Virtual coaching can work once you find out who, or what, is coaching you

Key takeaways

• In a 2025 Johns Hopkins trial, a fully automated coaching app matched human-led programs on the CDC’s diabetes prevention benchmark, 31.7% versus 31.9%, but participants rated the human programs as more satisfying.
• The CDC’s lifestyle change program rests on a 3,234-person federal trial in which lifestyle change cut the risk of type 2 diabetes by 58%.
• In CDC data on more than 330,000 participants, online programs produced less weight loss on average than in-person ones, but people who stayed for 22 sessions lost more than 5% of their weight in every format.
• Two widely cited app trials were paid for by the companies that sell the apps, and in one, fewer than half of the people assigned to the app finished the program.
• Before you pay, ask who the coach is, how they were trained, whether the program is CDC-recognized, how cancellation works and what happens to your data.

In a study published in JAMA in October 2025, Johns Hopkins researchers assigned 368 adults with prediabetes by chance to one of two kinds of coaching¹. One group got an app that ran without any human coach. A learning algorithm decided which tailored phone alerts each person received about weight, activity and diet¹.

The other group was referred to one of four remote, yearlong group programs led by human coaches¹.

After 12 months, the app held its own. In the app group, 31.7% met the CDC’s benchmark for diabetes prevention programs, compared with 31.9% in the human-led group¹. More app users started the program, and more finished it¹.

Then the team asked people how they felt about it. Satisfaction favored the human coaches in every area measured². Almost half of the app group, 46.5%, said they would have preferred a human coach².

That tension runs through virtual coaching. Programs by app, text and video can help people change habits. But “coach” can mean very different things, and the evidence depends on which one you get.

“Coach” can mean four very different things

Before you sign up, find out who, or what, is on the other end. This table sorts the main types.

Who is coaching you	What rules apply	What to check
A licensed clinician by video, such as a doctor, dietitian or therapist	Must be licensed or legally allowed to practice in the state where you are13	Their license in your state
A board-certified health coach	A private national credential with an exam developed with the National Board of Medical Examiners14	The credential, using the board’s online verification14
A CDC lifestyle coach	CDC recommends at least “12 hours or two days” of formal training, plus two hours of advanced training a year3	Whether the program is CDC-recognized
An app’s automated messages or AI	CDC doesn’t count chatbots or AI forums as coach contact in its online programs3	Whether a person ever reviews your plan

Credentials and coaching skill are covered in more depth in our article on health coaching. The key point here: a coach is not the same as a licensed health professional.

CDC’s standards set training requirements for lifestyle coaches, not licensing ones³. Medicare’s diabetes prevention program can use coaches who are employees, contractors or even volunteers¹².

That matters for what they can do. A coach can help you set goals, plan meals and stay accountable. Diagnosing a condition or adjusting medicine is a job for a licensed clinician.

The best-tested program started as a federal trial of 3,234 people

The strongest evidence for coaching comes from diabetes prevention. From 1996 to 2001, the National Institutes of Health ran the Diabetes Prevention Program at 27 U.S. centers⁵. It enrolled 3,234 people at high risk of type 2 diabetes⁵.

The lifestyle group aimed to lose 7% of their body weight and exercise 150 minutes a week. They met one-on-one with study staff at least 16 times in the first 24 weeks⁵.

The results were large. Over about three years, lifestyle change cut the risk of type 2 diabetes by 58%, and by 71% in people 60 and older⁵. The drug metformin cut it by 31%⁵. After 10 years, follow-up research found lifestyle change had still delayed diabetes by 34%⁵.

That’s worth pausing on if you value natural approaches. One of the best-proven prevention tools in modern medicine is better food, more movement and steady support from another person.

Today, the CDC’s National Diabetes Prevention Program offers a lifestyle change program that the agency says is proven to cut type 2 diabetes risk in half⁴. It runs for a year, meeting about weekly for six months and then once or twice a month⁴. It can be delivered in person, online, by live video or phone, or in a mix⁴.

Online versions work for people who stay

CDC researchers analyzed outcomes for 333,715 people who started the program between 2012 and 2018⁶. More of them used online programs than in-person ones⁶.

In-person participants attended an average of 15 sessions. Online participants attended about 13 and dropped out sooner⁶.

Average weight loss was 4.4% in person, 4.7% by live distance learning and 2.6% online⁶. The paper’s tables and text differ slightly, but roughly a third of in-person participants reached 5% weight loss, compared with about a fifth of online participants⁶.

Staying power changed the picture. Among people who reached 22 sessions, average weight loss topped 5% in every format⁶. The authors noted they couldn’t tell how virtual programs collected weights, and that motivation may differ between people who choose each format⁶.

A separate study of referrals in California, Florida and Colorado from 2015 to 2018 found the same pattern⁷. Online programs were better at getting referred people to enroll, 56% versus 51%⁷.

But by week 26, only 46% of online participants had attended at least nine sessions, versus 66% in person⁷. At one year, 23% of online participants had lost more than 5%, versus 35% in person⁷. Two authors worked for the referral company that supplied the data⁷.

Results also varied by who took part. In person, people 65 and older lost an average of 5.1%, while adults ages 18 to 34 lost 3.3%⁸.

Company trials: real results, paid for by the company

Some of the best-known virtual coaching programs are sold by companies that also fund their own research. That doesn’t make the findings wrong. It means you should read them the way you’d read a supplement company’s study.

Noom paid for a randomized trial of its coach-guided diabetes prevention app at Stony Brook Medicine on Long Island, starting in 2016⁹. Noom helped design the study, and two of the authors were Noom employees with stock options⁹.

The app group weighed 2.64 kilograms (about 5.8 pounds) less than the usual-care group at six months, and the gap held at a year⁹. But at six months, there was no difference in HbA1c, a blood test of average blood sugar⁹.

Engagement was the weak point. Of the people assigned to the app, 23 never downloaded it or activated their account. Of the 80 who did, 45 completed the program⁹.

Omada, another digital diabetes program, funded a randomized trial of 599 adults with prediabetes in the Omaha area, according to MobiHealthNews¹⁰. The early results came from a conference abstract presented at ObesityWeek 2020, not a peer-reviewed paper, and the company said journal publication was still pending¹⁰.

In those results, Omada users had lost 5.4% of their weight after a year. People who attended a single prevention class lost 2.0%¹⁰. Average HbA1c fell by 0.23 points versus 0.15¹⁰.

Both trials showed real but modest gains. Neither was compared against a full in-person CDC program, which is the comparison most buyers would want.

Human or algorithm? A rare head-to-head test

The Johns Hopkins trial is one of the few to pit an automated coach directly against human coaches¹. Federal research money from the National Institute of Diabetes and Digestive and Kidney Diseases paid for it, including participants’ program costs². The app company and the human programs were paid for their services, and the human programs didn’t see the overall results¹. One author and Johns Hopkins University receive royalties from a separate online diabetes prevention program, which the university reviewed under its conflict-of-interest rules¹.

The automated app had clear practical advantages. People started it sooner, a median of 11 days versus 26 days for the human programs². More people began it, 93.4% versus 82.7%, and more finished, 63.9% versus 50.3%¹.

But people liked the humans better. Satisfaction and acceptability scores favored the human programs in every area measured, with the biggest gap in overall satisfaction². The authors described a trade-off between reach and the quality of the experience².

The CDC has drawn its own line. In programs delivered fully online, a live coach must be available each week a participant is active. Two-way texts and emails count, but chatbots and AI forums do not³.

What the evidence shows

Type of virtual coaching	What studies found	Evidence level
CDC lifestyle program (the original design)	Cut diabetes risk 58% in a large federal trial5	Strong evidence
CDC-recognized online programs	Large real-world data: less average weight loss than in person, but more than 5% for people who stay 22 sessions67	Promising but early; few head-to-head trials, and dropout is the main risk
Company app-based coaching (Noom, Omada)	Modest weight loss in company-funded trials, one reported only as a conference abstract; mixed blood sugar results910	Promising but early
Fully automated AI coaching	Matched human programs on CDC goals in one trial; rated less satisfying12	Promising but early
Coaching that claims to diagnose or treat disease	Diagnosis and treatment require a licensed clinician13	Not enough evidence, and outside a coach’s role

What it costs, and who pays

Costs vary widely. The CDC says some lifestyle change programs are free, and employers or insurers may cover others⁴.

If you’re on Medicare and qualify, the Medicare Diabetes Prevention Program has no cost-sharing, though Medicare Advantage plans may set their own costs¹². You need Part B or a Medicare Advantage plan, no prior diabetes diagnosis and a body mass index of at least 25, or 23 if you identify as Asian¹². You also need a blood test in the prediabetes range within the past year, such as an A1C of 5.7% to 6.4%¹².

Medicare is expanding virtual options. Live distance learning is allowed through 2029. Starting in January 2026, Medicare is testing self-paced online delivery for four years to see whether results match other formats¹². In 2026, Medicare pays suppliers up to $624 per person for online delivery and up to $822 for in-person or live remote sessions¹².

Consumer apps can cost more than they first appear. In 2022, Noom agreed to a preliminary $62 million settlement of a class action, which needed a judge’s approval, Reuters reported: $56 million in cash and $6 million in subscription credits¹¹. Customers alleged that “risk-free” trials turned into hard-to-cancel renewals costing as much as $199¹¹. Noom denied wrongdoing¹¹.

Our article on wellness subscriptions explains how to protect yourself.

Red flags, and what to ask before you sign up

Federal rules apply to the claims, even when the coach isn’t licensed. The Federal Trade Commission’s 2022 guidance says health claims, including for apps, generally need randomized human trials behind them¹⁵. Customer testimonials don’t count as proof, and a “results not typical” disclaimer doesn’t fix a misleading success story¹⁵.

Watch for these warning signs:

• The coach offers a diagnosis or tells you to stop or change a medicine.
• The coach sells supplements or products and recommends them to you, without saying they profit.
• The program guarantees a specific amount of weight loss or a reversal of a disease.
• Success stories are the only evidence offered¹⁵.
• You can’t find out who the coaches are or how they were trained.
• Cancellation terms are hard to find, or a “risk-free” trial turns into a long paid plan¹¹.

Use these questions on a sales call or in a chat window:

1. “Is the person coaching me a licensed professional, a certified coach or software?”
2. “What training did my coach complete, and can I verify it?”³¹⁴
3. “Is this program recognized by the CDC?” Recognition means the program meets CDC’s standards and reports its results to the agency³. The CDC’s website also has a program finder.
4. “How often will I talk to a live person, and can I reach them between sessions?”³
5. “What published studies support this program, and who paid for them?”⁹¹⁰
6. “What does it cost after any trial period, and how do I cancel?”¹¹
7. “Who sees my health data, and can I delete it?”
8. “What happens if my blood sugar or blood pressure is out of range?”

The last question tests whether a program knows its limits. A good answer sends you to your clinician.

Questions readers ask

The bottom line

Virtual coaching works best when it follows a tested design, keeps you engaged and connects you with a person when you need one. The CDC’s diabetes prevention program has decades of evidence behind it, and people who stick with its online versions can reach the same 5% weight-loss goal. Company apps and AI coaches show promise, mostly in company-funded or single trials.

Before paying, find out who is coaching you, how they were trained, what the research shows, who paid for it and how to cancel. Let a coach help with habits, and keep diagnosis and medicine with a licensed clinician.

Related reading

• Health coaching is real, unregulated, and easy to get wrong
• Wellness apps help a little if you keep using them, and most people don’t
• Who paid for the study matters, even when the science is real
• The wellness subscription you meant to cancel
• What your provider is legally allowed to do

Sources

1. Johns Hopkins Medicine. “AI-Powered Diabetes Prevention Program Shows Similar Benefits to Those Led by People.” Johns Hopkins Medicine, October 2025 (on Mathioudakis N et al., JAMA, 2025). hopkinsmedicine.org
2. Lalani B, Eaton C, Zade D, Riekert K, Maruthur N, Dobs A, Mathioudakis N. “Patient engagement, acceptability, and preference of artificial intelligence versus human coaching for diabetes prevention.” npj Digital Medicine, 2026. nature.com
3. Centers for Disease Control and Prevention. “Diabetes Prevention Recognition Program: Standards and Operating Procedures.” CDC, June 2024. stacks.cdc.gov
4. Centers for Disease Control and Prevention. “About the Lifestyle Change Program.” CDC, 2024. cdc.gov
5. National Institute of Diabetes and Digestive and Kidney Diseases. “Diabetes Prevention Program (DPP).” NIH. niddk.nih.gov
6. Ely EK, Ng BP, Cannon MJ. “Delivering the National Diabetes Prevention Program: Assessment of Outcomes in In-Person and Virtual Organizations.” Journal of Diabetes Research, 2023. pmc.ncbi.nlm.nih.gov
7. Golovaty I, Wadhwa S, Fisher L, et al. “Reach, engagement and effectiveness of in-person and online lifestyle change programs to prevent diabetes.” BMC Public Health, 2021. link.springer.com
8. Ng BP, Ely E, Papali’i M, Cannon MJ. “Delivering the National Diabetes Prevention Program: Assessment of Retention, Physical Activity, and Weight Loss Outcomes by Participant Characteristics and Delivery Modes.” Journal of Diabetes Research, 2024. pubmed.ncbi.nlm.nih.gov
9. Toro-Ramos T, Michaelides A, Anton M, et al. “Mobile Delivery of the Diabetes Prevention Program in People With Prediabetes: Randomized Controlled Trial.” JMIR mHealth and uHealth, 2020. pubmed.ncbi.nlm.nih.gov
10. MobiHealthNews. “Omada reports blood sugar, weight improvements after 12 months of its digital diabetes prevention program.” MobiHealthNews, November 2020. mobihealthnews.com
11. Stempel J. “Noom diet app reaches $62 million settlement over automatic subscription renewals.” Reuters, via Euronews, February 2022. euronews.com
12. Centers for Medicare & Medicaid Services. “Medicare Diabetes Prevention Program Expanded Model.” MLN Booklet, CMS, February 2026. cms.gov
13. U.S. Department of Health and Human Services. “Getting started with licensure.” Telehealth.HHS.gov, 2024. telehealth.hhs.gov
14. National Board for Health & Wellness Coaching. “National Board Certified Health & Wellness Coaches.” NBHWC. nbhwc.org
15. Federal Trade Commission. “Health Products Compliance Guidance.” FTC, December 2022. ftc.gov