News · Heart & Metabolic
Digital nudges lifted statin refills from 12% to 14%, and a second nudge worked as well as the first
Health systems send reminders constantly and rarely test them. A JAMA trial randomized 20,604 patients who had stopped refilling statins, then re-randomized everyone who ignored the first message, and found repeating beat switching channel.
- A digital nudge lifted statin refills from 12% to 14% among people who had stopped.
- Sending a second message to the people who ignored the first worked about as well again.
- Switching channel, from portal to text to email, added nothing over repeating the same one.
- Two percentage points is small per person and large across 20,604 of them.
- This counts prescriptions refilled, not heart attacks avoided.
Every health system in the world sends reminders. Almost none of them know whether the reminders work, and fewer still know what to do when one is ignored.
Statin nonadherence is a leading, addressable driver of preventable cardiovascular events, and the standard response is a message. Writing in JAMA, the ADHERE-ASCVD investigators did something unusual with that message: they randomized it, and then randomized what happened next.
Among 20 604 participants who had stopped refilling a statin, initial digital outreach increased 14-day statin refill vs usual communication, 14.4% against 12.0%.
The design is the interesting part
Most trials randomize once and follow people to the end. This one randomized twice.
Participants were randomized 1:1:1:1 to receive secure portal messaging, SMS/text messaging, nonsecure email, or usual communication. Then the people who did nothing were randomized again: initial nonresponders within 14 days underwent prespecified second-stage randomization to repeat the same outreach, switch communication modality, or receive usual communication.
That second step is what makes the trial useful. Knowing that a message helps is mildly interesting. Knowing what to do about the four in five people who ignored it is the operational question, and it is normally answered by whoever runs the program having a hunch.
The hunch was wrong
Ask anyone who manages outreach what to do when a portal message is ignored, and they will tell you to try a text instead. Different channel, different chance of being seen.
Switching communication modalities did not improve refill vs repeating the same modality, 13.2% against 12.5%, with an interval that comfortably includes no difference at all.
Repeating did work. Among initial nonresponders, second-stage outreach further increased refill, 13.0% against 10.4% for those left alone. So persistence pays and channel-hopping does not, which is the opposite of how most programs are built.
What 2 percentage points buys
The effect sizes here are small in a way that would be unimpressive in a drug trial and are not unimpressive here.
Cumulative statin refill through 28 days was higher in those receiving any digital outreach vs usual communication, 24.9% against 21.2%. That is roughly four extra people refilling per hundred contacted.
The trial ran at Kaiser Permanente Northern California, an integrated health care delivery system serving more than 4.6 million members. A message costs a fraction of a cent to send. At that denominator and that price, a two-point difference is a large number of people back on a drug that prevents heart attacks.
Studies have shown that statins lower the risk of heart attack and stroke in people with high LDL cholesterol, which is why the refill matters. A high LDL level leads to the buildup of plaque in your arteries, and a statin only does its work while it is being taken.
What the outcome is not
Nobody in this trial had their cholesterol measured as an endpoint, and nobody was followed to a heart attack.
The outcome was a pharmacy refill, which is several steps removed from the thing that prevents disease. A person can collect a prescription and leave it in a drawer. The gap between refilling and swallowing is real and unmeasured here.
It is also worth noticing what four in five participants did: nothing. Even after two randomized attempts, most people who had stopped their statin stayed stopped. A trial that improves refill from 12% to 14% is describing a small movement in a large problem.
The reason people stop, which this does not address
The framing of outreach research is that non-adherence is a memory problem. Send a reminder, restore the behavior.
That is true for some people and not for most. The commonest reasons for stopping a statin are muscle symptoms, a belief that the drug is unnecessary, something read online, or a decision made after a side effect that was never discussed with anybody. Statins usually don’t cause side effects, but the person who stopped is by definition not in that majority, or believes they are not.
A text message does not address any of that. What it does is catch the subset whose reason for stopping was closer to drift than decision, and this trial suggests that subset is real and reachable, and small.
Results were consistent across prespecified subgroups, which means the effect was not concentrated in one group anyone could target more efficiently.
People also ask
What did the trial find?
Initial digital outreach increased 14-day statin refill vs usual communication (14.4% vs 12.0%; adjusted risk difference 2.3 percentage points, 95% CI 1.3-3.4; adjusted risk ratio 1.20, 1.10-1.30). Among initial nonresponders, second-stage outreach further increased refill (13.0% vs 10.4%). Switching communication modalities did not improve refill vs repeating the same modality (13.2% vs 12.5%; adjusted risk ratio 1.06, 0.94-1.17).
What is a sequential multiple assignment design?
A trial that randomizes people more than once. Everyone who did not respond to the first message was randomized again, which lets the trial answer what to do next rather than only whether the first attempt worked.
Why does the second-stage result matter?
Because it settles a question every health system guesses at. The instinct after an ignored message is to try a different channel. This trial says that instinct is wrong: repeating worked just as well, and switching added nothing.
Is a 2 percentage point improvement worth having?
At the scale these programs run, yes. Kaiser Permanente Northern California serves more than 4.6 million members, and a message costs almost nothing to send. Small effects on very large denominators are how population health programs are judged.
What was the outcome exactly?
A pharmacy refill within 14 days of the message, and cumulative refill through 28 days. Not cholesterol, and not heart attacks. It is a behavioral outcome that sits several steps upstream of the thing anyone actually cares about.
Does refilling mean taking?
No, and this is the honest limit. Collecting a prescription is the last observable step before swallowing a tablet, and the two are correlated but not identical. Nobody watched anyone take anything.
What does this change for a patient?
Very little directly; it is a finding about how health systems should design outreach. If you have stopped a statin, the useful conversation is why, since side effects and doubts about the drug are far more common reasons than forgetting. This is general information rather than medical advice.