumes | 6 | 49 | 3.37 (1.56) |
| Bake, broil, or grill foods, without adding solid fats (like butter or margarine) | 7 | 52 | 2.98 (1.53) |
| Cook meals at home | 9 | 52 | 4.29 (1.59) |
| Make choices that you think are healthy when eating out at restaurants | 11 | 47 | 2.57 (1.49) |
| Plan your meals for the week | 12 | 46 | 2.22 (1.55) |
| Make a grocery list and use it at the store | 13 | 49 | 2.45 (1.44) |
| Choose oils (like olive or canola) rather than solid fats, like butter or margarine | 14 | 49 | 3.57 (1.73) |
| Eat traditional tribal foods | 17 | 45 | 2.02 (1.22) |
| Try to make thoughtful choices about the foods you choose to eat | 18 | 50 | 4.12 (1.76) |
| Try to eat 800 mg or less of sodium (salt) at each meal | 19 | 41 | 3.56 (1.86) |
| Read the food label on packaged foods to choose foods with less salt | 20 | 49 | 2.71 (1.65) |
| Read the food label on packaged drinks to choose drinks with no sugar | 21 | 51 | 3.29 (1.89) |
| Fit sweet or sugary foods (like cookies or desserts) into the carb section of the Diabetes Plate | 22 | 46 | 2.28 (1.47) |
| Help someone in your family or community eat healthier | 23 | 40 | 2.68 (1.77) |
| Set goals for yourself focused on what you will eat | 24 | 43 | 3.33 (1.81) |
| McDonald Omega | 0.92 | ||
| Standardized Cronbach Alpha | 0.89 | ||
| Unhealthy Nutrition Behavior | |||
| Drink regular soda, juice, sweetened tea, or other sugary beverages | 5 | 52 | 2.58 (1.73) |
| Eat fried foods, such as fried chicken, fried fish, fried potatoes | 8 | 49 | 2.82 (1.01) |
| Eat out at restaurants | 10 | 47 | 2.47 (1.02) |
| Eat processed foods (like Spam, potato chips, Ramen noodles) | 15 | 52 | 2.62 (1.25) |
| Eat frozen or boxed meals (like frozen dinners, Rice-A-Roni, or boxed mac and cheese) | 16 | 50 | 2.08 (1.31) |
| McDonald Omega | 0.88 | ||
| Standardized Cronbach Alpha | 0.48 |
Participants responded to each item using a six-point scale: More than once a day (6), once a day (5), 5-6 days a week (4), 3-4 days a week (3), 1-2 days a week (2), and never (1).
Composite scores for the two behavioral categories were computed: Healthy Nutrition Behavior and Unhealthy Nutrition Behavior. At baseline, participants reported engaging in Healthy Nutrition Behavior at an average frequency of 3-4 days per week (x̄ = 3.08, SD = 0.96), while engaging in Unhealthy Nutrition Behavior at an average frequency of 2-3 days per week (x̄ = 2.52, SD = 0.72). Although the items included in the Healthy Nutrition Behavior score had strong internal consistency (standardized Cronbach alpha = 0.89 and McDonald omega = 0.92), the items in the Unhealthy Nutrition Behavior score appeared to have poor internal consistency according to Cronbach alpha (0.48) but high reliability as measured by McDonald omega (0.88). Given these inconsistent results, only the Healthy Nutrition Behavior score was included in the convergent validity analyses.
Convergent Validity
To evaluate the convergent validity of the self-efficacy items, analyses were conducted to examine the association of the two self-efficacy factors with diabetes nutrition behavior and the clinical indicators. As shown in Table 4, the self-efficacy factors – Confidence in Making Healthy Nutrition Choices and Confidence in Using the Diabetes Plate – were strongly correlated with the Healthy Nutrition Behavior score. Greater confidence in the ability to make healthy choices and to use the Diabetes Plate was associated with a higher frequency of making healthy dietary choices (r = 0.47, P = 0.0004; r = 0.46, P = 0.0005, respectively). Although neither of the self-efficacy factors was associated with BMI or blood pressure (Ps > 0.05), Confidence in Using the Diabetes Plate was significantly associated with A1C (r = −0.32, P = 0.02). Stronger self-efficacy was associated with lower A1C (i.e., better glycemic control).
Table 4.
Convergent Validity of the Diabetes Nutrition Self-efficacy and Behavior Measures Developed to Evaluate the “What Can I Eat?” Diabetes Nutrition Education Curriculum Adapted for American Indian and Alaska Native Adults with Type 2 Diabetes a
| Diabetes Nutrition Behavior |
Clinical Indicators | ||||
|---|---|---|---|---|---|
| Healthy Nutrition Behavior |
Body Mass Index (kg/m2) |
Systolic Blood Pressure (mmHg) |
Diastolic Blood Pressure (mmHg) |
Hemoglobin A1C (%) |
|
| Diabetes Nutrition Self-efficacy b | |||||
| Confidence in Making Healthy Nutrition Choices |
0.47 (0.0004) (N = 52) |
−0.25 (0.0825) (N = 48) |
0.22 (0.1155) (N = 52) |
0.14 (0.3106) (N = 52) |
−0.15 (0.2985) (N = 51) |
| Confidence in Using the Diabetes Plate |
0.46 (0.0005) (N = 53) |
−0.07 (0.6262) (N = 54) |
0.19 (0.1846) (N = 53) |
0.09 (0.5436) (N = 53) |
−0.32 (0.0243) (N =52) |
| Diabetes Nutrition Behavior c | |||||
| Healthy Nutrition Behavior | −0.23 (0.1233) (N = 48) |
−0.07 (0.6314) (N = 52) |
−0.04 (0.7639) (N = 52) |
−0.38 (0.0057) (N = 51) |
|
Table presents Pearson correlation coefficients (P values). Correlations that are significant (P < 0.05) are formatted in bold text.
The self-efficacy measures were computed as the mean of the items that loaded highly on the relevant factor. The self-efficacy items used a four-point scale: Not at all confident (0), not very confident (1), somewhat confident (2), and very confident (3).
The Healthy Nutrition Behavior measure was computed as the average frequency with which participants engaged in nutrition behaviors identified by the curriculum as healthy. Behavior items used a six-point scale: More than once a day (6), once a day (5), 5-6 days a week (4), 3-4 days a week (3), 1-2 days a week (2), and never (1).
Finally, analyses were conducted to examine the association of diabetes nutrition behavior with the clinical indicators. The Healthy Nutrition Behavior score was significantly associated with A1C (r = -0.38, P = 0.0057), such that participants who engaged in healthy dietary behaviors more frequently had lower A1C values. No other significant relationships were seen between the diabetes nutrition behavior score and the clinical indicators.
Sensitivity Analysis
To examine the impact of missing behavior data, the research team conducted a post-hoc sensitivity analysis to determine whether there were any baseline differences between participants with complete behavior data and those with any missing behavior data. As shown in Supplementary Table 5, participants with and without missing data did not differ on age, gender, education, or most clinical metrics. However, participants with any missing behavior data differed significantly from those with complete data on income, employment, and diastolic blood pressure. Those with missing data appeared to have higher income; were more likely to be working full or part time, disabled, or unemployed (and less likely to be retired or have “other” as their employment status); and had higher diastolic blood pressure.
Discussion
The analysis reported here provides evidence that the diabetes nutrition self-efficacy and behavior items developed for the AI/AN WCIE program are valid and reliable. The self-efficacy factor scores – Confidence in Making Healthy Nutrition Choices and Confidence in Using the Diabetes Plate – were characterized by strong internal consistency reliability and demonstrated convergent validity through their robust associations with diabetes nutrition behavior. As expected, participants with greater self-efficacy engaged in healthy dietary behaviors more frequently than those with less confidence in their ability to make healthy nutrition choices and to use the Diabetes Plate. Additional evidence of convergent validity was seen in the significant association of Confidence in Using the Diabetes Plate with A1C. Consistent with the Expanded Health Belief Model, stronger self-efficacy was linked with better glycemic control. Finally, the behavior measure capturing Healthy Nutrition Behavior demonstrated excellent internal consistency reliability and was strongly associated with A1C. Participants who engaged in healthy behaviors more frequently had lower A1C than participants with less optimal behavior.
Although the self-efficacy and behavior scores were associated with A1C, they were not significantly related to other clinical measures. Neither the self-efficacy factor scores nor the behavior score was correlated with BMI or blood pressure. This result may not be surprising as weight loss and blood pressure control are not the primary goals of the AI/AN WCIE curriculum and few self-efficacy and behavior items address dietary behaviors specifically designed to reduce weight or blood pressure. Likewise, given that the Diabetes Plate and the AI/AN WCIE curriculum itself were developed to improve blood glucose management, it is reasonable that the self-efficacy measure assessing Confidence in Using the Diabetes Plate and the Healthy Nutrition Behavior measure would be particularly well associated with A1C.
Although the healthy nutrition items showed strong internal consistency, results were inconsistent regarding the reliability of the unhealthy behavior items. Simulation studies have shown that, for ordinal data, the McDonald omega statistic performed worse than Cronbach alpha when the number of items and correlations among the items were low.42 Thus, most likely, the reliability of the unhealthy behavior items was not ideal, as indicated by the value of their Cronbach alpha. There may be less within-subjects consistency in how participants engage in unhealthy behaviors as opposed to healthy behaviors. Or, it may be that the small number of unhealthy behavior items included in the AI/AN WCIE Impact Survey (N = 5) restricted the magnitude of the Cronbach alpha statistic.43 Evaluating the internal consistency of the unhealthy nutrition items with a larger sample would be of value.
Additional evaluation of the self-efficacy items may also be of value. In the analyses reported here, two items were poorly correlated with the other items in the self-efficacy scale and one item loaded well on both self-efficacy factors. Additional validation work with a larger sample, including a confirmatory factor analysis, may provide insight into the performance of these items and may verify the validity of the factor scores.
The diabetes nutrition self-efficacy and behavior measures evaluated as part of this analysis will provide a valuable resource in the implementation and dissemination of the AI/AN WCIE curriculum. These valid and reliable measures – which will be used to assess the longitudinal impact of the curriculum in the present sample – can facilitate rigorous, theory-driven evaluation of the program by future researchers. Likewise, the availability of validated measures that are tailored to the AI/AN WCIE curriculum will enable the use of consistent measures of diabetes nutrition self-efficacy and behavior, facilitating comparison of results across studies.
Organizations participating in the Special Diabetes Program for Indians (SDPI) may also benefit from the AI/AN WCIE curriculum and its associated self-efficacy and behavior measures. Established by Congress in 1997, SDPI funds diabetes prevention and treatment activities through IHS-administered grants to health care organizations serving Native communities. In Fiscal Year 2020, SDPI supported diabetes services at 301 IHS, tribal, and urban health care programs in 35 states, collectively serving more than 780,000 people. Ninety-four percent of these programs offered diabetes nutrition-related services to AI/AN adults.44 For those sites that choose to implement the AI/AN WCIE program, the availability of validated measures of self-efficacy and behavior can provide a valuable tool for evaluating and demonstrating the impact of their efforts.
The work reported here may also influence the evaluation of the WCIE curriculum in other populations. The WCIE program was initially developed by the ADA for use in populations that are disproportionately impacted by T2D.25 Although the AI/AN WCIE curriculum has been culturally adapted for AI/AN communities, much of the content is relevant to other populations that are medically underserved and disproportionately affected by diabetes. As a result, many of the items developed to measure diabetes nutrition self-efficacy and behavior among AI/AN participants may be valuable for use in other populations. The research team would welcome adaptation and testing of the items in other communities.
Although the original protocol for the randomized controlled trial that was the source of the data used in this secondary analysis involved in-person classes and data collection visits, the COVID-19 pandemic required that this approach be drastically revised. The AI/AN population experienced some of the most severe consequences of COVID-19 of any racial or ethnic group in the U.S. Compared to non-Hispanic Whites, AI/AN people were substantially more likely to be infected with the coronavirus, to be hospitalized, to die, and to die at younger ages.45-47 To ensure the safety of participants and project staff, the curriculum was adapted for virtual delivery and the team developed methods for recruitment and data collection that avoided person-to-person contact. Limitations of this approach include the possibility that participants might have been less accurate in measuring weight, blood pressure, and A1C than a study staff member might have been. This could be the reason there was not a significant relationship between the self-efficacy and behavior measures and the BMI and blood pressure measures. To ensure that participants were able to accurately measure their clinical indicators, they received written instructions and virtual guidance from the local study coordinator on using the body weight scale, blood pressure monitor, and home A1C test kits. Despite this potential limitation, adapting the curriculum and the study to a virtual format allowed this work to continue during the pandemic, ensuring that AI/AN people in participating communities were able to receive tailored diabetes nutrition education even while the communities had “safer at home” protocols in place. Going forward, the availability of a virtual diabetes nutrition education program may be of great value, particularly in communities that are geographically remote or do not have a local RDN to provide such educational offerings.
The research described here has important strengths. A key strength of the project was the involvement of a diverse set of AI/AN-serving health care organizations. The research team partnered with five clinical sites across the country, representing both tribal and urban health care programs as well as a range of tribal groups and geographic regions. The diversity in these clinical sites allowed the research team to incorporate some of the variability that exists in the AI/AN population as well as some of the complexities of the Indian health care system.
The study also had limitations. Although the trial included diverse types of AI/AN-serving health care organizations in different regions of the country, it could not capture the vast diversity of the Indian health care system or the AI/AN population. Hence, it is not clear whether the results of this study can generalize to other Native communities and other types of health care settings (e.g., IHS facilities).
The relatively high degree of missing data for the diabetes nutrition behavior items was also a limitation. It is unclear why missing data were more common for these items. The behavior scale may have been somewhat taxing for participants because it included a fairly large number of items (24), each with seven response options. As is well-established in the literature and is a salient problem for evaluations of dietary intake, collection of dietary behavior data is difficult. Participants may not always be able to recall their nutrition behavior, may not know the composition of their food or how it was prepared, and may be reluctant to report some dietary behaviors. These factors may have made selecting a response for some behavior items challenging.48,49 That participants with complete behavior data differed significantly from those with any missing behavior data on income, employment, and diastolic blood pressure suggests that the results of the behavioral analyses may not generalize to all AI/AN people.
An additional limitation of the study is the small sample size (N = 60), which contributed to challenges in conducting an EFA with the behavior items, made it infeasible to conduct a confirmatory factor analysis to confirm the results of the self-efficacy EFA, and may have reduced power to detect significant associations among measures. Although the research team intended to enroll 180 participants, 60 people were successfully recruited. The pandemic significantly delayed study recruitment and implementation of the pilot test. Early in the pandemic, some participating sites were forced to shut down after having already begun in-person classes. The resulting delay – while the team developed “touch-free” recruitment and data collection methods and adapted the curriculum for virtual delivery – cut into the time available to conduct the study. Likewise, the shutdown of non-essential activities in the collaborating partner clinics reduced the time site-based study coordinators were able to allocate to study recruitment and data collection. Further, the unexpected need to purchase and ship class kits – including data collection instruments such as body weight scales, blood pressure monitors, and A1C test kits – reduced the number of participants the team could afford to enroll. The inability to conduct the planned behavior EFA and to test the factor structure found in the self-efficacy EFA using a confirmatory factor analysis calls for future larger studies to confirm the factors identified in this trial. Despite the unprecedented disruption posed by the pandemic, this work reflects a rigorous trial of the AI/AN WCIE curriculum and offers support for diabetes nutrition education offered remotely to AI/AN communities.30
A final limitation is that the diabetes nutrition self-efficacy and behavior items did not undergo evaluation prior to their implementation in the randomized controlled trial. Ideally, a pilot test of the items would have been conducted, with validity and reliability assessed, prior to the beginning of the trial. Pilot testing would have identified items that were not well correlated with other items in their respective scales, loaded equally well on multiple factors, and might have provided insight into the causes of missing responses to the behavior items. Although pilot testing was not conducted before trial implementation, the results reported here suggest that the self-efficacy and behavior items can support the accurate evaluation of the impact of the AI/AN WCIE program on diabetes nutrition self-efficacy and behavior.
Conclusions
The reported analysis supports the reliability and validity of the diabetes nutrition self-efficacy and behavior items developed for the AI/AN WCIE program. These items – and the factor and composite scores computed from them – can facilitate rigorous and consistent measurement of diabetes nutrition self-efficacy and behavior by researchers and health care programs implementing the AI/AN WCIE curriculum in Native communities. Likewise, the items and methods used for assessing reliability and validity can inform efforts to develop and evaluate similar measures for WCIE programs addressing the needs of other populations.
Supplementary Material
Research Snapshot.
Research Question:
Are the diabetes nutrition self-efficacy and behavior items developed for the What Can I Eat? program for American Indian and Alaska Native (AI/AN) adults valid and reliable?
Key Findings:
Secondary analyses of baseline data from a randomized controlled trial conducted with 60 AI/AN adults with type 2 diabetes showed that the self-efficacy items were characterized by strong internal consistency reliability, were directly associated with healthy nutrition behavior, and were inversely associated with A1C. The behavior measure capturing Healthy Nutrition Behavior showed strong internal consistency and was inversely associated with A1C.
Acknowledgments
We thank the Shakopee Mdewakanton Sioux Community and the American Diabetes Association for their vision in instigating the cultural adaptation of the WCIE curriculum for American Indian and Alaska Native peoples. We are grateful to the study participants for their time and insight. We also thank the WCIE Study Group for their many contributions to this project. The WCIE Study Group includes Gemalli Austin, Nilofer Couture, Ulela Harris, Heather Garrow, Nancy O’Banion, Dawn Pavli, Heather Pontius, and Jessie Yanson. The authors received permission from those named in the Acknowledgments.
Funding/Support
Support for this work was provided by the American Diabetes Association (Grant 4-18-SMSC-01; PI: K. R. Moore) and the National Institute of Diabetes and Digestive and Kidney Diseases (Grant P30DK092923; Multiple PIs: S. M. Manson & A. G. Brega). The content of this report is solely the responsibility of the authors and does not necessarily represent the official views of the American Diabetes Association, the National Institutes of Health, or the Indian Health Service.
Footnotes
Conflict of Interest Disclosure
In 2021, Dr. Moore was a paid consultant for Novo Nordisk, Inc. All other authors have no conflicts of interest to report.
Clinical Trials Registration
The randomized controlled trial that was the source of data for this secondary analysis was registered on ClinicalTrials.gov (Trial Registration Number NCT04254653).
Contributor Information
Angela G. Brega, Centers for American Indian & Alaska Native Health, Colorado School of Public Health, University of Colorado Anschutz Medical Campus, Nighthorse Campbell Native Health Building, 13055 East 17th Avenue, Mail Stop F800, Aurora, CO 80045.
Sarah A. Stotz, Department of Food Science and Human Nutrition, Colorado State University, 1571 Campus Delivery, Fort Collins, CO 80523-1571.
Kelly R. Moore, Centers for American Indian & Alaska Native Health, Colorado School of Public Health, University of Colorado Anschutz Medical Campus, Nighthorse Campbell Native Health Building, 13055 East 17th Avenue, Mail Stop F800, Aurora, CO 80045.
Monica C. McNulty, Adult & Child Center for Outcomes Research & Delivery Science, School of Medicine, University of Colorado Anschutz Medical Campus, 1890 North Revere Court, Mailstop F443, Aurora, CO 80045.
Luohua Jiang, Department of Epidemiology & Biostatistics, Susan and Henry Samueli College of Health Sciences, University of California Irvine, 856 Health Sciences Quad, Room 3412, Irvine, CA 92697-3957.
References
- 1.Centers for Disease Control and Prevention. National Diabetes Statistics Report. https://www.cdc.gov/diabetes/data/statistics-report/appendix.html#tabs-1-3. Accessed January 13, 2024.
- 2.Zhang Y, Galloway JM, Welty TK, et al. Incidence and risk factors for stroke in American Indians the Strong Heart Study. Circulation. 2008;118(15):1577–1584. doi: 10.1161/CIRCULATIONAHA.108.772285 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Guo Y, Huang Z, Sang D, Gao Q, Li Q. The Role of Nutrition in the Prevention and Intervention of Type 2 Diabetes. Front Bioeng Biotechnol. 2020;8:575442. doi: 10.3389/fbioe.2020.575442 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Evert AB, Dennison M, Gardner CD, et al. Nutrition Therapy for Adults With Diabetes or Prediabetes: A Consensus Report. Diabetes Care. 2019;42(5):731–754. doi: 10.2337/dci19-0014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Warne D, Wescott S. Social determinants of American Indian nutritional health. Curr Dev Nutr. 2019;3(Suppl 2):12–18. doi: 10.1093/cdn/nzz054 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gurney RM, Caniglia BS, Mix TL, Baum KA. Native American Food Security and Traditional Foods: A Review of the Literature. Sociol Compass. 2015;8:681–693. [Google Scholar]
- 7.Smith-Morris CM. Reducing diabetes in Indian country: Lessons from the three domains influencing Pima diabetes. Hum Organ. 2015;63(1):34–46. doi: 10.17730/humo.63.1.mfp4tupu29tcvnqm [DOI] [Google Scholar]
- 8.Hsu CC, Lee CH, Wahlqvist ML, et al. Poverty increases type 2 diabetes incidence and inequality of care despite universal health coverage. Diabetes Care. 2012;35(11):2286–2292. doi: 10.2337/dc11-2052 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.US Department of Health and Human Services. Profile: American Indian/Alaska Native. Office of Minority Health. https://minorityhealth.hhs.gov/omh/browse.aspx?lvl=4&lvlid=33. Published 2019. Accessed April 26, 2024. [Google Scholar]
- 10.Bower KM, Thorpe RJ, Rohde C, Gaskin DJ. The intersection of neighborhood racial segregation, poverty, and urbanicity and its impact on food store availability in the United States. Prev Med (Baltim). 2014;58:33–39. doi: 10.1016/j.ypmed.2013.10.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Nikolaus CJ, Johnson S, Benally T, et al. Food Insecurity among American Indian and Alaska Native People: A Scoping Review to Inform Future Research and Policy Needs. Adv Nutr. 2022:1–18. doi: 10.1093/advances/nmac008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Satterfield D, Frank M. Health promotion and diabetes prevention in American Indian and Alaska Native communities -Traditional foods project, 2008-2014. Morb Mortal Wkly Rep Suppl. 2016;65(1):4–10. [DOI] [PubMed] [Google Scholar]
- 13.Wilson C, Brown T, Acton K, Gilliland S. Effects of clinical nutrition education and educator discipline on glycemic control outcomes in the Indian Health Service. Diabetes Care. 2003;26(6):2500–2504. [DOI] [PubMed] [Google Scholar]
- 14.Miller CK, Edwards L, Kissling G, Sanville L. Nutrition education improves metabolic outcomes among older adults with diabetes mellitus: Results from a randomized controlled trial. Prev Med (Baltim). 2002;34(2):252–259. doi: 10.1006/pmed.2001.0985 [DOI] [PubMed] [Google Scholar]
- 15.Guthrie JF, Stommes E, Voichick J. Evaluating food stamp nutrition education: issues and opportunities. J Nutr Educ Behav. 2006;38(1):6–11. doi: 10.1016/j.jneb.2005.11.001 [DOI] [PubMed] [Google Scholar]
- 16.Weinstein E, Galindo RJ, Fried M, Rucker L, Davis NJ. Impact of a focused nutrition educational intervention coupled with improved access to fresh produce on purchasing behavior and consumption of fruits and vegetables in overweight patients with diabetes mellitus. Diabetes Educ. 2014;40 VN-r(1):100–106. doi: 10.1177/0145721713508823 [DOI] [PubMed] [Google Scholar]
- 17.Jiang L, Manson S, Beals J, et al. Translating the diabetes prevention program into American Indian and Alaska Native communities: Results from the special diabetes program for Indians diabetes prevention demonstration project. Diabetes Care. 2013;36(7):2027–2036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Homenko DR, Morin PC, Eimicke JP, Teresi JA, Weinstock RS. Food insecurity and food choices in rural older adults with diabetes receiving nutrition education via telemedicine. J Nutr Educ Behav. 2010;42(6):404–409. doi: 10.1016/j.jneb.2009.08.001 [DOI] [PubMed] [Google Scholar]
- 19.Seligman HK, Lyles C, Marshall MB, et al. A pilot food bank intervention featuring diabetes-appropriate food improved glycemic control among clients in three states. Health Aff. 2015;34(11):1956–1963. doi: 10.1377/hlthaff.2015.0641 [DOI] [PubMed] [Google Scholar]
- 20.Taylor-Powell E. Evaluating food stamp nutrition education: a view from the field of program evaluation. J Nutr Educ Behav. 2006;38(1):12–17. doi: 10.1016/j.jneb.2005.11.007 [DOI] [PubMed] [Google Scholar]
- 21.Fleischhacker S. Emerging opportunities for registered dietitian nutritionists to help raise a healthier generation of Native American youth. J Acad Nutr Diet. 2016;116(2):219–225. doi: 10.1016/j.jand.2015.10.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Jackson Y, Broussard B. Cultural Challenges in Nutrition Education Among American Indians. Diabetes Educ. 1997;13(1):47–50. doi: 10.1177/014572178701300110 [DOI] [PubMed] [Google Scholar]
- 23.Shaw JL, Brown J, Khan B, Mau MK, Dillard D. Resources, roadblocks and turning points: A qualitative study of American Indian/ Alaska Native adults with type 2 diabetes. J Community Health. 2013;38(1):86–94. doi: 10.1007/s10900-012-9585-5 [DOI] [PubMed] [Google Scholar]
- 24.Schure M, Goins RT, Jones J, Winchester B, Bradley V. Dietary Beliefs and Management of Older American Indians with Type 2 Diabetes. J Nutr Educ Behav. 2018;000(000). doi: 10.1016/j.jneb.2018.11.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.American Diabetes Association. What Can I Eat? Healthy Choices for People with Type 2 Diabetes. Summary of Impact for Funders.; 2017. [Google Scholar]
- 26.American Diabetes Association. What Is the Diabetes Plate Method? American Diabetes Association - Connected for Life. https://www.diabetesfoodhub.org/articles/what-is-the-diabetes-plate-method.html. Published 2020. Accessed August 16, 2022.
- 27.Stotz SA, Brega AG, Henderson NJ, Lockhart S, Moore KR. Food insecurity and associated challenges to healthy eating among American Indians and Alaska Natives with type 2 diabetes: Multiple stakeholder perspectives. J Aging Health. 2021;33(7-8_suppl):31S–39S. doi: 10.1177/08982643211013232 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Stotz S, Brega A, Gonzales K, Hebert L, Moore K. Facilitators and barriers to healthy eating among American Indian and Alaska Native adults with type 2 diabetes: stakeholder perspectives. Curr Dev Nutr. 2021;5(4):22–31. 10.1093/cdn/nzaa114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Stotz S, Brega AG, Lockhart S, et al. An online diabetes nutrition education programme for American Indian and Alaska Native adults with type 2 diabetes: perspectives from key stakeholders. Public Health Nutr. 2020:1–11. doi: 10.1017/S1368980020001743 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Stotz S, Moore K, McNulty M, et al. Evaluation of a synchronous, online diabetes nutrition education program for American Indians and Alaska Natives with type 2 diabetes: Facilitators and participants’ experiences. J Nutr Educ Behav. 2023;55(2):114–124. doi:doi.org/ 10.1016/j.jneb.2022.10.013 [DOI] [PubMed] [Google Scholar]
- 31.Burns A. The expanded health belief model as a basis for enlightened preventive health care practice and research. J Health Care Mark. 1992;12(3):32–45. [PubMed] [Google Scholar]
- 32.Johnson E, Rhoades E. The history and organization of Indian Health Services and systems. In: American Indian Health: Innovations in Health Care, Promotion and Policy. Baltimore, MD: Johns Hopkins University Press; 2000:74–92. [Google Scholar]
- 33.Warne D, Frizzell LB. American Indian Health Policy: Historical trends and contemporary issues. Am J Public Health. 2014;104(SUPPL. 3):263–267. doi: 10.2105/AJPH.2013.301682 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Leggat-Barr K, Goldman N, Uchikoshi F. COVID-19 risk factors and mortality among Native Americans. Demogr Res. 2021;45(39):1185–1218. doi: 10.4054/DemRes.2021.45.39 [DOI] [Google Scholar]
- 35.Cunningham JA, Kypri K, McCambridge J. Exploratory randomized controlled trial evaluating the impact of a waiting list control design. BMC Med Res Methodol. 2013;13(1). doi: 10.1186/1471-2288-13-150 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)-A metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377–381. doi: 10.1016/j.jbi.2008.08.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Hayes AF, Coutts JJ. Use Omega Rather than Cronbach’s Alpha for Estimating Reliability. But…. Commun Methods Meas. 2020;14(1):1–24. doi: 10.1080/19312458.2020.1718629 [DOI] [Google Scholar]
- 38.Nunnally J, Bernstein I. Psychometric Theory. New York: McGraw Hill; 1994. [Google Scholar]
- 39.Stevens J. Applied Multivariate Statistics for the Social Sciences. 2nd ed. L Erlbaum Associates; 1992. [Google Scholar]
- 40.Watkins M. Exploratory Factor Analysis: A Guide to Best Practices. J Black Psychol. 2018;44(3):219–246. doi: 10.1177/0095798418771807 [DOI] [Google Scholar]
- 41.SAS Institute Inc. SAS Statistical Software. Version 9.4 Cary, NC. [Google Scholar]
- 42.Orcan F. Comparison of Cronbach’s alpha and McDonald’s omega for ordinal data: Are they different? Int J Assess Tools Educ. 2023;10(4):709–722. doi: 10.21449/ijate.1271693 [DOI] [Google Scholar]
- 43.Carmines E, Zeller R. Quantitative Applications in the Social Sciences Series. In: Reliability and Validity Assessment. SAGE Publications; 1978. [Google Scholar]
- 44.Indian Health Service. Special Diabetes Program for Indians 2020 Report to Congress: Changing the Course of Diabetes: Charting Remarkable Progress. Rockville, MD; 2020. [PubMed] [Google Scholar]
- 45.Arrazola J, Masiello M, Joshi S. COVID-19 Among American Indian and Alaska Native Persons —. Morb Mortal Wkly Rep. 2020;69(34):1–4. doi: 10.15585/mmwr.mm6949a3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Hatcher SM, Agnew-Brune C, Anderson M, et al. COVID-19 among American Indian and Alaska Native persons. Morb Mortal Wkly Rep. 2020;69(34):1–4. doi: 10.15585/mmwr.mm6934e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Aschmann HE, Riley AR, Chen R, et al. Dynamics of racial disparities in all-cause mortality during the COVID-19 pandemic. Proc Natl Acad Sci. 2022;119(40):e2210941119. doi: 10.1073/pnas.2210941119 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Grandjean AC. Dietary intake data collection: challenges and limitations. Nutr Rev. 2012;70(suppl_2):S101–S104. doi: 10.1111/j.1753-4887.2012.00545.x [DOI] [PubMed] [Google Scholar]
- 49.Guthrie JF, Smallwood DM. Evaluating the effects of the dietary guidelines for Americans on consumer behavior and health: methodological challenges. J Am Diet Assoc. 2003;103(12, Supplement):42–49. doi: 10.1016/j.jada.2003.09.036 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
