Experimental pre-tests of public health communications on the COVID-19 vaccine: A null finding for medical endorsement, risk and altruism

OBJECTIVE: Vaccination campaigns against COVID-19 will only be successful if enough people want to take the vaccine. We tested a government communications intervention to encourage uptake. DESIGN: A pre-registered randomised controlled trial. METHODS: A large, nationally representative sample were randomly assigned to see one of eight posters. The posters varied by image (general practitioner or two hospital doctors) and message (control with public health guidance not related to vaccination, endorsement of the vaccine from the pictured doctor, endorsement with information about COVID-19 risk, endorsement with information about risk and appeal to get vaccinated to protect friends and family). The posters were presented as part of a larger study. The main outcomes were intention to be vaccinated and how soon people would be willing to be vaccinated. RESULTS: The posters induced different reactions indicating that participants had engaged with them. The hospital image was generally preferred to the GP image. Perhaps critically, all intervention messages were trusted less than a control message which did not mention the vaccine (Control Poster Mean = 5.65, SE = 0.09 vs. Poster M Mean = 5.18, SE = 0.09, p <.001; vs. Poster M + R Mean = 5.11, SE = 0.09, p <.001; vs. Poster M + R + F Mean = 5.33, SE = 0.09, p =.01). There were no effects of poster type on intention to take the vaccine or how soon people were willing to take it. CONCLUSION: Although the intervention messages were based on the strongest correlates of vaccine hesitancy identified by contemporaneous surveys, none was effective. More recent research suggests that focusing on the risk of COVID-19 may be less effective than focusing on the benefits of vaccination. Null findings can be as important as positive findings for designing public health campaigns. This study informed government communications about the COVID-19 vaccine.