Framing and Financial Incentives to Increase Influenza Vaccination Uptake: A Randomized Controlled Trial in Chinese Communities

T0: Control group. Participants receive no additional vaccination information.
T1: Neutral information group. Participants receive objective factual information on where to get vaccinated, timing, costs, safety standards, and appointment procedures.
T2: Positive-framing group. Participants receive messages emphasizing the benefits of vaccination, such as protecting oneself and family members, reducing missed work or school, and contributing to herd immunity.
T3: Negative-framing group. Participants receive messages emphasizing the risks of non-vaccination, such as infection risk, complications, medical costs, and transmission to vulnerable household members.
Monetary incentive module. Participants are additionally assigned to either no subsidy or to a randomly determined monetary incentive/subsidy amount between RMB 10 and RMB 100. The study materials also describe this incentive module as a subsidy/lottery notification.

Primary hypotheses
H1. Subsidy effect on current uptake.
Higher subsidy amounts will increase verified flu-vaccination uptake in the current season.
H2. Information effect on current uptake.
Receiving any vaccine-information message will increase current-season vaccination uptake relative to receiving no information.
H3a. Complementarity between information and subsidy.
Information messages will strengthen the behavioral response to subsidies, so the marginal effect of subsidy on uptake will be larger in the information arms than in the control arm.
H3b. Substitutability between information and subsidy.
Information messages will substitute for the motivating effect of subsidies, so the marginal effect of subsidy on uptake will be smaller in the information arms than in the control arm.
H4. Framing effect of message content.
Message framing will matter for current-season vaccination uptake. In particular, framed messages (positive/gain or negative/loss) will have larger effects than a neutral informational message.
H5. Loss framing versus gain framing.
Negative/loss framing will generate a larger increase in current-season vaccination uptake than positive/gain framing.
H6. Framing effect on future intention. Information framing will affect stated future vaccination intention, including willingness to vaccinate in the next influenza season and willingness to vaccinate under lower out-of-pocket cost.
H7. Behavioral and attitudinal change. Information and subsidy treatments will shift beliefs about vaccine effectiveness, side-effect risk, and influenza risk, and will also affect behavioral margins such as stated intention, perceived barriers, and self-reported reasons for vaccination or non-vaccination.
H8. Health consequences of uptake. Higher vaccination uptake will be associated with lower influenza-like illness, lower healthcare utilization, and fewer missed work or school days over the influenza season.


Longer-run hypotheses
H9. Persistence of intervention effects.
Exposure to current-season subsidies and/or information will increase stated willingness to vaccinate in the following flu season even when next-season vaccination is not subsidized.
H10. Information sustains future motivation.
Information messages paired with subsidies will produce more persistent future vaccination intention than subsidies alone.

Mechanism hypotheses
H11. Belief updating.
Information messages will improve vaccine-related beliefs, including perceived effectiveness, perceived safety, and/or perceived benefits of vaccination.
H12. Beliefs as a mediator.
Part of the effect of information messages on current-season uptake and future intention will operate through changes in vaccine-related beliefs.

2025-10-15

2025-12-15

Vaccine uptake: whether the participant reports having received the current season’s influenza vaccine at the midline follow-up.

Vaccination behavior:
• Vaccination date, timing of uptake, out-of-pocket vaccination cost, vaccine brand, and self-reported reasons for vaccination.
• For participants not yet vaccinated, self-reported reasons for non-vaccination and remaining barriers to uptake.
Current and future intentions:
• Current vaccination intention at midline (intent_f1).
• Vaccination intention under full reimbursement or zero price (intent_free_f1).
• Endline stated likelihood of vaccinating in the next influenza season.
Beliefs and attitudes:
• Perceived vaccine effectiveness.
• Perceived side-effect risk.
• Perceived influenza risk for the general population.
• Perceived own influenza risk and perceived relative risk compared with peers.
• Changes in attitudes toward the vaccine and toward influenza over the season.
Message exposure and treatment-response measures:
• Whether the participant saw the study message.
• Perceived framing of the message.
• Message credibility, persuasiveness, irritation, and sharing behavior.
• Exposure to and participation in the subsidy or lottery notification.
Health and utilization outcomes:
• Any influenza-like illness symptoms.
• Symptom duration.
• Healthcare utilization for influenza-like illness.
• Missed work or school days.
• Overall impact of influenza on daily life.
Additional behavioral follow-through outcomes:
• Alignment between baseline intention and realized vaccination behavior.
• Self-reported reasons for intention-behavior mismatch.
• Main stated driver of the final vaccination decision.

This is an individually randomized controlled trial. The design materials target an overall sample of approximately 1,300 participants, with about 300 participants assigned to each of the four information arms. If recruitment falls short of the target, the design calls for keeping cell sizes as balanced as possible while preserving balance on key baseline characteristics. Random assignment is stratified using baseline characteristics including gender, age, whether the participant was vaccinated in the previous influenza season, and baseline vaccination intention. The study collects three survey waves: a baseline survey before treatment, a midline survey 2–3 weeks after the intervention, and an endline survey at the end of the influenza season, approximately 3–4 months after the intervention.

Computerized randomization was conducted in the survey platform prior to message delivery.

Individual.

No

50 cities

1200 Individuals

300 Individuals for each arm