Can Four Health Texts Lower Blood Pressure? PUSHME Says No

A weekly text about walking or salt intake is cheap, easy to send, and easy to ignore. In the 2025 PUSHME randomized trial, four one-way health messages did not lower blood pressure more than usual care in the overall group of Swedish primary-care patients with hypertension. The result is a useful corrective to the idea that any digital intervention with a notification badge qualifies as treatment.

That does not mean phones cannot help with blood-pressure control. It means the design matters. Programs that ask patients to measure blood pressure, return data, receive feedback, or work with clinicians are doing more than broadcasting reminders. PUSHME tested the broadcast version, and its primary endpoint was null.

What PUSHME actually tested

The PUSHME trial enrolled 401 adults aged 40 to 85 with ICD-10-coded hypertension at 10 Swedish primary health-care centers. In this six-month parallel-group randomized controlled trial, the intervention group received four weekly, one-way SMS messages. The messages addressed brisk walking for at least 150 minutes per week, tobacco cessation, sodium reduction, and cardiovascular health and alcohol. The control group received usual care.

The investigators designed the study to detect a 4 mmHg between-group blood-pressure difference, assuming a standard deviation of 13 mmHg, with 80% power and 186 people per group. That makes the null result more informative than a tiny pilot study that simply cannot distinguish a small effect from noise.

At six months, the intervention did not improve systolic or diastolic blood pressure versus usual care in the full study population. The available report does not provide extractable confidence intervals for the primary blood-pressure estimates here, so readers should not infer a precise zero effect. The defensible conclusion is narrower: this brief, non-personalized, one-way SMS program did not beat usual primary care on its prespecified primary outcome.

Interesting subgroups are not a clinical recommendation

Two subgroup findings deserve attention without being promoted into proof. Among sedentary participants, diastolic blood pressure fell by 5.2 mmHg in the text-message group and 2.4 mmHg in controls. Among participants who rated their health poorly, the corresponding changes were 4.5 and 1.4 mmHg. Those differences are large enough to be worth testing again, especially because PUSHME was powered around a 4 mmHg between-group difference.

But subgroup results can mislead when investigators examine several slices of a trial. The available publication information does not establish whether these analyses were prespecified. Until a new trial targets these groups prospectively, the signals are hypotheses, not a reason to enroll every patient with hypertension in a reminder-text program.

"Mobile health" is too broad to answer the question

The broader digital-health literature looks more favorable because it often evaluates a different intervention. A 2024 umbrella review of mobile-phone interventions combined 34 meta-analyses covering 235 randomized trials, 48,957 participants, 52 countries, and 10 chronic diseases. It found positive signals across disease areas, but it pooled apps, interactive messaging, and clinical decision-support systems with simpler tools. A pooled result from that mix cannot tell us whether four one-way texts lower blood pressure.

The distinction is more than a technicality. A 2025 systematic review of smartphone apps for blood-pressure control found that interactive tools can reduce blood pressure. Their likely active features include self-monitoring and feedback loops—features PUSHME did not include. Similarly, a 2021 observational cohort study of 1,633 hypertension-clinic patients found better blood-pressure management with a physician-supervised app integrated with electronic medical records. App users chose whether and how often to upload readings, and clinicians could respond. Because the study was observational, it cannot prove the app caused the improvement, but its architecture bears little resemblance to passive SMS.

Context may also matter. A systematic review of mHealth interventions in African settings reported benefits for diabetes and hypertension management, where baseline access to care can differ substantially from Swedish primary care. The included interventions were heterogeneous, so it does not validate one-way SMS. It does support a reasonable possibility: a prompt may add more where routine follow-up and monitoring are less available.

Even SMS evidence for adherence has a low ceiling

Text messages can affect some behaviors without producing a measurable blood-pressure benefit. A 2024 Cochrane review included 18 randomized trials and 8,136 people receiving secondary prevention for established cardiovascular disease. Ten trials reported improved medication adherence, with relative risks for being adherent from 1.10 to 1.14. Yet the review rated that evidence very low certainty, and fatal cardiovascular events did not differ meaningfully: odds ratio 0.83, 95% confidence interval 0.47 to 1.45.

This is neither a contradiction nor a rescue of PUSHME. Medication adherence in people who already have cardiovascular disease is a different outcome in a different population. It does show that a message can sometimes change a small part of behavior. It does not show that messages alone lower blood pressure in unselected primary-care patients.

What moves blood pressure requires participation

The most striking comparison in this evidence is not between texts and another app. It is between passive prompts and training. In a double-blind, sham-controlled trial, Craighead and colleagues assigned 36 adults aged 50 or older with above-normal blood pressure to high-resistance inspiratory muscle strength training or sham training. The active protocol was 30 breaths at 75% of maximal inspiratory pressure, about five minutes per day for six weeks. Systolic blood pressure fell by approximately 9 mmHg, alongside changes in endothelial function, nitric-oxide bioavailability, and oxidative stress.

The sample was small, so that estimate needs replication in larger trials. Still, the sham-controlled design limits expectation effects, and the result is far larger than any overall PUSHME signal. A 2025 meta-analysis of resistance training in older adults also found consistent, clinically meaningful blood-pressure reductions. Exercise asks more of the participant than reading a text. That is probably part of why it delivers more.

Supplements and dietary additions may have a place, but the evidence described here puts them in a more modest category. A 2024 meta-analysis of beetroot juice in people with hypertension found statistically significant but limited blood-pressure effects. A 2025 review of prebiotics found blood-pressure modulation through gut-microbiome pathways, again with modest absolute effects. A vitamin network meta-analysis reported variable findings across vitamin classes, while a 2025 Cochrane review of riboflavin found insufficient randomized-trial evidence to conclude that vitamin B2 lowers blood pressure. Insufficient evidence does not prove no effect; it does not justify a blood-pressure recommendation either.

Where PUSHME applies—and where it does not

PUSHME excluded people with blood pressure above 180/110 mmHg, systolic pressure below 120 mmHg, serious illness, cognitive impairment, and people without smartphones. Its results therefore do not answer whether similar outreach helps people at the highest risk or those with the least access to care. Six months may also be too short for some lifestyle changes to become durable.

Swedish primary care offers relatively high baseline access and blood-pressure management. A simple reminder has less room to improve care when patients already receive routine follow-up. That could help explain why an SMS program might perform differently in lower-resource settings, though it remains an untested inference for this exact design.

Four generic texts over four weeks did not lower blood pressure beyond usual care for the average PUSHME participant. Digital care can work, but monitoring, feedback, and clinical response appear more plausible ingredients than message delivery alone.

For a person trying to lower blood pressure, the practical ranking is clear enough from this evidence. Treat one-way reminders as an optional nudge, not therapy. Put more weight on interventions that require and measure real behavior—such as structured exercise—and use supplements with claims proportionate to their modest or uncertain evidence. A phone can remind you to train. It cannot do the breaths, lift the weights, or take the walk.