Lowering the Impact of Food Insecurity in African American adults with Type 2 Diabetes Mellitus (LIFT-DM) – Study Protocol for a Randomized Controlled Trial

Abstract

Background:

There is strong evidence that disparities in the burden of diabetes exist by both race and poverty. Food insecurity, or an inability to or limitation in accessing nutritionally adequate food, is an important modifiable social determinant of health, particularly in adults with chronic disease. African Americans are more likely to be diagnosed with diabetes and more likely than whites to be food insecure.

Methods:

We describe a 4-year ongoing randomized controlled trial, which will test the separate and combined efficacy of monthly food vouchers and monthly mailed food stock boxes layered upon diabetes education in improving glycemic control in low income, food insecure, African Americans with type 2 diabetes mellitus using a 2×2 factorial design. Three hundred African American adults with clinical diagnosis of diabetes and HbA1c≥8% will be randomized into one of four groups: 1) diabetes education alone; 2) diabetes education plus food vouchers; 3) diabetes education plus stock boxes; and 4) diabetes education plus combined food vouchers and stock boxes. Our primary hypothesis is: among low-income, food insecure, African Americans with type 2 diabetes, those receiving diabetes education enhanced with food supplementation (food vouchers alone, mailed stock boxes alone, or combination) will have significantly greater reduction in HbA1c at 12 months compared to those receiving diabetes education only.

Discussion:

Results from this study will yield valuable insight currently lacking on how best to design and deliver diabetes interventions to low-income, food insecure, African Americans with diabetes that takes into account both clinical and social determinants of health.

Trial registration:

This study was registered on November 29, 2019 with the United States National Institutes of Health Clinical Trials Registry (ClinicalTrials.gov identifier# NCT04181424).

Introduction

Background

Diabetes is the 7th leading cause of death, and the leading cause of kidney disease, amputation, and blindness in adults in the United States [1]. Diabetes disproportionately affects African Americans compared to whites with a prevalence of 11.7% in African Americans compared to 7.5% in Non-Hispanic Whites [1]. Among individuals with diabetes, African Americans and those living below the federal poverty level have poorer metabolic control, higher complication rates and higher mortality compared to Whites and those living at the highest income level [1,2,3]. Hence, there is a need for research focused on addressing modifiable factors, such as diet, leading to this disparity.

Food insecurity, defined as an inability to or limitation in accessing nutritionally adequate foods impacts approximately 11.1% of US households, including households with both low and very low food security [4]. Food insecurity is a modifiable determinant of health and health disparities [4,5]. For instance, evidence suggests that US adults reporting very low food security have a two-fold increased likelihood of having diabetes compared to food secure adults [5]. In addition, African Americans are almost three times more likely than Whites to be food insecure [4].

Food insecurity also has detrimental effects on diabetes management [6]. Food insecure individuals with diabetes report difficulty with carbohydrate counting and tracking servings due to financial constraints and inconsistent income [7]. In addition, food insecurity is related to greater diabetes distress [8], low self-efficacy [9], and those with diabetes who are food insecure have worse glycemic control compared to those who are food secure [9–11]. The underlying mechanisms driving this association are not well known; however, evidence suggests that multidimensional factors such as competing needs and reliance on affordable calorically dense, nutrient poor, meals may be important driving factors for diabetes incidence and poor self-management among food insecure individuals [9].

Evidence suggests that nutritional education and diabetes management delivered via telephone improves access to information and skills for food insecure individuals with diabetes; however, little impact has been seen on food security status or outcomes [12]. Additionally, the use of farmers markets and food vouchers have been shown to improve access to nutritious foods for individuals with food insecurity and diabetes, however little evidence exists on their impact on diabetes management and outcomes [13]. A number of federal and local resources exist to assist food insecure individuals, such as the Supplemental Nutrition Assistance Program (SNAP), the Women, Infants, and Children (WIC) program, and food pantries. While these programs improve caloric, macronutrient and micronutrient intake, some individuals may not be eligible due to income guidelines and the food choices available may not be ideal for those with diabetes [6,14,15]. Specifically, these programs may not provide an adequate variety and quantity of foods to keep blood glucose under control. In addition, these programs do not account for the complexities that a diabetes diet presents and may not be sufficient to reduce the growing burden that food insecurity creates for individuals with diabetes due to variable or scarce nutritious options [16,17].

As diabetes is a complex disease requiring self-management and access to a nutrient rich diet, interventions to address the co-occurrence of food insecurity and diabetes are highly warranted. The 2017 American Diabetes Association Standards of Care calls for assessing food insecurity and associated competing needs at the clinic level when developing treatment plans and making clinical decisions for patients with diabetes [18]. While these recommendations signify upstream approaches to addressing food insecurity and diabetes at the clinic level, there is a paucity of evidence for downstream solutions for treatment and interventions once food insecurity is identified as a barrier for diabetes management [6]. This limitation necessitates the development of tailored interventions to address food insecurity and diabetes, such as prescribed meals or medically tailored food and nutrition interventions [19].

In this paper, we describe the rationale, study aim and objectives, research design and methods of an ongoing four-year randomized controlled clinical trial to test the separate and combined efficacy of monthly food vouchers to healthy food and monthly food stock boxes layered upon diabetes education in improving glycemic control in low income, food insecure, African Americans with Type 2 Diabetes (T2DM) using a 2×2 factorial design.

Rationale

There is strong evidence that disparities in the burden of diabetes exist by both race and poverty [2,20–22]. African Americans are more likely to be diagnosed with diabetes and more likely than whites to be food insecure [1,4]. Food insecure individuals with diabetes have worse glycemic control, report lower dietary quality and more difficulty following a healthy diet compared to those who are not food insecure [7,9,23,24]. Evidence suggests food insecurity better predicts chronic disease than income, highlighting its importance as a modifiable factor [25]. Currently tested strategies to address food insecurity include food supplementation in the form of: 1) vouchers to cover cost of food; 2) vouchers/coupons that can only be used at farmer’s markets; and 3) shipments of pre-packaged food (stock boxes) to participants homes [6,23,26,27]. However, within food supplementation options, it is unclear whether mailed stock boxes are superior to food vouchers or the combination (mailed stock boxes and food vouchers) are superior to either food supplementation option alone in achieving glycemic control. Also, there is limited evidence on whether diabetes education in combination with food supplementation leads to improved clinical outcomes compared to education alone. This study provides an opportunity to address this gap in the knowledge.

Study Aims and Objectives

The primary aim of this study is to test the separate and combined efficacy of monthly food vouchers to farmers market and monthly mailed food stock boxes layered upon diabetes education in improving glycemic control in low income, food insecure, African Americans with Type 2 Diabetes (T2DM). The primary outcome is HbA1c at 12 months post-randomization and the secondary outcomes are blood pressure (BP) control, low density lipoprotein (LDL), quality of life (QoL), and resource utilization/cost at 12 months post-randomization.

Methods

Funding for this four-year study was received in May 2019 with an anticipated end date of December 2023. This 4-arm RCT using a 2X2 factorial design is currently ongoing with randomization of individual study participants and blinded outcomes assessments at baseline, 3, 6, 9 and 12months.

Location and setting

The study site is the Medical College of Wisconsin (MCW), Milwaukee, USA. Recruitment will take place in the Milwaukee city community and communities throughout Milwaukee County.

Ethics and Trials Registration

This study is funded by grant the National Institute for Minority Health and Health Disparities (R01MD013826). The Institutional Review Board (IRB) of the Medical College of Wisconsin approved this trial on December 6, 2018, PRO #00033749. The trial is registered (registration date, November 29, 2019) on the United States National Institutes of Health Clinical Trials Registry (ClinicalTrials.gov identifier# NCT04181424), available online at: http://clinicaltrials.gov/ct2/show/NCT04181424

Trial Population and Recruitment

A total of 300 food insecure African American adults with diagnosed T2DM will be randomized to one of four groups: 1) diabetes education alone; 2) diabetes education plus food vouchers; 3) diabetes education plus stock boxes; and 4) diabetes education plus combined food vouchers and stock boxes.

Recruitment of eligible participants in collaboration with community partners and food pantries is underway using two complementary approaches. First, recruitment targets community partners that provide services for food insecure individuals. We have worked with multiple food pantries in low-income African American communities in Milwaukee to recruit for current studies. In addition, Hunger Task Force, Milwaukee’s free and local food bank, has partnered with our team to provide access to pantries and USDA stock box delivery locations that target low-income seniors in Milwaukee. Community partners serving low-income food insecure individuals who are interested in partnering with us display IRB approved recruitment flyers in prominent locations. We also offer recruitment pre-screening sessions at community sites where interested individuals can receive information on the study and criteria for eligibility. Individuals who are identified as likely to be eligible who are interested in participating in the study then complete eligibility assessment and enrollment.

Second, we are targeting the 10 zip codes in Milwaukee with the highest percentage of African Americans living at an income less than $15,000 per year. We have established relationships with a number of community sites in these target zip codes, including churches, barber shops, beauty salons, and YMCAs through an ongoing project to build capacity in low-income African American communities in Milwaukee. We distribute and/or display IRB approved recruitment flyers in prominent locations in public housing locations and community venues. We share the goals of the study and inclusion/exclusion criteria with community leaders and directly with community members. We offer pre-screening sessions at community sites interested in serving as partners with our recruitment process using the same methods as screening at food pantry screening sessions.

In an effort to assure study participants meet criteria for study participation, research study staff obtain a written informed consent and complete screening for eligibility irrespective of recruitment method. The study procedure and risks are explained to the participants and the consent form is signed according to standard clinical research practice. Participants who meet eligibility criteria complete the remainder of the assessments (see Table 1).

Table 1:

Data Collection Schedule

Questionnaires/Measurements	Screening Visit	Baseline Visit	3-month Visit	6-month Visit	12-month Visit
Primary Outcome Measure
HbA1c	X		X	X	X
Secondary Outcome Measures
Resource Utilization/Cost		X	X	X	X
Quality of Life (SF – 12)		X	X	X	X
Quality Adjusted Life Years (EQ-5D)		X	X	X	X
Blood Pressure		X	X	X	X
LDL-Cholesterol		X	X	X	X
Exploratory Analyses Measures
Behavioral Skills		X	X	X	X
Empowerment		X	X	X	X
Self-efficacy		X	X	X	X
Food Frequency Questionnaire		X	X	X	X
Covariates
Patient Demographics	X
Food Insecurity		X	X	X	X
Health Literacy		X
Depression (PHQ-9)		X	X	X	X
Medical Comorbidity		X

Randomization

A permuted block randomization method is used to assign subjects to one of the four intervention groups (cells within the 2×2 factorial design): Group 1 (Diabetes Education Only); Group 2 (Diabetes Education Plus Monthly Food Vouchers); Group 3 (Diabetes Education Plus Monthly Stock Boxes); and Group 4 (Diabetes Education Plus Combination of Monthly Food Vouchers and Stock Boxes). Block size are varied to protect blinding. The randomization is stratified by baseline HbA1c levels (8–10% vs. >10%). Research study staff collect eligibility information using Research Electronic Data Capture and once eligibility is confirmed, intervention assignment is made by a pre-programmed randomization scheme. All participants who are randomized are entered into the study database and analyzed according to Consolidated Standards of Reporting Trials guidelines.

Description of Intervention Groups

Group 1: Diabetes Education

Diabetes education is based on materials developed by our group that has been tested in prior studies in African Americans with T2DM [28]. It consists of culturally tailored diabetes education and skills training modules delivered one-on-one via telephone by a trained nurse educator. Modules were developed based on American Diabetes Association (ADA) guidelines for diabetes education. Content was developed based on the principles of adult learning theory and was designed to be relevant, person-centered, and using language at or below a seventh-grade level. Education modules were reviewed by an endocrinologist, primary care physician, diabetes educator, and nurse practitioner and revised based on feedback. Weekly sessions cover: 1) Overview of ABC’s of diabetes; 2) Self Blood Glucose Monitoring; 3) Medications; 4) Basics of Eating; 5) Meal Planning; 6) Carbohydrate Counting; 7) Exercise; 8) Blood Pressure; 9) Cholesterol; 10) Foot and Skin Care; and 11) Stress. The twelfth session provides a summary of prior session materials. Behavioral skills are addressed through individualized problem-solving and self-monitoring strategies. Sessions will stress four behaviors – physical activity, diet, medication adherence, and glucose self-monitoring.

Two registered nurses will be trained to deliver the diabetes education modules. Training will consist of two full days of information and role playing prior to the start of the study, and one day booster sessions each year to minimize drift in intervention delivery. Nurse educators will also have weekly supervision meetings with a nurse practitioner to answer questions and address any concerns. One nurse educator will be assigned to deliver sessions to an individual participant over time. Using the participant’s current barriers and preferences, target behavior goals will be set and each of the behaviors will be covered in 3 sessions over the course of the intervention. Twelve 30-minute weekly sessions will be delivered via the telephone, with two 30-minute booster sessions at months 6 and 9.

Group 2: Diabetes Education and Monthly Food Vouchers

Participants randomized to this group receive monthly food vouchers in addition to diabetes education as detailed above. Monthly food vouchers are intended to provide additional access to healthy food beyond that which participants currently have. Individuals will not be prevented from obtaining food from other programs such as WIC, SNAP, or food pantries, but the supplementation from the intervention will not be increased to cover family size. This decision was based on the need to ensure an equivalent intervention cost across participants given the individual basis of the intervention design. Family size will be measured within the patient demographics to allow further investigation after the study. Vouchers will be limited to purchasing food at the Hunger Task Force’s Mobile Market, which is a grocery store that travels throughout Milwaukee and surrounding counties during the week to visit neighborhoods with limited access to fresh and healthy foods. Specifying use of vouchers for purchase of food at that Mobile Market is intended to support increasing purchase of fruits and vegetables, lean meats, and dairy items. The market does not sell canned items or processed foods. Each month $50 will be redeemable on the voucher for use prior to the end of the month.

Group 3: Diabetes Education and Monthly Stock Boxes

Participants randomized to this group receive monthly stock boxes in addition to diabetes education as detailed above. Monthly pre-packaged food boxes are intended to provide additional access to healthy food beyond that which participants currently have. Individuals will not be prevented from obtaining food from other programs such as WIC, SNAP, or food pantries, but the supplementation from the intervention will not be increased to cover family size. Stock boxes have foods selected based on ADA diet guidelines to include fresh produce, whole and unrefined grain products, low sodium canned vegetables, sauces, and soups, low-fat dairy, lean meats, and beans. Items such as spices, cooking oil, and utensils for cooking will also be included in boxes to facilitate individuals introducing new products into their diet. The Hunger Task Force creates similar boxes through the USDA commodities program, and has worked with our team to design stock boxes for diabetes specific diets. While stock boxes are not changed to meet food preferences of participants, contents of the boxes were discussed with community members through focus groups prior to the study to ensure they are culturally acceptable to participants. Stock boxes with foods worth $50 are available for pick-up each month for 12 months at convenient community locations.

Group 4: Diabetes Education, Monthly Food Vouchers and Monthly Stock Boxes

Individuals assigned to this group will receive a combination of diabetes education, monthly food vouchers for use at the Mobile Market, and monthly stock boxes with diabetes appropriate food items available to pick up at the Mobile Market as detailed above. Members of the community participating in focus groups conducted prior to the study indicated an interest in having an option to select some foods through a voucher and provide other foods through a stock box. Therefore, the study was designed to incorporate this feedback from the community and test the separate and combined efficacy of two types of food supplementation. Each month for 12 months participants will have $25 loaded on their voucher for use at the Mobile Market and have a stock box with food worth $25 available for pick up at convenient community locations.

Study Instruments and Data Collection Schedule

Data will be collected at baseline, 3-months, 6-months, and 12-months. The study design, study flow, data collection schedule, measures and instruments are shown in Tables 1–3 and Figure 1.

Table 3:

Data Collection Instruments

Measure	Data Collected	Method
Process and Behavioral Measures	Behavioral Skills	The Summary of Diabetes Self-Care Activities scale [31], a brief, validated self-report questionnaire of diabetes self-care will measure behavioral skills.
	Empowerment	This will be measured by the 8-item Diabetes Empowerment Scale-Short Form, which has a Cronbach’s alpha of 0.85 [32].
	Self-efficacy	This will be measured by the 8-item perceived diabetes self-management scale [33], which has Cronbach’s alpha of 0.83 indicating internal consistency.
	Food Frequency Questionnaire	Dietary intake will be measured by the NCI food frequency questionnaire[34]. This has been validated in multiple culture groups and used regularly in national surveys [34].
Covariates	Patient Demographics	Demographic characteristics will be captured using previously validated items from the National Health Interview Survey.
	Food Insecurity	This will be measured using a 6-item scale which classifies household food security status and is valid for households with and without children [35].
	Health Literacy	A 3-item literacy scale noting capacity to obtain, process, and understand basic health-related decisions will be used to measure health literacy [36].
	Depression	The Patient Health Questionnaire-9 is a brief questionnaire that scores each of the nine Diagnostic and Statistical Manual of Mental Disorders (4th revision) criteria for depression [37].
	Medical Comorbidity	This will be assessed using previously validated items from the Behavioral Risk Factor Surveillance System [38].

Figure 1:

Study Design and Study Flow

Primary Outcome Measure

The primary outcome is glycemic control measured by HbA1c at 12 months of follow-up.

Sample Size Determination and Power Analysis

The primary (HbA1c) and secondary outcomes (BP, LDL, QOL) are measured at 3-, 6-, and 12 months post-randomization with the primary time point at 12 months post randomization. For hypotheses involving 3 comparisons (e.g the combination food supplementation group has superior efficacy to each of the individual interventions alone (vouchers, stock boxes, diabetes education)), with 60 participants in each group, there will be 85% power to detect a 0.48sd between-group standardized effect size. [39] This assumes 3 post-randomization measurement time points, level of significance alpha=0.02 (two-tailed) to adjust for 3 individual comparisons, and correlation between pairs of measurements within participants (interclass correlation) no larger than rho=0.5, and AR(1) covariance structure. For example, assuming the standard deviation (sd) for the primary outcome (HbA1c) is 1.9, the raw HbA1c effect sizes that can be detected for these comparisons is 0.95 percentage points; for the secondary outcomes, the raw effect sizes for SBP (sd:19.9 mmHg), LDL (sd: 20 mg/DL), SF-12 (sd: 12 Points) are 9.3 mmHg, 9.4 mg/DL, and 5.6 points, respectively.

To account for the “fraction of missing” information that must be imputed in the intent-to-treat (ITT) sample and the dilution effect of ITT analyses [40], we increase the sample size by 20% to achieve a final sample size of 75 participants randomized to each treatment group (total N=300).

Data Analysis

For the analyses of primary (HbA1c) and secondary (SBP, LDL, QOL) outcomes, a longitudinal mixed model approach will be used as the general analytic framework. Each primary/secondary outcome will be used separately as the dependent variable with intervention groups, time, and time by groups interaction as the primary independent variables and baseline for each outcome as an adjustment covariable. The magnitude and direction of the effect sizes for the a priori specified comparisons will be estimated using 95% confidence intervals for model-based contrasts of least squares means (LSM). In a set of secondary analyses, additional adjustment covariables (e.g. age, sex, education, income and comorbidity) will be added to the model, if indicated.

Exploratory Analyses

Intervention effects on self-care behaviors (diet, exercise, self-monitoring, medication adherence) will be explored using the longitudinal mixed modeling procedure (with appropriate link function) described above, with each exploratory outcome used separately as the dependent variable in the model. The role of diabetes empowerment, self-efficacy, and dietary intake as mediators of intervention effects on HbA1c, BP, LDL and QOL will be explored following the method of Baron & Kenney (1986) and Preacher & Hayes, 2004 [41,42]. Possible mediating effect (partial or full) will be evaluated using structural equation modeling (SEM) [43].

Cost and Cost Effectiveness Analyses

Determination of organizational and system costs, and ultimate cost-effectiveness, associated with implementing the food supplementation interventions for diabetes involves a series of steps. The methods approved by the World Health Organization [44] and the US Guidelines for Cost-Effectiveness Analysis for Cost-Effectiveness Analysis [45] will be used to guide cost effectiveness analyses. From the provider and payer perspective, we will calculate the total cost of the food supplementation interventions per participant using the resource cost method which includes personnel, overhead, supplies, and equipment costs necessary to provide the food supplementation interventions for diabetes. From the payer perspective, we will estimate the cost of reimbursing the provider who delivers the food supplementation intervention or diabetes education component to the patient. This will include the $600 cost per participant for the food supplementation interventions whether combined, food box only, or food voucher only. From the patient perspective, we will calculate the patient’s time cost and any additional cost burden for the patient related to improving their diabetes self-care via self-funded food improvement activities. Effectiveness will be measured based on changes in HbA1c, Quality of Life (SF-12), Quality Adjusted Life Years (QALYs), health services utilization and costs, and patient productivity (working hours and income). Individuals will report health service utilization of primary care, emergency room, and inpatient hospital stays using standard questions from the Medical Expenditure Panel Survey (MEPS) as part of the follow-up survey.[46] The patient-level analyses of cost and outcomes described above will be aggregated to produce overall cost-effectiveness measures including incremental cost-effectiveness ratios. For the lifetime perspective, we will estimate gains in QALYs for alternative trajectories of the disease based on the EuroQoL [47], health services utilization and costs and patient productivity measured in hours worked and income [48].

Discussion

There are currently no large randomized trials indicating whether food supplementation is beneficial in improving glycemic control in food insecure African Americans with diabetes, or whether incorporating diabetes education with food supplementation is essential for improving health outcomes. This study will yield valuable insight currently lacking into how best to design and deliver diabetes interventions to low-income, food insecure, African Americans with type 2 diabetes that take into account both the clinical and social determinants of health. By 1) focusing on low-income, urban, African Americans with diabetes, 2) incorporating insight from food pantries and food insecure populations on barriers and facilitators to diabetes care, 3) answering a series of questions using a controlled randomized trial design essential to guide policy and program development for food insecure populations with diabetes, 4) measuring glycemic control as the primary outcome, and 5) designing the study to test separate and combined efficacy of food vouchers, food stock boxes, and diabetes education, this study will answer questions necessary to shift the field of food insecurity and diabetes to implementation of best practices at the community and health care policy level.

Trial Status

The study was funded on May 3, 2019. Study recruitment began in January 2020 but was suspended between March 16, 2020 and June 15, 2020 because of the COVID-19 pandemic. As of November 2, 2020, 65 patients have been enrolled and randomized. Of the number randomized, 27 have completed the 12-week phone interventions and the 3-month follow-up assessment.

Table 2:

Data Collection Measures

Outcome	Test	Measurement
Primary Outcome Measure	Hemoglobin A1c	Blood samples will be collected at baseline, 3, 6 and 12 months by trained phlebotomists or nurses
Secondary Outcome Measures	Resource Utilization/Cost	The perspective of cost will be that of the payer. Previously validated questions on resource utilization will be administered at baseline, 3, 6 and 12 months. The questionnaires capture information on hospitalizations, physician visits, and emergency room visits. [46]
	Quality of Life (SF – 12)	The SF-12 [29] is a valid and reliable instrument to measure functional status and reproduces 90% of the variance in PCS-36 and MCS-36 scores. This will be used to assess quality of life at baseline, 3, 6 and 12 months.
	Quality Adjusted Life Years (EQ-5D)	The EQ-5D is a validated measure to assess health status developed by the EuroQol group, an international team of researchers. This scale assesses health status across 5 dimensions using 26 items [30] and will be administered at baseline, 3, 6 and 12 months.
	Blood Pressure	Blood pressure readings will be obtained using automated BP monitors (OMRON IntelliSense™ HEM-907XL) at baseline, 3, 6 and 12 months. The device will be programmed to take 3 readings at 2-minute intervals and give an average of the 3 BP readings.
	LDL-Cholesterol	Blood samples will be collected at baseline, 3, 6 and 12 months by trained phlebotomists or nurses

List of abbreviations

BP

Blood Pressure

EQ-SD

Standardized measure of health-related quality of life

FFQ

Food Frequency Questionnaire

HbA1c

Hemoglobin A1c

IRB

Institutional Review Board

ITT

Intent-To-Treat

LDL

Low Density Lipoprotein

LSM

Least Squares Means

MCW

Medical College of Wisconsin

NCI

National Cancer Institute

PHQ

Patient Health Questionnaire

QALYs

Quality Adjusted Life Years

QOL

Quality of Life

RCT

Randomized Controlled Trial

SBP

Systolic Blood Pressure

SEM

Structural Equation Modeling

SF-12

Short Form - 12

SNAP

Supplemental Nutrition Assistance Program

T2DM

Type 2 Diabetes Mellitus

USA

United States of America

USDA

United States Department of Agriculture

WIC

Women, Infants, and Children

YMCA

Young Men’s Christian Association