Home-based hypertension care cut blood pressure 8 mm Hg in rural South Africa

Hypertension is the rare global health problem where the drugs are cheap, off-patent and effective, and the failure is almost entirely one of delivery. A randomised trial published in September in the New England Journal of Medicine tested the delivery side directly: it took hypertension care out of the clinic in rural South Africa and moved it into people's homes [s1].

The result was a mean systolic blood pressure at six months 7.9 mm Hg lower than standard clinic-based care (95% CI, −10.5 to −5.3; P<0.001) in the arm using community health workers, and 9.1 mm Hg lower (95% CI, −11.7 to −6.4; P<0.001) in an enhanced arm where blood-pressure machines transmitted readings automatically [s1].

What the trial did

The IMPACT-BP trial was open-label and randomised 774 adults with hypertension to one of three groups [s1]. The home-based care group monitored their own blood pressure, received home visits from a community health worker for data collection and medication delivery, and had remote, nurse-led decision-making supported by a mobile application [s1]. The enhanced group received the same intervention, but with blood-pressure machines that transmitted readings automatically rather than relying on manual collection [s1]. The third group received standard clinic-based management [s1].

The population was not a young, healthy convenience sample. Mean age was 62 years, 76.0% were women, 13.6% had diabetes and 46.5% were living with HIV [s1] — a comorbidity mix that reflects the actual patient load of rural South African primary care rather than an idealised one.

The primary outcome was systolic blood pressure at six months; secondary outcomes were systolic pressure at 12 months and hypertension control at 6 and 12 months [s1].

The control numbers are the more useful result

Blood-pressure differences of 8 or 9 mm Hg are meaningful, but the categorical result is easier to act on. At six months, 32.5% of the standard-care group had controlled hypertension, compared with 57.4% in the community health worker group (relative risk, 1.76; 95% CI, 1.40 to 2.13) and 61.3% in the enhanced group (relative risk, 1.89; 95% CI, 1.51 to 2.27) [s1]. The improvements in both systolic pressure and control appeared to persist at 12 months [s1].

Retention in care was observed in more than 95% of participants in both home-based groups [s1] — which, in a condition whose defining problem is people falling out of follow-up, is arguably the headline finding.

Severe adverse events occurred in 2.7% of participants and deaths in 1.0%, at similar rates across all three groups [s1].

Why an 8 mm Hg delivery effect matters globally

The global hypertension picture is one of enormous, largely untreated prevalence. The NCD Risk Factor Collaboration's pooled analysis of 1,201 population-representative studies covering 104 million participants found that the number of people aged 30–79 with hypertension doubled between 1990 and 2019 — from 331 million women and 317 million men to 626 million women and 652 million men — despite a stable age-standardised prevalence [s2]. In 2019, 59% of women and 49% of men with hypertension reported a previous diagnosis [s2].

Against that backdrop, the standard-care control rate of 32.5% in this trial is not an indictment of South African clinics specifically; it is roughly what clinic-based hypertension care achieves in most of the world. What IMPACT-BP shows is that a substantial part of that gap is addressable without any new drug.

What the trial does not establish

Several limits are worth stating plainly. The trial was open-label — participants and staff knew their assignment — which matters more for a self-monitoring intervention than it would for a pill, because measurement behaviour itself is part of the intervention. Blood pressure is also a surrogate: the trial measured pressure and control rates, not strokes, heart attacks or deaths, and was neither sized nor long enough to do so.

The difference between the two home-based arms was small. Automatic transmission of readings produced a systolic pressure 1.2 mm Hg lower than manual collection by community health workers, and a control rate 3.9 percentage points higher [s1] — differences well inside the confidence intervals around each estimate and not, on this evidence, a reason to require the more expensive equipment.

Nor does the trial answer the cost question. Home visits, medication delivery and nurse time are real recurring costs, and the paper's outcomes do not tell a health ministry what the intervention costs per patient controlled or whether it scales beyond the trial's supported setting.

The trial was supported by the National Institutes of Health and others, and is registered as NCT05492955 and as DOH-27-112022-4895 in the South African National Clinical Trials Register [s1].

What to watch: whether the 12-month persistence holds over longer follow-up, and whether any health system attempts this outside a trial's supervision.

This article is informational and is not medical advice.

Sources

• [s1] "Home-Based Care for Hypertension in Rural South Africa," New England Journal of Medicine, published online September 2025. https://doi.org/10.1056/NEJMoa2509958
• [s2] NCD Risk Factor Collaboration, "Worldwide trends in hypertension prevalence and progress in treatment and control from 1990 to 2019," The Lancet, 24 August 2021. https://doi.org/10.1016/S0140-6736(21)01330-1