The village health worker programme still worked three years after the money stopped

Design

cluster-randomised controlled trial of 326 villages with 3-year post-trial extended follow-up (CRHCP)

Population

33995 adults aged 40 or over in rural China with blood pressure 140/90 or above, or 130/80 or above if at high cardiovascular risk or on treatment; 31334 entered post-trial follow-up

Primary outcome

composite of myocardial infarction, stroke, hospitalisation for heart failure and cardiovascular death

Effect

7-year 2.4% vs 3.0% per person-year, HR 0.76 (95% CI 0.72 to 0.81); 3-year post-trial 3.4% vs 4.2%, HR 0.79 (0.73 to 0.85); blood pressure difference -13.5/-5.4 mmHg

The China Rural Hypertension Control Project randomised 326 villages to a non-physician community health provider-led intensive blood pressure programme, targeting below 130/80 mmHg, or to usual care. Trained providers initiated and titrated antihypertensives under a standard protocol with physician supervision, coached home monitoring, lifestyle and adherence, and the trial supplied discounted or free medicines, extra training and performance incentives. Those four supports were withdrawn at four years while the providers continued caring for the same patients. This is the seven-year report.

Of 33995 enrolled, 31334 entered the post-trial period. At seven years, blood pressure was 138.8/80.7 mmHg in the intervention group against 152.3/86.1 in usual care, a difference of -13.5/-5.4 mmHg, and 33.9% against 10.5% were below 130/80. Over the whole seven years the composite of myocardial infarction, stroke, heart failure hospitalisation and cardiovascular death ran at 2.4% against 3.0% per person-year, hazard ratio 0.76 (95% CI 0.72 to 0.81). During the three years after the supports were withdrawn it was 3.4% against 4.2% per person-year, hazard ratio 0.79 (0.73 to 0.85). Hypotension (risk ratio 1.58) and mild hypokalaemia (1.38) were commoner in the intervention group.

The post-trial period is the finding. Programmes like this routinely evaporate when the project funding ends; this one did not, because the thing that persisted was the trained provider and the protocol rather than the subsidy. For Indian practice that is the directly relevant part - the workforce this model needs already exists in the ASHA and ANM cadre and in the National Programme's NCD clinics, and the trial's design says what has to sit around them: a written titration protocol, physician supervision that is reachable, and a supply of medicines. The harms are the price and should be quoted: more hypotension, more mild hypokalaemia, in a programme chasing 130/80.

• The transferable components are a written titration protocol, reachable physician supervision, and reliable medicine supply - not the incentives.
• Expect a real difference in achieved control, not just in process: 33.9% versus 10.5% below 130/80 at seven years.
• Budget for the harms: hypotension 1.58 times as common, mild hypokalaemia 1.38 times.
• Check potassium as part of the protocol if you target 130/80 with these drug classes.
• Measure achieved blood pressure and control rates at the programme level; that is what predicted the event reduction here.

The statistics, in plain English

A hazard ratio of 0.79 with a confidence interval of 0.73 to 0.85 during the post-trial period, on more than 31000 participants, is both large and precisely estimated - the kind of result cluster randomisation at this scale can deliver and a small trial cannot. Note that event rates rose in both groups after the trial phase (3.4% and 4.2% per person-year against 2.4% and 3.0% over the whole period), which is what an ageing cohort does; the relative benefit held even as absolute risk climbed. Cluster randomisation of villages also means the comparison is between communities rather than individuals, so the effect includes whatever the programme did to the surrounding environment, not just to the person prescribed a tablet.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.