Field Testing the Transferability of Behavioural Science Knowledge on Promoting Vaccinations

"Effectively translating behavioural science research to practices and identifying behavioural interventions that can change consequential behaviours across settings requires concerted efforts..."

The uptake of COVID-19 booster shots has stalled in many countries. Conducting randomised controlled trials (RCTs) in this context can help generate policy-relevant knowledge about the impact of behavioural science in promoting vaccination. This context also offers an environment for investigating whether interventions built on prior research produce consistent findings in the field. Across three pre-registered RCTs (N = 314,824), the researchers field tested text-message-based reminders that either increased vaccinations in prior field work (RCT1), elevated vaccination intentions in an online study (RCT2), or were favoured by scientists and non-experts (RCT3).

In the RCTs, the researchers varied the presence and language of reminders, based on the notion that behaviour change involves two stages: (i) establishing intentions to act and (ii) turning these intentions into action. To help patients turn their intentions to receive the COVID-19 bivalent booster into action, all messages in the three RCTs sought to address one common barrier to follow-through - forgetfulness - by reminding patients of their eligibility for the booster and encouraging them to get it.

The first RCT (N = 136,452) delivered intention- and action-oriented interventions based on strategies that previous field experiments identified as effective for promoting initial COVID-19 vaccinations. In a prior RCT involving the same population as the current RCTs, reminders that induced a sense of psychological ownership by asking patients to "claim their dose" were the best-performing reminders, leading to higher vaccine uptake than basic reminders that simply prompted action with a link to schedule appointments. To examine whether the effectiveness of reminders with ownership language would translate to booster uptake, the first RCT included messages containing the "claim your dose" language. Two additional messages combined doctor endorsement with the ownership language and appointment scheduling links. This first RCT also compared two strategies for providing appointment scheduling links.

The second RCT (N = 159,195) examined interventions that the researchers developed based on a survey of COVID-19-related beliefs and that they found effective at changing hypothetical booster intentions in a concurrent online experiment. This RCT compared a basic message, which simply told patients they could now get the COVID-19 bivalent booster, with five messages containing additional content aimed at elevating patients' vaccination intentions.

The third RCT (N = 90,968) examined interventions derived from the behavioural literature and forecasted to be effective by experts. These interventions were designed to remind patients that they could get a flu shot alongside the COVID-19 booster shot.

The results of RCT1 showed that reminders and psychological ownership language increased booster uptake, replicating prior findings. For example, among 13 messages with links to vaccination venues, the messages containing the "claim your dose" language led to significantly higher booster uptake by 0.78 percentage points, compared with other reminders lacking such ownership language. However, strategies deemed effective by intention (RCT2) or prediction surveys (RCT3), such as addressing misconceptions or encouraging the bundling of COVID-19 boosters and flu shots, yielded no detectable benefits over simple reminders.

Thus, "While hypothetical surveys and self-reports are undoubtedly valuable for providing foundational evidence on the mechanisms of human behaviour," the findings suggest "they may not always translate to complex real-world situations where various factors can affect behaviour".

That said, the research shared here does provide one data point on the robustness of interventions identified as effective in prior field evaluations: sending reminders, leveraging psychological ownership to make individuals feel the vaccine is theirs, and providing information about doctors' endorsement of the vaccine. While the effect sizes of text reminders and ownership framing are modest, they are noteworthy, in part because participants had been repeatedly exposed to COVID-19 vaccination messages, such as text messages encouraging COVID-19 booster uptake from pharmacies. "These findings provide some evidence that text reminders and ownership framing can remain effective even when deployed multiple times within the same population, which may have broader implications for motivating repeated behaviour change."

In conclusion: "the findings highlight the importance of accumulating knowledge about intervention impact in the real world, which could help practitioners save critical resources when attempting to influence consequential behaviours."