Community Servings: Food as Medicine for Diabetes

a. Historical background The Center for Disease Control and Prevention (CDC) estimates that
there are currently 29 million people with diabetes and 86 million people with pre-diabetes
in the U.S. One in 10 Americans has diabetes now, and, if current trends continue, 1 in 3
Americans will have diabetes by 2050. This chronic disease significantly impacts both quality
of life and rapidly rising national healthcare costs. The estimated cost of diabetes in the
U.S. in 2014 was $265 billion with $176 billion in direct medical costs and $89 billion is
indirect medical costs (disability, work loss, premature mortality). Medical expenses for
people with diabetes are 2.3 times higher than for people without diabetes.

Food insecurity, defined as limited access to nutritious food due to cost, has been
associated with increased prevalence of diabetes and worse diabetes control. Food insecurity
may worsen diabetes by decreasing consumption of fresh fruits and vegetables and increasing
consumption of inexpensive, calorie-dense food, and which in turn leads to greater Hemoglobin
A1c, an indicator of hyperglycemia, over time.

c. Rationale behind the proposed research, and potential benefits to participants and/or
society

Approximately 20% of diabetes patients report food insecurity, a number that increases to
over 25% among those with the worst metabolic control.5 The prevalence of food insecurity is
also 20% in the MGH Population we surveyed (data not yet published). Hyperglycemia is
particularly responsive to dietary changes,8 yet few interventions have attempted to address
food insecurity in diabetes care. Prior studies have examined the impact of the Supplemental
Nutrition Assistance Program (SNAP, formerly the Food Stamp Program), but have not found
important improvements in diabetes outcomes for participants9. This may be because
neighborhood access to produce and other high quality food is low for many SNAP participants,
or because making healthy food choices is difficult in resource-constrained environments.
Additionally, recent sociological work has shown that expecting low-income women to cook
healthy meals for their families induces a significant burden, and the burden of these
expectations may drive less healthy food choices. Additionally, while significant time is
needed for healthy food preparation, low-income patients often face limited leisure time, and
multiple competing demands for both time and financial resources. Alternatively, direct
provision of healthy foods was incidentally noted to improve diabetes outcomes in a prior
randomized controlled trial, but this study was not conducted with the goal of addressing
food insecurity.

In this study, we propose to test whether home delivery of freshly prepared meals
specifically tailored to the needs of diabetes patients improves their dietary quality. We
hypothesize that the delivery of the meals will help them eat more healthily and improve the
food security of participants. Secondary outcomes in this pilot study will be weight and
metabolic control, along with psychological aspects of diabetes care.

Aim 1: To evaluate the effectiveness of receiving Community Servings meals on dietary quality
for food insecure diabetes patients with severe hyperglycemia (HbA1c > 8.0%) H1. Primary
outcome. Healthy Eating Index 2010 (HEI) score: We hypothesize that the CS group will
demonstrate greater improvements in dietary quality, as assessed by HEI score, at 12 weeks,
compared with usual care. The sample size of 50 provides 80% power to detect a 5 point
difference between the CS and usual care groups, assuming an 11 point standard deviation and
accounting for a 10% drop-out rate.

H1b. Secondary exploratory outcomes. Medical outcomes: We hypothesize that compared with
usual care, CS group participants will improve HbA1c, blood pressure, weight, and lipids from
baseline at the end of the intervention.

H1c. Behavioral and psychosocial outcomes: Because meal provision will reduce stress related
to procuring healthy meals, and free up household resources that would otherwise be spent on
food, we hypothesize that compared with usual care, the CS groups will have greater
improvements from baseline in patient-reported outcomes of diabetes distress and material
need security.

Aim 2: To evaluate the feasibility of providing meals and patient experience with the CS
program, particularly focusing on factors that determine acceptability, continuation, and
scalability We will use a mixed methods approach using participant structured interviews and
surveys to assess engagement and satisfaction with the program, and participant interviews or
focus groups to compare responders and non-responders. We will also collect quantitative
indicators of feasibility and implementation such as percent of meals delivered and consumed,
enrollment and persistence with the program, and logistical issues in order to plan for a
future full-scale intervention.

Inclusion Criteria:

- • Diagnosis of type 2 diabetes

- Age 18 years or older

- HbA1c level >8.0%

- Report food insecurity as indicated by the 2-item USDA Food Security Survey
Module13

- Willing to commit to random assignment to either receive CS meals immediately or
as a waitlist control

- Stable health, with no severe medical comorbidities that might interfere with
their ability to participate in the intervention, such as severe psychiatric
illness or imminent hospitalization

- Be willing to keep a food diary

- Be willing to attend and complete a baseline, 12 week, and 24 week assessment at
MGH

- Be able to understand and communicate effectively in English

- Have a primary care physician within the MGH practice based research network

- Live in an area where Community Servings can deliver meals

- Ability to store and prepare Community Servings meals

Exclusion Criteria:

- • Must not be pregnant or planning pregnancy in the next year

- Currently enrolled in another diabetes study Food allergy that would prohibit
consumption of meals

- Receiving episodic treatments that may increase blood glucose levels (e.g.
prednisone)