Prediabetes: Independent Evidence on Diagnosis and Prevention

Table of contents

• Evidence summary
• What prediabetes is, and why the definitions differ
• Symptoms, diagnosis and red flags
• Causes, risk factors and what happens next
• Treatments that work
• Medicines
• Supplements and lifestyle claims
• Risks of common treatments
• Who to see
• Follow the money: who funds the research and who profits
• Related research
• Frequently asked questions
• Sources and funding notes

Evidence summary

Independent evidence and credibility scorecard

Independence tiers: 1 = regulator or government body; 2 = academic or Cochrane team with no declared commercial funding for the work; 3 = independent authors with minor declared industry ties, or a body evaluating or promoting its own programme or definition; 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D.

Note the clustering: 11 of the 12 sources are based in the US or UK, and the founding trial is American. The only large long-term trial showing fewer deaths, Da Qing, comes from China. The evidence base is therefore shaped mostly by two health systems with very different thresholds. The strongest point of agreement is between sources with opposing interests: a sceptical BMJ critique, a WHO-funded Cochrane review and the ADA-linked DPP investigators all agree that lifestyle change delays diabetes, and that proof of fewer heart attacks is lacking.

What prediabetes is, and why the definitions differ

Prediabetes means that blood glucose is higher than normal but not high enough to diagnose diabetes. The NHS describes it as a state that puts you at higher risk of getting diabetes in the future (NHS). Clinicians also call it intermediate hyperglycaemia, impaired fasting glucose (IFG), impaired glucose tolerance (IGT) or, in England, non-diabetic hyperglycaemia (NDH).

There are three ways to measure it, and the bodies that set thresholds disagree on where to draw the lines.

Sources: ADA Standards of Care 2026, Table 2.2; NICE PH38; WHO thresholds as summarised in the Cochrane review and the Huang 2016 BMJ meta-analysis. The underlying WHO report is Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia (WHO and the International Diabetes Federation).

Why the thresholds matter

The Cochrane authors point out that the ADA lowered its fasting glucose threshold to 5.6 mmol/L in 2003. This "substantially increased" the number of people with impaired fasting glucose. The ADA then lowered the HbA1c threshold from 6.0% to 5.7% in 2010 (Richter 2018). In a BMJ analysis, John Yudkin and Victor Montori used Chinese survey data to show the effect. Impaired glucose tolerance alone covered 8.3% of adults, while the combined ADA criteria covered about half of all Chinese adults (Yudkin & Montori 2014). They argue that the cut-offs mark no biological turning point. Risk rises gradually, and the three tests identify different people.

The numbers are now large on both sides of the Atlantic. The CDC currently estimates that 115.2 million US adults have prediabetes, including 31.3 million people aged 65 or over (CDC). In England, 3,615,330 people registered with a GP had a recorded diagnosis of non-diabetic hyperglycaemia in 2023, up from 3,065,825 in 2022 (NHS figures from the National Diabetes Audit). Part of the difference reflects England's stricter HbA1c threshold.

The overdiagnosis debate

Critics say the label turns a risk factor into a condition. Yudkin and Montori argue that more than half of the people identified will be free of diabetes 10 years later, and that public health measures would do more than labelling individuals (BMJ 2014). Data on older adults support that concern. In the US ARIC study, 73% of adults with an average age of 76 met either the HbA1c or fasting glucose criterion. Over about 5 years, only 9% of those with a raised HbA1c developed diabetes, while 13% returned to normal and 19% died (Rooney 2021). The authors concluded that prediabetes may not be a robust diagnosis in older age.

Supporters reply that prediabetes is linked to more than diabetes. A meta-analysis of 53 cohort studies and 1.6 million people found modestly higher cardiovascular risk. The relative risk was 1.13 for impaired fasting glucose by the ADA definition and 1.30 for impaired glucose tolerance (Huang 2016). That is an association, though, not proof that lowering glucose at this stage prevents heart disease.

Symptoms, diagnosis and red flags

Most people with prediabetes have no symptoms. The NHS says most do not have the symptoms of type 2 diabetes, such as passing more urine, tiredness and unplanned weight loss (NHS). It is usually found on a routine blood test, an NHS Health Check (offered in England to people aged 40 to 74 without existing conditions), or testing because of risk factors.

How it is diagnosed

NICE recommends a two-stage approach in England: first a risk score, then a blood test (fasting glucose or HbA1c) for those at high risk (NICE PH38). In the US, the Preventive Services Task Force recommends screening adults aged 35 to 70 who have overweight or obesity (USPSTF 2021). The ADA goes further. It advises testing at any age for adults with overweight plus one risk factor, and for everyone from age 35 (ADA 2026).

The tests are imperfect. Measured against the oral glucose tolerance test, HbA1c picked up only about half of people with prediabetes (sensitivity 0.49), and fasting glucose about a quarter (0.25) (Barry 2017). The authors warned that many people will be wrongly labelled, while others will be falsely reassured. Because levels fluctuate, a single borderline result is best confirmed and then rechecked. The ADA advises monitoring at least once a year (ADA Section 3).

Red flags: when it may already be diabetes

• See a GP soon if you develop thirst, frequent urination, unexplained weight loss, marked tiredness, blurred vision, slow-healing cuts, or genital itching or repeated thrush (NHS).
• Get emergency help if those symptoms come with vomiting, stomach pain, deep breathing, fruity breath, drowsiness or confusion. These are signs of diabetic ketoacidosis (NHS).
• Rapid weight loss and thirst in a younger, slim person may point to type 1 diabetes rather than prediabetes. Type 1 tends to develop more quickly (NHS).

Causes, risk factors and what happens next

Prediabetes reflects two problems that build up over years: the body responds less to insulin (insulin resistance), and the insulin-making beta cells in the pancreas start to fail. A Lancet review notes that both are present before glucose changes can be detected (Tabák 2012).

The NHS lists the main risk factors for type 2 diabetes as being white and over 40, or over 25 and of South Asian, Chinese, Black African or Black Caribbean background. Others are a parent or sibling with diabetes, overweight with low activity, and conditions such as high blood pressure (NHS). The ADA also highlights previous gestational diabetes, which is one group it singles out for metformin (ADA Section 3). Statins slightly raise the risk of diabetes. The ADA advises monitoring glucose in people taking them, but says statins should not be stopped because of this (ADA Section 3).

How many progress?

The Lancet review estimated that 5–10% of people with prediabetes progress to diabetes each year, and roughly the same share return to normal glucose (Tabák 2012). Risk depends heavily on the definition. Combined impaired fasting glucose and impaired glucose tolerance carries the highest risk. An HbA1c of 6.0–6.4% carries more risk than 5.7–5.9% (Richter 2018). The Cochrane authors advised caution about "any active intervention" for people diagnosed on these grounds, because glucose status can move in both directions even after years.

Treatments that work

Every major guideline puts structured lifestyle change first. NICE recommends intensive lifestyle-change programmes for people with fasting glucose of 5.5–6.9 mmol/L or HbA1c of 42–47 mmol/mol (NICE PH38). The ADA gives its top grade (A) to referral to a programme aiming at 5–7% weight loss and at least 150 minutes a week of moderate activity. It gives grade B to certified digital programmes delivered through apps, websites or telehealth (ADA Section 3).

The Diabetes Prevention Program (US)

The DPP randomised 3,234 adults to placebo, metformin 850 mg twice daily, or an intensive lifestyle programme. The lifestyle goals were at least 7% weight loss and 150 minutes of activity a week. Participants had both raised fasting glucose and impaired glucose tolerance, with an average BMI of 34. Diabetes developed in 11.0, 7.8 and 4.8 people per 100 person-years on placebo, metformin and lifestyle respectively (DPP 2002). This was a selected, very high-risk group. Applying the result to everyone with an HbA1c of 5.7% stretches it.

After the trial, everyone was offered group lifestyle sessions, and follow-up continued in the DPP Outcomes Study. At 15 years, 62% of the original placebo group had diabetes, compared with 56% on metformin and 55% on lifestyle. Overall, rates of eye, kidney and nerve complications did not differ between groups. Women in the lifestyle group had fewer complications, and people who stayed free of diabetes had 28% fewer complications (DPPOS 2015). After a median of 21 years, neither intervention reduced major cardiovascular events (Goldberg 2022). The authors suggest that statin use, lifestyle sessions for all and people switching treatments may have diluted any effect. A 2025 re-analysis that allowed for these switches still found no clear metformin effect on any outcome except diabetes itself (Knowler 2025).

Finland and China

The Finnish Diabetes Prevention Study (522 adults with impaired glucose tolerance) also found a 58% reduction with individual counselling. Diabetes developed in 11% vs 23% after four years (Tuomilehto 2001). In the Da Qing study in China, 6 years of diet and exercise in 576 adults delayed diabetes by a median of almost 4 years. After 30 years, the intervention groups had fewer cardiovascular deaths (HR 0.67) and deaths from all causes (HR 0.74), and lived on average 1.44 years longer (Gong 2019). This is the strongest signal for hard outcomes, but it comes from one small cluster-randomised trial with a lean population that differs from most Western patients.

The NHS Diabetes Prevention Programme (England)

The NHS DPP began in 2016. It offered at least 13 group sessions over at least 9 months, totalling at least 16 hours of contact. By December 2018, 324,699 people had been referred and 96,442 had attended at least one group session. Of the 32,665 who started the sessions and had enough time to finish, 17,252 completed the programme. Completers lost an average of 3.3 kg, and HbA1c fell by 2.04 mmol/mol (Valabhji 2020). That means fewer than one in five people referred finished the course.

Independent NIHR-funded evaluations found benefit at scale. Areas that adopted the programme early saw about 6% fewer new cases than expected (McManus 2022). A matched study of 18,470 referred patients found a 20% lower conversion rate. At 36 months, 87.3% of referred people were still diabetes-free, compared with 84.6% of those not referred (Ravindrarajah 2023). An NHS England-led analysis found programme completers had about half the odds of diabetes at 24 months compared with non-attenders. The authors said they could not rule out bias, because people who complete tend to differ from those who do not (Barron 2025).

Medicines

Metformin

Metformin is the only medicine both NICE and the ADA recommend for prediabetes. NICE says to "use clinical judgement" on offering it when blood results worsen despite an intensive lifestyle programme, or when a person cannot take part, particularly with a BMI over 35. It notes that some metformin products are now licensed for this use after 3 to 6 months of lifestyle change (NICE PH38). NICE also tells clinicians to explain that metformin may be needed for life, and that long-term lifestyle change can be more effective than drugs.

The ADA says to consider metformin especially for people aged 25–59 with a BMI of 35 or more, higher fasting glucose (for example 6.1 mmol/L or above) or a higher HbA1c (for example 6.0% or above), and for women with previous gestational diabetes (ADA Section 3, grade A). The DPP dose was 850 mg twice daily (DPP). That is the dose studied, not a recommendation. The Cochrane review found metformin roughly halves diabetes risk compared with standard advice, but is no better than intensive diet and exercise. It found data on deaths, heart disease and quality of life "sparse or missing" (Madsen 2019). Yudkin and Montori argue that metformin delayed diabetes by around two years rather than preventing it (BMJ 2014). Some of the effect may be lower glucose readings while the drug is being taken.

Weight-loss medicines (GLP-1 and related drugs)

Newer injectable weight-loss drugs have large effects on glucose in industry trials. In a Lilly-funded 3-year trial in 1,032 people with obesity and prediabetes, tirzepatide cut weight by 12–20%. Diabetes developed in 1.3% vs 13.3% on placebo, and the gap stayed wide 17 weeks after stopping (2.4% vs 13.7%) (Jastreboff 2025). In pooled semaglutide trials co-authored by Novo Nordisk staff, 84–90% of people with prediabetes had normal glucose after 68 weeks, against 48–70% on placebo (Perreault 2022). These drugs are prescribed for obesity. NICE PH38 does not cover them. The ADA's prevention section does not name these drugs in any graded recommendation, but it gives grade A to considering medicines (for example for weight management) to support individual care goals, and its supporting text calls GLP-1-based therapies for weight management in people with overweight or obesity "highly beneficial" (ADA Section 3). No independent, publicly funded long-term trial of these drugs in prediabetes was found.

Orlistat and other drugs

NICE says GPs should consider orlistat, with a low-fat diet, for people at high risk who cannot lose weight through lifestyle change alone (NICE PH38). For people with prediabetes, the ADA recommends screening for and treating cardiovascular risk factors such as blood pressure and cholesterol (ADA Section 3). This matters because heart risk is the main long-term threat, and glucose-lowering at this stage has not been shown to reduce it.

Supplements and lifestyle claims

Supplements are widely sold for "blood sugar support". Independent evidence is thin. The US National Center for Complementary and Integrative Health (NCCIH) says some studies suggest that chromium, cinnamon or berberine might help blood sugar control, but that limitations in most of those studies lead to "no clear conclusions" (NCCIH).

For a wider review, see our diabetes supplements evidence guide. No supplement has been shown to match a structured lifestyle programme. A supplement should not replace re-testing.

Risks of common treatments

• Stomach upset with metformin. In the DPP, gastrointestinal symptoms occurred at 77.8 events per 100 person-years on metformin vs 30.7 on placebo (DPP full text).
• Low vitamin B12 with metformin. After 13 years, low or borderline-low B12 affected 20.3% on metformin vs 15.6% on placebo. Anaemia was more common on metformin, and nerve symptoms were more common in those with low B12 (Aroda 2016). The ADA advises periodic B12 checks in long-term users (ADA Section 3).
• Muscle and joint problems with exercise programmes. Musculoskeletal symptoms were highest in the DPP lifestyle group (24.1 per 100 person-years) (DPP full text). Serious harms were rare in the Cochrane trials (Hemmingsen 2017).
• Gastrointestinal effects with GLP-1-type drugs. These were the most common side effects in the tirzepatide trial, mostly during dose increases (Jastreboff 2025). Weight regain and loss of effect after stopping are possible, as diabetes rates rose once treatment ended.
• The label itself. Critics argue that a diagnosis brings worry, more testing and medicine use for people who would never have developed diabetes (Yudkin & Montori). We found no large trial that measured these harms directly.

Who to see

In the UK, start with your GP practice, which can arrange blood tests and referral to the Healthier You NHS Diabetes Prevention Programme where it is available. The NHS website also points to the Diabetes UK risk tool and the free NHS Health Check for adults aged 40 to 74 (NHS). A practice nurse, dietitian or pharmacist can help with diet, weight and medicines. In the US, there is the CDC-recognised National Diabetes Prevention Program. For people on Medicare, the Medicare Diabetes Prevention Program offers 16 weekly sessions followed by 6 monthly maintenance sessions, in person or online (CMS). See a doctor promptly for any of the red-flag symptoms above. Ask for a recheck rather than accepting a single borderline result as permanent.

Follow the money: who funds the research and who profits

Who profits if more people are labelled

Prediabetes is a large market because so many people meet the criteria. Several industries gain when thresholds are low and screening is wide. Makers of glucose-lowering and weight-loss drugs are one. Makers of tests and glucose-monitoring devices are another. Providers of prevention programmes, in person and digital, are a third. On the other side, fewer parties gain from narrower thresholds. Public payers such as the NHS save money by targeting fewer people, which is one reason NICE's band is tighter.

The American Diabetes Association

The ADA sets the most widely used definition. It states that its Standards of Care are "funded by ADA general revenue". It says no industry money supports the guideline, members with a relevant conflict cannot vote on a recommendation, and no committee member worked for a drug or device company in 2026 (ADA methods). The ADA's general revenue does include industry money, however. Its own unaudited disclosure lists $24,844,486 from drug and device makers in 2023 (ADA 2023 disclosure). The largest payments were:

• Eli Lilly: $7,272,646. Lilly also makes tirzepatide.
• Abbott Laboratories / Abbott Diabetes Care: $6,397,698.
• Novo Nordisk: $2,740,266. Novo also makes semaglutide.
• Genentech: $2,002,500.
• Xeris: $1,852,580.
• Sanofi: $1,120,437.
• Dexcom: $729,396.

Diagnostic companies including Roche Diagnostics, Bio-Rad and EKF Diagnostics also appear on the list, with smaller sums. The ADA also co-funded the original DPP trial (DPP acknowledgements). None of this shows that the ADA's thresholds are wrong. It does mean that the body with the broadest definition is partly funded by companies that sell to people with diabetes and prediabetes.

Who funded the key trials

• DPP (US): led by the NIH's National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) with other US agencies. It was co-funded by the ADA, Bristol-Myers Squibb and Parke-Davis. Lipha Pharmaceuticals supplied the metformin and placebo, and companies including LifeScan, Merck, Nike, Slim-Fast Foods and Quaker Oats provided materials and equipment (DPP 2002). The long-term follow-up was funded by NIDDK, with no conflicts declared (DPPOS).
• Da Qing (China): US CDC, WHO, the Chinese CDC, the World Bank, the Chinese Ministry of Public Health and Chinese hospitals (Gong 2019).
• Cochrane progression review: funded by the World Health Organization (Richter 2018). One author of the Cochrane lifestyle review declared fees from several diabetes-drug companies (Hemmingsen 2017).
• NHS DPP evaluations: the DIPLOMA studies were funded by NIHR (McManus 2022). The early-outcomes paper was funded by NHS England. Its lead author was the programme's clinical lead. Another author had chaired the NICE guideline group and co-directs the Leicester Diabetes Centre, whose programme one provider used as the basis of its course (Valabhji 2020). The 2025 completer analysis included authors employed by an analytics firm paid by NHS organisations (Barron 2025).
• GLP-1-type drug trials: SURMOUNT-1 was funded by Eli Lilly (Jastreboff 2025). The semaglutide STEP analysis had Novo Nordisk employees as co-authors (Perreault 2022).

Prevention programme companies

Diabetes prevention has become a business. In England, the first wave of the NHS DPP was delivered by four providers: Reed Momenta, ICS Health and Wellbeing, Ingeus UK and Living Well Taking Control (Valabhji 2020). In the US, Medicare pays approved suppliers through attendance-based and weight-loss-based payments. From 2026 it also covers asynchronous online delivery (CMS).

Omada Health, based in San Francisco, launched its first programme, in diabetes prevention and weight health, in 2012 and filed to list its shares in 2025. Its filing reports revenue of $169.8 million in 2024 and $122.8 million in 2023, with net losses of $47.1 million and $67.5 million. It also reports more than 679,000 members and "full recognition" from the CDC for some deployments of its prediabetes programme (Omada S-1). The company now trades on Nasdaq. For 2025 it reported revenue of $260 million, up 53%, a net loss of $13 million and 886,000 members at year end (Omada 2025 results). The independent check on such programmes came from a 2016 US review. Digital programmes with human coaches were judged better than usual care, while fully automated ones lacked enough evidence (ICER/CTAF 2016). ICER says it is funded mainly by non-profit foundations. Drug companies and insurers contribute less than 25% of its revenue, through its policy forum (ICER).

Documented integrity concerns

We found no regulatory action against the main prevention programmes. The documented concerns are structural. Definitions have widened over time (Cochrane). The body with the widest definition receives large industry income (ADA). National programmes are partly evaluated by the people who run them (Valabhji 2020). A 2019 investigation in Science examined the ADA's industry ties and the prediabetes label (Piller 2019), but we could not access its full text to check its findings.

Related research

For wider, evidence-graded context, see these Pure City Research guides:

• Type 2 diabetes: prevention and management evidence
• Diabetes and heart disease: cardiovascular risks and warning signs
• Diabetes supplements: benefits, limits and safety
• Berberine: glucose evidence, formulations and safety
• Psyllium husk: cholesterol, blood sugar, IBS and safety
• Vitamin D: D2 vs D3, deficiency, benefits and safety
• Myo-inositol: PCOS, fertility and pregnancy trials
• MASLD and MASH: fatty liver diagnosis and treatment
• High cholesterol: assessment, treatment and evidence funding
• High blood pressure: prevention and management evidence

Frequently asked questions

Sources and funding notes

What we could not verify: We could not open the full text of the 2019 Science investigation "Dubious diagnosis" (the publisher blocked access), so we did not use its claims. The WHO 2006 report and WHO's 2011 statement on HbA1c could not be downloaded, so WHO thresholds are taken from the Cochrane and BMJ reviews that quote them. NHS England's programme pages and NHS Digital's National Diabetes Audit tables blocked automated access. The 2023 diagnosis count therefore comes from an NHS networks summary of the audit, and we could not confirm current NHS DPP providers, digital providers, costs per participant or the latest referral numbers. We did not obtain the funding statement of the Barry 2017 BMJ review or the full list of other D2d funders. Omada Health's major shareholders were not identified in the sections we read. Individual ADA committee members' disclosures and NICE PH38 committee interests were not reviewed in full. The ADA's 2023 industry figures are labelled as an unaudited draft.

1. NICE PH38: Type 2 diabetes: prevention in people at high risk (updated September 2017) — UK guideline; NICE is funded by the UK Department of Health and Social Care.
2. ADA. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes—2026 — definitions; ADA general revenue.
3. ADA. Prevention or Delay of Diabetes: Standards of Care in Diabetes—2026 — prevention recommendations; ADA.
4. ADA. Introduction and Methodology: Standards of Care in Diabetes—2026 — guideline funding and conflict policy; ADA.
5. ADA. Revenue received from pharmaceutical/device makers, 2023 (unaudited draft) — industry income; ADA self-disclosure.
6. WHO/IDF. Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia (2006) — WHO definition; WHO (member-state and voluntary funding).
7. CDC. National Diabetes Statistics Report — US prevalence; US government.
8. NHS networks: More people at risk of type 2 diabetes (June 2024) — National Diabetes Audit figures; NHS.
9. NHS: Symptoms of type 2 diabetes and how it's diagnosed and NHS: Diabetic ketoacidosis — patient information; NHS England.
10. Knowler et al. NEJM 2002 (DPP) and full text — founding trial; NIH-led, co-funded by ADA, Bristol-Myers Squibb and Parke-Davis; drugs from Lipha.
11. DPP Research Group. Lancet Diabetes Endocrinol 2015 (DPPOS 15-year) — long-term follow-up; NIDDK.
12. Goldberg et al. Circulation 2022 — 21-year cardiovascular outcomes; DPP Research Group (NIH).
13. Knowler et al. Diabetes Care 2025 — 22-year metformin re-analysis; support services from McKesson BioServices and Matthews Media Group under contract.
14. Aroda et al. JCEM 2016 — metformin and vitamin B12; DPP Research Group.
15. Richter et al. Cochrane 2018 — progression and regression; funded by WHO.
16. Madsen et al. Cochrane 2019 — metformin for prevention; no conflicts known.
17. Hemmingsen et al. Cochrane 2017 — diet and activity; one author declared drug-company fees.
18. Yudkin & Montori. BMJ 2014 and full text — overdiagnosis critique; no relevant interests declared.
19. Rooney et al. JAMA Intern Med 2021 — older adults (ARIC); NIH grants.
20. Barry et al. BMJ 2017 — screening accuracy and screen-and-treat; University of Oxford; funding not verified.
21. Tabák et al. Lancet 2012 — review of prediabetes; authors declared no conflicts.
22. Huang et al. BMJ 2016 — cardiovascular risk meta-analysis; no support for the submitted work declared.
23. US Preventive Services Task Force. JAMA 2021 — screening recommendation; US federal panel.
24. Tuomilehto et al. NEJM 2001 (Finnish DPS) — lifestyle trial; Finnish public health institute; funding statement not accessed.
25. Gong et al. Lancet Diabetes Endocrinol 2019 (Da Qing 30-year) — long-term outcomes; US CDC, WHO, Chinese CDC, World Bank and others.
26. Valabhji et al. Diabetes Care 2020 and full text — NHS DPP early outcomes; NHS England.
27. McManus et al. Lancet Reg Health Eur 2022 — population impact; NIHR.
28. Ravindrarajah et al. PLoS Med 2023 — referral and conversion; University of Manchester team; no competing interests.
29. Barron et al. Nat Med 2025 — completers and long-term conditions; NHS England authors and an NHS-paid analytics firm.
30. Jastreboff et al. NEJM 2025 (SURMOUNT-1) — tirzepatide; funded by Eli Lilly.
31. Perreault et al. Diabetes Care 2022 (STEP) — semaglutide; Novo Nordisk co-authors.
32. Pittas et al. NEJM 2019 (D2d) — vitamin D; NIDDK and others.
33. NCCIH: Diabetes and dietary supplements (November 2023) — supplement evidence; US NIH.
34. MedlinePlus: Metformin — safety information; US National Library of Medicine.
35. CMS: Medicare Diabetes Prevention Program Expanded Model — US payer programme; US government.
36. ICER/CTAF: Diabetes Prevention Programs final report (July 2016) and ICER funding — independent assessment; mainly foundation-funded, with under 25% from industry and insurer dues.
37. Omada Health Form S-1 (2025) — company financials; self-reported to the US Securities and Exchange Commission.
38. Omada Health: fourth quarter and full year 2025 results (March 2026) — company-reported revenue, net loss and members; Omada Health press release.
39. Piller. Science 2019, "Dubious diagnosis" — investigative news article; record only, full text not accessed.

Educational research reviewed 7 October 2026. This article is not medical advice; talk to a qualified clinician or pharmacist before starting, stopping or combining treatments.