We conducted a single-arm explanatory sequential mixed methods (24) pilot study to examine the feasibility, acceptability, and preliminary effectiveness of a VLC-DPP among Veterans with prediabetes. We examined change in weight, hemoglobin A1c (HbA1c), and patient-reported measures of food cravings, stress eating, perceived health status, and motivation to prevent type 2 diabetes over the 12-month study period. Qualitative interviews were conducted at 6 months to explore participants’ experiences with the program and strengthen our understanding of quantitative findings (25). The study was approved by the VA Ann Arbor Healthcare System (VAAAHS) Institutional Review Board and conducted from October 2020 to September 2021. Trial registration: clinicaltrials.gov, NCT04881890. Registered May 11, 2021, https://clinicaltrials.gov/ct2/show/NCT04881890.

The intervention was intended to be delivered through in-person, group-based sessions. However, due to the COVID-19 pandemic, the intervention was delivered remotely using the VA Video Connect platform (26). VA Video Connect allows for secure virtual visits with Veterans and their care team; it can accommodate multiple simultaneous video feeds to allow for group interactions, and it can be accessed from any computer, tablet, or mobile device (e.g., Smartphone) with internet connection. To better facilitate group interaction on a virtual platform, we offered two small, closed groups rather than a single in-person group. At baseline, participants self-selected group participation on either Monday afternoons or Wednesday mornings.

The intervention was conducted among individuals who received primary care within the VAAAHS, which includes the Lieutenant Colonel Charles S. Kettles VA Medical Center (VAMC) and six other freestanding VA outpatient health clinics serving a diverse patient population throughout southeast Michigan and Toledo, Ohio (27). Study inclusion criteria were: (a) Overweight, defined as body mass index (BMI) ≥ 25 kg/m²; (b) prediabetes, defined as HbA1c ≥ 5.7% and ≤ 6.4% drawn within 6 months of the anticipated study start date; (c) willingness and ability to participate in group-based, online classes using video; (d) ability to engage in at least light physical activity; and (e) willingness to self-weigh at least once weekly and report these data to the study team. Individuals without access to video-enabled technology or a home scale were provided with a webcam and/or scale by the study team. Study exclusion criteria were: (a) history of type 1 diabetes; (b) use of anti-hyperglycemic medications other than metformin; (c) current participation in another lifestyle or behavior change program or research study; (d) use of anti-obesity medications; (e) adherence to a vegetarian or vegan dietary pattern; (f) inability to read, write, or speak English; (g) inability to provide informed consent; or (h) pregnant or intention to become pregnant during the intervention period.

Potentially eligible Veterans were identified using VA Health Services Research & Development Corporate Data Warehouse. Study invitation letters were sent to 167 individuals. The letters briefly described the intervention and provided individuals with study team contact information if they desired to enroll and/or learn more about the study. Approximately 2 weeks after letters were sent, study staff reached out to Veterans by telephone to further describe the study, answer questions, and screen interested Veterans for eligibility. Interested and eligible Veterans completed written informed consent, a baseline survey, and laboratory tests drawn at the Lieutenant Colonel Charles S. Kettles VAMC or another freestanding VA outpatient health clinic within the VAAAHS. For participants who chose to have laboratory tests obtained at the Lieutenant Colonel Charles S. Kettles VAMC, a study team member (CH) met participants to complete written informed consent and the baseline survey prior to participants’ blood draw. For participants who chose to have laboratory tests obtained at another VA outpatient health clinic, written informed consent and the baseline survey were sent by postal mail and completed prior to laboratory testing; a study team member (CH) was available by phone to answer questions during the consent process.

Our team previously developed and pilot tested a very low-carbohydrate adaptation of the CDC’s NDPP among a non-Veteran population (23). The CDC’s NDPP is similar to MOVE! in structure and content, and prior work has demonstrated similar 12-month weight change outcomes (8). Like MOVE!, the NDPP teaches participants to follow a low-fat, calorie-restricted eating pattern and to engage in at least 150 min of moderate intensity physical activity per week with the goal of achieving and maintaining 5 to 7% body weight loss (28).

A detailed summary of VLC-DPP intervention components is reported using the Template for Intervention Description and Replication (TIDieR) checklist in Appendix 1 (29). In brief, we substantially adapted the content of the NDPP’s dietary change sessions to teach participants to follow a very low-carbohydrate eating pattern rather than a low-fat, calorie-restricted eating pattern. We minimally altered NDPP sessions focused on other topics such as physical activity or behavior change. Appendix 2 shows NDPP and VLC-DPP session topics and denotes sessions with substantial versus minimal content modifications. Both programs consist of 16-weekly sessions delivered over 6 months (i.e., core phase) followed by 7 bimonthly or monthly sessions (i.e., maintenance phase) (30).

VLC-DPP participants were taught to consume 20–35 grams of net carbohydrates per day, defined as total grams of carbohydrate minus grams of dietary fiber. Participants were not provided with an explicit calorie limit or protein intake goal. Rather, participants were encouraged to eat when hungry and stop when satisfied, and they were instructed to keep protein intake similar to baseline. Participants were taught to change one meal per week from a high- to a very low-carbohydrate option, starting with breakfast and aiming for approximately 10 grams of net carbohydrates per meal or less. Participants were encouraged to consume foods such as meat, fish, poultry, eggs, cheese, seeds, nuts, leafy greens, non-starchy vegetables, and some fruits (e.g., berries). Participants who preferred plant-based sources of protein were encouraged to consume tofu and tempeh. Participants were taught to use low-carbohydrate substitutes in place of common high carbohydrate foods (e.g., cauliflower rice in place of grain rice).

Participants were encouraged to explore a variety of low-carbohydrate foods that appealed to their preferences and budget constraints. Participants were not provided with specific meal plans but rather provided with a variety of low-carbohydrate alternatives to high-carbohydrate foods and meals. For example, low-carbohydrate breakfast options included scrambled eggs, an omelet with broccoli and cheese, full-fat unsweetened Greek yogurt with pecans, a crustless quiche, low-carbohydrate pancakes, waffles or muffins, or a low-carbohydrate shake. Participants were taught to search for recipes online using search teams like “low-carb pancakes” and reviewing nutrition facts to select options with approximately 5 grams of net carbohydrates per serving. Participants were encouraged to share their favorite low-carbohydrate foods and meals with classmates during group sessions.

During the maintenance phase, participants were advised that they could maintain their very low-carbohydrate eating pattern or they could gradually liberalize their daily carbohydrate intake. Participants desiring to add carbohydrates to their eating plan were advised to increase their daily carbohydrate intake by 5 grams of net carbohydrates and maintain that change for at least 1 week. For example, individuals consuming 30 grams of net carbohydrate per day during the core phase were advised to increase to 35 grams of net carbohydrate per day and maintain that level for at least 1 week before making another change. Participants were advised to continue to self-weigh and to use increases in body weight, hunger, or food cravings as an indication that the carbohydrate intake was too high. In contrast, weight stability or continued weight loss without excessive hunger or cravings would indicate a well-tolerated carbohydrate level.

The intervention was delivered by a VA Registered Dietician with certification in diabetes education. She was trained to deliver low-carbohydrate education through self-guided review of the program curriculum and selected online resources (31). A primary care provider (PCP) with training in obesity medicine and experience with low-carbohydrate nutrition counseling (the principal investigator: DHG) provided additional training, as needed, and was present at all sessions to assist with answering participants’ questions, if necessary.

PCPs were notified of their patient’s participation in the study through electronic health record messaging. PCPs were also notified of participants’ weight changes and laboratory results at 6 and 12 months. Additional communication with PCPs occurred as needed. For example, when participants occasionally disclosed medical or psychosocial needs to a study team member, this information was communicated to PCPs to facilitate appropriate care for individual participants.

Potentially eligible study participants were identified by HbA1c within the prediabetes range drawn within 6 months of the program’s anticipated start date. After providing written informed consent, individuals were asked to complete baseline laboratory and survey measures. Laboratory measures included HbA1c and fasting lipids. Individuals were not excluded from study participation if their baseline HbA1c was outside the prediabetes range (<5.7% or > 6.4%) provided they met other study criteria. These laboratory measures were repeated at 6 and 12 months.

At baseline, 6, and 12 months, participants were invited to complete a paper survey. A study team member (CH) met participants to complete surveys on the day of laboratory testing, if testing was obtained at the Lieutenant Colonel Charles S. Kettles VAMC. If laboratory tests were obtained at another VA outpatient health clinic, surveys were sent by postal mail. Survey responses were subsequently entered into REDCap (32) by a study team member. At baseline, participants were asked to report demographic and socio-economic information. At each time point, we used validated survey measures to assess key patient-centered outcomes. Specifically, food cravings were assessed using the 8-item Control of Eating Questionnaire subscales for savory and sweet food cravings (33). Stress eating was assessed using the 4-item coping subscale of the Palatable Eating Motives Scale (34), and a 2-item assessment of participants’ usual response to a stressful event (35). Patient-reported physical and mental health was assessed using the 10-item Patient-Reported Outcomes Measurement Information System Global Health survey (36). Sources of motivation to prevent type 2 diabetes (i.e., intrinsic vs. extrinsic) were assessed using the 15-item Treatment Self-Regulation Questionnaire (37). Survey items, response options, and scoring details are shown in Appendix 3.

To evaluate possible side effects from a very low carbohydrate diet, we also assessed self-reported physical symptoms, including bad breath, abdominal bloating, nausea, constipation, diarrhea, dizziness, headaches, foot pain or numbness, acid reflux (i.e., heartburn), joint or muscle pain, and muscle cramps; response options were: “not at all,” “1 day a week,” “2–3 days a week,” “4–5 days a week” and “6–7 days a week.” All participants were invited to participate in a semi-structured interview at 6 months to explore their experiences with the program and solicit feedback for improvement. Veterans were compensated with a $25 gift card for completing each study measure at 6 months and at 12 months.

We initially planned to measure participants’ weight prior to each session. However, due to the COVID-19 pandemic, we were required to deliver the program virtually and collect self-reported weight, which has previously demonstrated concordance with measured weight across diverse populations (38). Participants were instructed to self-weigh at least once per week using a home scale. A study team member called all participants approximately 1 h prior to each session to collect self-reported weight data. Participants who could not be reached prior to the session were contacted within 24 h following the session; a total of 3 attempts were made to reach each participant per session. Participants were asked to self-report weight data even when they could not attend that day’s session.

Consistent with guidelines and recommendations for conducting pilot studies, (39, 40) our sample size was selected to provide sufficient data regarding the feasibility and acceptability of the intervention and the methods and procedures anticipated for use in a larger scale trial (41). Specifically, we aimed to understand whether we could (1) recruit the target population, (2) retain participants in the study, (3) deliver the group-based intervention via a virtual platform, and (4) collect data at prespecified timepoints. Based on our prior pilot evaluation of the VLC-DPP among a non-Veteran population (23) and our subsequent efficacy trial informed by these pilot data (42), we considered an enrollment target of 22 total participants sufficient to provide feasibility and acceptability data. Participants were divided between 2 virtual groups, as this is the maximum number of participants typically enrolled in VA virtual weight management programs.

Demographic and socioeconomic characteristics were assessed at baseline (Table 1). Thirteen (72%) participants were men, approximately half (56%) of participants were white, and the average age was 59 years. At baseline, mean BMI was 36 kg/m2 and mean HbA1c level was 6.1%. Four participants progressed from prediabetes to type 2 diabetes prior to the start of the study period.

Intervention uptake is shown in Figure 1.

Mean session attendance: Participants attended a mean (SD) of 12.4 (2.9) of 16 core sessions and 3.6 (2.4) of 8 maintenance sessions. Sixteen participants attended ≥8 core sessions.

Study retention: Seventeen participants completed the 6-month survey and 15 completed the 12-month survey, resulting in 94.4 and 83.3% retention at each respective timepoint.

Adherence with self-weighing: Participants self-reported weight data for a mean (SD) of 14.5 (1.95) of 16 core sessions and 5.06 (2.37) of 8 maintenance sessions.

Table 2 shows weight outcomes at 6 and 12 months compared to baseline. On average, participants lost 11 kilograms at 12 months, corresponding with a mean percent weight loss of 9.45%. At 12 months, 9 participants (50%) achieved ≥5% weight loss and 7 (38.9%) achieved ≥10% weight loss. All participants with ≥5% weight loss at 12 months had achieved ≥4% weight loss by 6 months. Figure 2 shows individual participants’ weight change at 6 and 12 months compared to baseline.

Table 3 shows change in HbA1c and lipids at 6 and 12 months compared to baseline. Change in HbA1c from baseline was not statistically significant at 6 months but was at 12 months (value of p < 0.01). Of the four participants with HbA1c ≥ 6.5% at baseline, 3 (75%) had an HbA1c <6.5% at 12-months, with two of these participants achieving ≥10% weight loss and one achieving nearly 5% weight loss. Two participants progressed from prediabetes to type 2 diabetes during the study period; one achieved <5% weight loss at 12 months and the other achieved approximately 5% weight loss at 12 months.

Sixteen participants participated in 6-month interviews. The baseline characteristics of interviewees are shown in Table 1. Mean weight change among interviewees was 8% at 6 months and 9% at 12 months and 8 (50%) achieved ≥5% weight loss at 12 months. Among interviewees who achieved ≥5% weight loss, all reported consistent adherence to the very low-carbohydrate eating pattern, which was facilitated by (a) enjoyment of low-carbohydrate foods; (b) careful monitoring of carbohydrate intake through food tracking and nutrition label reading; and (c) reduced hunger and food cravings. Representative quotes are shown in Table 5.

Among interviewees who achieved <5% weight loss at 12 months (n = 6), all reported factors that hindered adherence to the very low-carbohydrate eating pattern, including (a) food cravings, particularly for sweets, (b) challenges with maintaining a food log, and (c) difficulty with meal planning. Representative quotes are shown in Table 6.

Interviewees who achieved ≥5% weight loss at 12 months also discussed health benefits beyond weight loss, including a reduction in symptoms of acid reflux, increased energy, and improved glycemic control. One participant noted, “when I first started…I had heartburn, no matter how much or how little I ate, no matter when I ate…I was living off of antacids…I thought maybe it was because I was eating so much or I was eating a particular food but it was, it had to have been the sugar because since I started doing the low-carb I have not had heartburn at all. I have four bottles now of antacids that are sitting in my medicine cabinets and in my medicine closet and I do not even use them anymore (Participant 8).” Another participant noted, “I feel that this program has given me a new lease on life. I’m a lot more positive, I have a lot more energy, I feel better about myself, I have, you know all the way around I think I’m a better person because of this program (Participant 15).”

Regardless of weight change, participants acknowledged challenges with navigating the high-carbohydrate food environment, particularly when “you cannot control the meal that’s being served…if you get out on the run and have to get something to eat or if you are invited as a guest to an event or someone’s house…(Participant 17).” Two interviewees noted higher food costs when following a low-carbohydrate meal plan but did not feel that this was a barrier to dietary adherence. Three interview participants reported stable food costs and three reported a reduction in food costs. One interview participant noted a higher cost of individual low-carbohydrate food items but a reduction in total spending on food, “because I’m eating less. By choosing healthier low-carb options such as vegetables, meats things like that, I’m staying fuller, I feel like I’m staying fuller longer off one meal versus when I was eating processed snacks and you know fast food and quick food items…I’m not buying so much junk food that you eat it and then you are hungry, you know a couple minutes later (Participant 17).”