Archive · 6 Oct 2026
Personalised Flu-Risk Messages Slightly Increase Vaccination Rates
Researchers wanted to see if a personal warning could encourage people to get a flu shot.
They studied about 90,000 patients in a US healthcare system.
A computer tool helped identify patients at higher risk of flu or its complications.
Some patients received messages explaining their higher risk, while others received a general reminder or no message.
People who got a high-risk message were a little more likely to get vaccinated.
The increase was small, but it appeared across the trials.
Telling patients that an algorithm helped assess their risk did not noticeably change vaccination rates.
The study was published in Nature Human Behaviour on 5 October.
A study of about 90,000 patients in a US healthcare system tested personalised flu-vaccination messages.
Researchers used a previously validated machine-learning algorithm to identify patients at higher risk of flu and complications.
High-risk messages raised vaccination rates by 1.1 to 1.4 percentage points compared with general reminders.
Compared with receiving no message, high-risk messages raised vaccination rates by 1.7 to 3.5 percentage points.
Disclosing algorithm use made little difference to vaccination rates.
- Who
- Researchers, including first author Gail N. Rosenbaum, studied about 90,000 patients in a US healthcare system.
- What
- Three randomised field trials tested whether personalised flu-risk messages encouraged vaccination.
- When
- The study was published on 5 October; the article does not specify the year.
- Where
- A US healthcare system; the article identifies Geisinger as Pennsylvania-based.
- Why
- Researchers wanted to find out whether informing patients of higher influenza risk would make them more likely to get vaccinated.
Risk communication
Algorithm disclosure
Personalised warnings versus general reminders
Risk communication
Messages telling patients they were at higher risk modestly increased vaccination compared with general reminders.
Algorithm disclosure
The article reports a small effect, not evidence that personalised messages produce a large increase.
Disclosing algorithm use
Risk communication
The study found little difference in vaccination rates when messages disclosed algorithm involvement or risk explanations.
Algorithm disclosure
The findings do not show that disclosure itself encouraged or discouraged vaccination.
- Study
- “Algorithm-assisted personalised risk communication to encourage flu vaccination in the USA: three randomised field trials”
- Journal
- Nature Human Behaviour
- Publication date
- 5 October; year not specified in the article
- Participants
- About 90,000 patients
- Risk identification
- A previously validated machine-learning algorithm
- Increase versus general reminder
- 1.1 to 1.4 percentage points
- Increase versus no message
- 1.7 to 3.5 percentage points
- Effect of disclosing algorithm use
- Little difference in vaccination rates