Who ran each study, how many people were in it, and what it can and cannot show.
Each entry names who ran the study, how many people were in it, and what it can and cannot show. Click a study to read the original. Our full reasoning, with limits stated first, is on The Research, and the evidence ledger covers the business case.
20 studies
| Study | People | What it shows | Type |
|---|---|---|---|
| Gao et al., JAMA Network Open, 2024 Sitting and movement | 481,688 adults | In a Taiwan cohort, mostly-sitting jobs were linked to higher cardiovascular death (hazard ratio 1.34). Alternating sitting with moving showed no added risk. | Observational |
| Ajufo et al., JACC, 2024 (UK Biobank) Sitting and movement | 89,530 adults | Risk of heart failure and cardiovascular death rose past about 10.6 hours of sitting a day. | Observational |
| Ekelund et al., Lancet, 2016 Sitting and movement | 1,005,791 people | About 60 to 75 minutes a day of moderate activity removed the extra risk of death linked to sitting more than 8 hours a day. | Pooled observational |
| Chu et al., Obesity Reviews, 2016 Sitting and movement | 21 pooled studies, 4,568 workers | Programs combining several steps cut sitting by about 89 minutes per 8-hour workday. Education alone cut about 16 minutes. | Meta-analysis |
| Edwardson et al., BMJ, 2022 (SMART Work & Life) Sitting and movement | 756 employees, 78 workplaces | At 12 months, sitting fell about 22 minutes a day with the program and about 64 minutes with a height-adjustable desk added. | Randomized |
| Pereira et al., Int J Behav Nutr Phys Act, 2020 (Stand and Move at Work) Sitting and movement | 630 employees, 24 worksites | Program plus sit-stand desks cut sitting about 59 minutes per 8-hour day. The heart-health risk score did not change significantly. A small exploratory subgroup with prediabetes or diabetes (95 people) showed larger effects. | Randomized |
| Raichlen et al., JAMA, 2023 Brain and sitting | 49,841 older adults | Dementia risk was 63% higher at 12 hours of sitting a day than at a typical 9 hours. | Observational |
| Wu et al., Molecular Psychiatry, 2025 Brain and sitting | 73,411 adults | More sedentary time, measured by wearable, was linked to higher risk of dementia, depression and other brain conditions. | Observational |
| Carter et al., Journal of Applied Physiology, 2018 Brain and sitting | 15 adults | Four hours of uninterrupted sitting lowered brain blood-flow velocity. Two-minute walks every 30 minutes prevented it. Eight-minute walks every 2 hours did not. | Small experiment |
| Qaseem et al., Annals of Internal Medicine, 2016 Sleep | Guideline | American College of Physicians: CBT-I is the initial treatment for long-term insomnia. | Guideline |
| Song & Baicker, JAMA, 2019 Workplace programs | 32,974 employees, 160 worksites | Only exercise and weight-management habits improved. No significant change in clinical measures, spending, absence or job performance at 18 months. | Randomized |
| Jones, Molitor & Reif, Quarterly Journal of Economics, 2019 Workplace programs | 4,834 employees | No significant effects on medical spending, productivity or other health behaviors after more than 2 years. | Randomized |
| Swadźba et al., BMC Health Services Research, 2026 Workplace programs | 56 studies | Favorable economics appeared almost only where screening led to follow-up by a physician or coordinator. The authors say it is not supported as a stand-alone cost saver. | Review of observational studies |
| Caloyeras et al., Health Affairs, 2014 (PepsiCo) Workplace programs | One employer | Disease management accounted for 87% of savings while 13% of employees took part. Lifestyle management returned about 50 cents per dollar. | Observational |
| Sieber et al. (NIOSH), American Journal of Industrial Medicine, 2014 Desk jobs and drivers | 1,670 long-haul drivers | Higher rates of obesity (68.9% vs 30.5%), diabetes (14.4% vs 6.8%) and smoking than other U.S. working adults. The desk-driver comparison on this site is an illustration. | Survey |
| Loehr & Schwartz, Harvard Business Review, 2001 Energy and performance | Concept | The corporate-athlete idea about managing energy that EPX3 draws on. EPX3 is independent. | Concept paper |
| Brynjolfsson, Li & Raymond, Quarterly Journal of Economics, 2025 AI at work | 5,172 support agents | About 15% more issues resolved with AI help, and about 30% for less experienced workers. | Field study |
| Dell'Acqua et al., Harvard Business School, 2023 AI at work | 758 consultants | On a task outside AI's strengths, consultants using it were 19% less likely to produce correct solutions. | Randomized |
| Brynjolfsson, Chandar & Chen, Stanford Digital Economy Lab, 2025 AI at work | Payroll data | August 2025 version: about 13% relative employment decline for ages 22 to 25 in the most AI-exposed jobs. Later updates report larger declines. The authors call this descriptive, not causal. | Observational |
| World Economic Forum, Future of Jobs Report, 2025 AI at work | 1,000+ companies surveyed | 170 million jobs created and 92 million displaced by 2030, a net gain of 78 million. About 41% of employers plan workforce reductions as AI automates tasks. | Survey |
Observational studies show links, not proof of cause. A program that helps one group may not help another. These are published findings about other programs, not EPX3 results, and they do not predict yours. EPX3 does not promise savings or a return on investment.
DSS is a proposed clinical framework developed by Dr. Michael Majette. It is not an established medical diagnosis, and it is not recognized in the ICD or DSM. Its components are individually well documented; the framework and the EPX3 program built on it are being tested.