Effectiveness of a Ketogenic Diet and Virtual Coaching Intervention for Patients with Diabetes

The health and wellness industry is big business. In 2017, estimated global expenditures totaled $4.5 trillion, $702 billion of which were spent on “healthy eating, nutrition, and weight loss.” In the United States alone, the diet and weight loss market is valued at $78 billion annually. This investment simultaneously reveals two truths — Americans want to lose weight and Americans need to lose weight.

The harmful effects of being overweight or obese are well-documented and many. One potential consequence is Type 2 diabetes (T2D). In 2018, nearly 33 million Americans had T2D, and diabetes was the seventh leading cause of death. The economic impact of diabetes is similarly grim — a quarter of all health care expenditures go to diabetes-related costs.

Conventional wisdom holds that the primary intervention for managing diabetes is to achieve and maintain a healthy weight through diet and exercise. One diet that has gained traction in recent years and, in response, has spawned numerous weight loss products and services, is the ketogenic (keto) diet. Keto diets emphasize the severe restriction of carbohydrates and increased consumption of fats for the purpose of initiating ketosis, a metabolic state in which the body burns fat as its primary energy source.

Historically, keto diets have been used to manage a variety of health conditions, most notably epilepsy. More recently, there has been burgeoning interest in whether keto diets may be beneficial to other chronic diseases, in particular diabetes, including at the Department of Veterans Affairs (VA). The prevalence of diabetes within the Veteran population is disproportionately high compared to the general population (25% vs. 9%, respectively) and VA has long been a pioneer in diabetes research. Further inquiry is undoubtedly needed – diet research is notoriously difficult to conduct and existing research into keto diets is relatively limited, especially long-term studies.

New Research

In a recent study published in Diabetes, Obesity and Metabolism, a research team from VA and Boston University School of Public Health (BUSPH) evaluated the effectiveness of a keto diet and virtual coaching intervention on treating diabetes.

(Affiliations of the authors for this study include Kiersten L. Strombotne, PhD, Nambi J. Ndugga, MPH, Steven D. Pizer, PhD, and Austin B. Frakt, PhD, of VA Boston Healthcare System (VABHS) and BUSPH, Department of Health Law, Policy & Management; Jessica Lum, MA, VABHS; Paul R. Conlin, MD, VABHS and Harvard Medical School, and Anne E. Utech, PhD, Veterans Health Administration.)

This evaluation was developed in response to a Virta Health virtual diabetes coaching intervention pilot within VA. Virta Health, a private sector digital health company based in San Francisco, provides a diabetes intervention that emphasizes strict adherence to a ketogenic diet, health data monitoring, and personalized coaching. Beginning in April 2019, a sample of VA patients (n=454) were given access to this pilot program, cost free, on a first-come-first-served basis. To be included, participants were required to be actively enrolled in VA medical benefits, have a diabetes diagnosis, and be prescribed at least one diabetes medication. An additional 867 patients expressed interest in participating but were instead placed on a waitlist due to capacity constraints. The aim of the evaluation was to ascertain what impact the intervention had on metabolic health and regulation, diabetes-related health care utilization, and the use and costs associated with prescription drugs.

The authors employed a difference-in-differences approach that allows for a pre- and post-intervention comparison between the treatment group (those in the Virta program) and the control group (those on the treatment waitlist). The authors measured changes in observational data for the following outcomes: HbA1c, body mass index, blood pressure, emergency department encounters, outpatient visits, hospitalizations, insulin prescriptions, medication prescriptions, and costs associated with prescription medication. Outcome data were collected at five months post-intervention and the two groups were compared in parallel.


When comparing the treatment group to those in the control group, the authors’ analyses recorded several improved outcomes. First, participants in the treatment group saw significant improvements in both their HbA1c and body mass index during the study period. Significant reductions were also observed relative to insulin and medication prescriptions. On average, the decrease in monthly prescription use resulted in average savings of $34.54 per patient. The analysis also found a small reduction in monthly outpatient visits but no significant changes in emergency department encounters or hospitalizations.

This study was not without its limitations. One is that, although the waitlist control group allowed the researchers to control for similarities in motivation, the study was not randomized. The study design also limited the authors’ ability to monitor the two groups. Specifically, it was not possible to know how closely individuals in the treatment group followed the treatment protocol or whether individuals in the control group attempted or sustained dietary changes because these data were not captured in VA electronic health records. Therefore, it was unclear whether the improvements were attributable to adherence to the diet, personalized coaching, individual motivation, or a combination of these factors. Lastly, given that the study only looked at five months post-intervention, the authors were unsure whether the improved outcomes would be sustained over time.


Casting aside all other reasons to shed a few pounds, diabetes alone demonstrates that there is both a clinical and economic imperative to promote sustained weight loss. The favorable results of this study suggest that a ketogenic diet may be a useful part of a treatment regimen for diabetes. However, there is not enough evidence to say that a keto diet by itself leads to improved outcomes. Other studies of virtual diabetes interventions have found comparable effects, regardless of diet type. Additional research is needed to understand the extent that other factors (e.g., intensive coaching, patient motivation) contribute to improved outcomes.

Further, the popularization of keto diets is not without its controversy. For starters, keto may not be right for every person, and may in fact be harmful for patients with certain clinical profiles. Medical professionals caution that keto diets should be undertaken in consultation with a treating physician. Given the rising profile of keto diets and the increasing market penetration of keto-related products and services, further evaluation and research as to its efficacy and risks is needed.

Hidden information below


Email Address*