The research behind Exercise Breaks

This is the plain-language summary of the evidence that Exercise Breaks is built on. It walks through what's known, outcome by outcome, with the caveats that matter.

The technical version — every paper cited, with DOIs, effect sizes, and per-paper funding sources — is on the bibliography page.

If you only have time for one section, the four outcomes are blood pressure, mood and energy, post-meal glucose, and the mortality association. Each has its own evidence base and its own caveats; none of them collapse into a single claim.

Roughly 12 minutes to read.

The question

Most office workers sit for the better part of eight hours a day. Many of those workers also exercise — a morning run, a gym session, a lunchtime walk. The question this app is built around is whether that morning workout fully covers the cost of the eight hours of sitting that follows it, or whether something else happens during the workday itself that exercise can't reach.

The short answer the research supports is: mostly the workout helps, but not entirely. Some of what sitting does to the body shows up in physiological systems that respond to what's happening right now, not to what you did at 7 AM. For those systems, the useful intervention is interrupting the sitting itself.

That's what this app is — a reminder to interrupt sitting, every hour, for three minutes. The rest of this document walks through the evidence one outcome at a time.

What movement breaks actually do

Blood pressure

This is the cleanest finding. A 2023 dose-response trial from Keith Diaz's lab at Columbia (Duran et al.) had middle-aged adults sit for eight hours under five different break schedules — including the minimum dose of one minute of light walking every hour. Every break schedule they tested produced a measurable drop in systolic blood pressure of roughly 3 to 5 mm Hg. Even the smallest dose worked.

A separate twelve-week trial in 2026 (Fang et al., BMC Public Health) tested almost exactly the protocol this app delivers — 3-minute breaks every hour during the workday, light-to-moderate bodyweight exercises — in office workers in China. Their result on blood pressure: systolic down 3.9 mm Hg, diastolic down 1.5 mm Hg over twelve weeks compared to a control group that kept sitting.

For context, a 3 to 5 mm Hg systolic reduction is in the range of what blood pressure medication or several months of regular aerobic exercise produce. The app is not a substitute for either of those things, but the effect is real and it shows up at the lowest doses the research has tested.

Caveat: the Duran trial was small (eleven participants, cut short by COVID) and used healthy adults. The Fang trial was larger (86 participants) but ran in a single cultural context with a predominantly overweight sample. The direction of the effect is consistent across both, but the precise number isn't pinned down.

Mood, energy, and afternoon fatigue

The same Duran trial measured mood and fatigue across its five break schedules. Every schedule that involved at least three minutes of movement reduced fatigue and improved mood compared to uninterrupted sitting. The one schedule that didn't clearly help was the minimum dose — one minute of walking every hour. That tells you the app's 3-minute default isn't arbitrary; it's right at the threshold where the mood and energy effects start showing up.

The Fang twelve-week trial reported larger effects on the same outcomes. Participants doing 3-minute hourly breaks reported substantially less afternoon fatigue, more daytime energy, and better self-rated work productivity than the control group. In their data, afternoon fatigue was the largest effect in the entire study — bigger than blood pressure, bigger than glucose.

Caveat that matters here: these are self-reported outcomes from a trial that couldn't be blinded. People who knew they were in the intervention group reported feeling better. Some of that effect is probably real — the same pattern shows up in the Duran trial with objective fatigue measures — but some of it is the placebo and expectation effects you'd expect in any wellness intervention. Honest framing: people who do this report feeling better, and there's reason to think the feeling tracks something real.

Post-meal blood sugar

This is where the picture gets more nuanced and where the app's defaults are a compromise rather than the strongest possible research-backed choice.

If you measure blood sugar acutely — meaning, over a single eight-hour day in a lab — the strongest evidence for movement-break benefits sits at intervals of every 20 to 30 minutes, not every hour. A 2024 network meta-analysis (Gao et al.) ranked the every-30-minutes schedule as the most likely best frequency for blood sugar control. The Duran dose-response trial found significant glucose attenuation at the every-30-minute dose but not at hourly.

But chronic adaptation — what happens over weeks, not hours — appears to look different. The Fang twelve-week trial did find significant improvements in fasting glucose, post-meal glucose, and insulin sensitivity at the hourly interval. The body seems to respond to repeated exposure to hourly breaks even if the single-day acute effect is smaller. In Fang's trial, 56% of participants who started with prediabetes had returned to normal blood sugar by week twelve, compared to 13% in the control group.

The honest framing: for acute post-meal glucose specifically, breaks every 30 minutes have stronger evidence than breaks every hour. For longer-term metabolic health, the hourly interval is supported. This app uses hourly because that's the interval most office workers can actually sustain across a full workday — but if acute glucose control is the specific outcome someone cares about, shorter intervals have a stronger case.

The mortality association

A 2021 pooled analysis in the British Journal of Sports Medicine (Chastin et al.) looked at six accelerometer studies covering more than 130,000 adults. They found that 3 minutes of moderate-to-vigorous activity per hour of sitting was associated with roughly 30% lower odds of early death over the four-year follow-up.

This is the most often-cited number in this literature, and it's useful as a way to communicate that the hourly cadence isn't a made-up cadence — it's how one of the largest pooled analyses in this field chose to express its finding. But two things to keep in mind: this is associational evidence (it can't prove the breaks themselves caused the mortality reduction — sicker people sit longer, which confounds the analysis), and the activity intensity in the data was moderate-to-vigorous. The app's protocol is light-to-moderate. So the 30% figure isn't a number the app's specific protocol can claim — it's the population-scale signal that movement-during-sitting matters at all.

Why hourly, and not every 30 minutes

Every reader who pays attention will notice that some of the strongest evidence in this literature points to 30-minute intervals, not 60. Why does this app default to hourly?

Two reasons, one of them honest, one of them pragmatic.

The honest reason: every-30-minute breaks have stronger acute evidence (especially for blood sugar). Every-hour breaks have stronger chronic evidence (the Fang twelve-week trial), strong evidence for blood pressure at any dose, and strong evidence for mood and energy at 3 minutes or more. The hourly cadence is well-supported for most of the outcomes this app is built around. It is not the strongest possible choice for acute glucose specifically.

The pragmatic reason: sixteen interruptions across an eight-hour workday is more than most desk workers can sustain through real meetings, deadlines, and focus blocks. Eight interruptions is hard enough. The Fang trial measured 82% adherence at the hourly interval over twelve weeks; no comparable trial has shown that level of adherence at 30-minute intervals in a free-living workplace.

The app picks the interval that pairs research support with realistic adherence. The default is hourly. The interval can be configured between 45 and 75 minutes if your workday makes a strict hour impractical, though the directly tested protocol is hourly. People who want to push toward 30 minutes can take extra breaks on top of the prompts. The defaults are the compromise.

Does being reminded actually help

The app is a reminder system, so a fair question is whether reminders themselves change behaviour, or whether people who would move anyway are the only ones who follow them.

A 2025 meta-analysis (Leppe-Zamora et al., International Journal of Behavioral Nutrition and Physical Activity) pooled 18 randomized trials of computer-prompt software vs. control or alternative intervention. Across 1,164 office workers, automated prompts produced an average reduction in workday sitting of about 12 minutes per day and an increase of about 1,000 steps per day. The secondary cardiometabolic outcomes trended favourably but didn't reach statistical significance.

Twelve minutes a day is modest. Honest framing: prompts produce a real, measurable change in behaviour, but they're not transformative on their own. They work best paired with some understanding of why the breaks matter — which is part of why this app includes the explanatory content you're reading.

What this app does not claim

A few stronger claims circulate in this space. None of them are defensible against a careful reader, and none of them appear in the app's copy.

"You can't compensate for sitting with exercise." Two large meta-analyses (Ekelund 2016 in The Lancet, Ekelund 2020 in BJSM) find that 30 to 75 minutes per day of moderate activity substantially reduces or eliminates the all-cause mortality risk associated with high sitting time. The strong version of "exercise doesn't compensate" isn't supported. The narrower claim — that exercise doesn't fully cover the acute physiological costs of sitting (post-meal glucose spikes, hourly blood pressure elevation, afternoon fatigue) — is supported, and it's the claim this app makes.
VILPA-scale mortality reductions. Studies of vigorous intermittent activity (Stamatakis 2022 and others) have found large mortality reductions — 38 to 49% — from short bursts of vigorous movement. The app uses light-to-moderate intensity, which is a different intensity regime. Those effect sizes don't transfer.
"Automated prompts are five times more effective than willpower." This figure circulates in AI-generated summaries and doesn't appear in any peer-reviewed source we could locate. It looks like a fabrication. The honest pooled effect from Leppe-Zamora 2025 is the 12-minute-per-day reduction described above.
"Reduces diabetes risk." The Fang 2026 trial found improvements in glucose, insulin sensitivity, and the proportion of prediabetic participants who returned to normal blood sugar over twelve weeks. Twelve weeks is too short to claim reduced diabetes incidence; what the trial supports is improvements in metabolic markers that are associated with diabetes risk.
That the 45-to-75-minute interval window itself is research-validated. The directly tested intervals in this literature are 20, 30, and 60 minutes. The app's default is 60 minutes, which matches Fang's protocol exactly. The 45-minute lower bound and 75-minute upper bound the app allows are product-design choices that accommodate real workdays — not directly tested intervals. Hourly is what the literature supports; the surrounding window is flexibility for users whose meetings and focus blocks don't fit neat hourly slots.

Who funded this research

A reasonable question to ask of any research-backed product is who paid for the research. The answer for the load-bearing papers behind this app:

The Fang 2026 twelve-week trial was unfunded — no grant from any public, commercial, or not-for-profit source. The Duran 2023 dose-response trial was funded by the Robert N. Butler Columbia Aging Center, an internal Columbia research center with no industry ties. The Chastin 2021 pooled cohort analysis was funded by 17 grants from the US National Institutes of Health, the British Heart Foundation, and the UK Medical Research Council — all public-sector sources. The Ekelund 2016 Lancet meta-analysis was partly funded by the UK Medical Research Council. The Leppe-Zamora 2025 prompt-software meta-analysis was authored by Chilean and Colombian academics with no declared conflicts of interest. None of these papers report industry funding. None of the authors declare financial conflicts of interest relevant to sedentary-break interventions.

Worth saying clearly: the absence of financial conflicts of interest is not the same as the absence of bias. Researchers who have built careers in this field have professional and reputational stakes in the literature continuing to find that movement breaks help, the same way nutritionists who built careers around a food category have stakes in that category continuing to look good. The structural protection against this kind of bias is replication across independent groups in different countries — which the load-bearing research does have. The same direction of effect shows up in trials run in China (Fang), the United States (Duran, Chastin's pooled cohorts), the United Kingdom (Chastin's pooled cohorts), Norway (Ekelund's international consortium), and Chile/Colombia (Leppe-Zamora). Cross-country, cross-institutional replication of the same direction of effect is what makes this literature trustworthy beyond the individual papers.

If you've read this far, you've read more of the evidence than most people who confidently make claims about sitting and exercise ever do. Thanks for caring about the details. Back to the top ↑