Frame side-effect warnings so patients keep taking their blood-pressure drugs
A one-page script changes how nurses and pharmacists describe common side effects (most are mild, and most feelings of "side effects" occur equally on a dummy pill), to cut the share of patients who stop treatment.
Proposed combination, not yet trialled as one Every source verified Health · checked 2026-11-18
What is new
Taken from: statin trials in high-income countries that measured the nocebo effect.
Applied to: primary-care hypertension programmes in low- and middle-income countries.
Blinded n-of-1 and trial-extension studies of statins found that most symptoms people report on treatment also appear on a dummy pill, and that unblinded patients report more. Telling people what to expect shapes what they feel (the nocebo effect). Stopping treatment because of feared side effects is a major reason blood pressure stays uncontrolled; a framing script costs almost nothing to add to a protocol that already exists.
The parts are supported by the studies below. The combination is a hypothesis, which is why the first test below is designed to be able to fail.
Evidence for each part
| Source | What it shows | Role | Check |
|---|---|---|---|
| Wood et al., New England Journal of Medicine, 2020 doi:10.1056/nejmc2031173 | In a blinded n-of-1 trial, symptom intensity was nearly the same on statin and on placebo and much lower in months with no tablet: almost the whole increase appeared on the dummy pill too (figures are in the paper body, not its abstract). | size of the nocebo effect | title and year found |
| Gupta et al., The Lancet, 2017 doi:10.1016/s0140-6736(17)31075-9 | Muscle symptoms were reported 41% more often on statin when patients knew they were taking it, and not more often when blinded. | unblinded patients report more | title and year found; quoted figure 41 found in the abstract |
| Colloca and Barsky, New England Journal of Medicine, 2020 doi:10.26226/morressier.5cf632c0af72dec2b0554738 | Review of how expectations and the wording of warnings shape side effects and treatment adherence. | the mechanism, reviewed | title and year found |
| Abegaz et al., Medicine, 2017 doi:10.1097/md.0000000000005641 | Systematic review of non-adherence to blood-pressure drugs worldwide. | size of the adherence gap | title and year found |
Closest prior work
The nocebo studies are mostly statin and high-income. The searches below look for a trial that tests expectation framing on blood-pressure drug adherence in low- and middle-income primary care.
The prior-work search was unavailable at the last check, so nothing is claimed about whether this has been tried. It reruns daily.
Benefit and cost
Estimated benefit: 1.05 WELLBYs per USD 1,000 (middle estimate; 5% to 95% range 0.1 to 6.58), counting side effects and spill-overs. Direct effect only: about USD 111 to 2,000 per healthy year of life gained. Chance the net effect is positive: 98%.
100% of this estimate rests on assumptions or weak evidence (grades C and D), not on a measured result for this combination. USD 5,000 for about 10,000 patients on treatment across 20 clinics: script design, one training session per clinic, printed cards (assumption). The pilot below exists to replace these assumptions with measurements.
For scale, the catalogue's cash-transfer entry is 7.55 WELLBYs per USD 1,000, from a published analysis. WELLBYs are one common measure, not the only value of an intervention. How the estimate is made.
| Link | Kind | Grade | Source |
|---|---|---|---|
| Years of healthy life gained (DALYs): 250 to 750 more patients still on treatment out of 10,000, each worth 0.01 to 0.06 DALY a year | direct | D | assumption: 25% baseline stopping, 10-30% relative reduction from framing; DALY value per retained patient-year as in the barbershop entry |
| Real side effects under-reported (WELLBYs lost per DALY gained) | unintended | D | assumption |
How to do it
Who: National hypertension programmes using WHO HEARTS, primary-care trainers, pharmacist associations.
What it takes: A one-page counselling script, an hour of nurse training, a printed card for the patient, and a question at refill visits that records why anyone stopped.
Cheapest first test: 20 clinics in one district, randomised 10 and 10 within a protocol that already runs. Intervention clinics use the script at the first prescription and at each refill. Primary outcome: share of new patients still collecting drugs at 6 months, from refill records. Safety outcome: serious adverse events reported to the clinic, to confirm that honest framing does not hide real harm. Stop rule: the pilot fails if retention at 6 months is not at least 5 points higher, or if serious events are higher in intervention clinics.
What could go wrong
- The script could be read as dismissing a real reaction; the pilot has patients flag any symptom for a clinician anyway.
- Counselling time is scarce; the script is meant to replace, not add to, the standard warning.
- Refill records may not capture patients who switch to another pharmacy.
The nocebo evidence is strongest for statins, not blood-pressure drugs, and for patients in rich countries. Framing might reduce honest reporting of a real adverse effect, which is why the pilot measures serious events. The size of the effect on retention is unknown until a trial runs.
Machine-readable: JSON. Related catalogue entries: hypertension. Found an error? Tell us.