Behavioural nudges increase COVID-19 vaccinations

The study addresses how to increase uptake of COVID-19 vaccines, a pressing public health priority given persistent vaccine hesitancy and follow-through failures. Although vaccines are effective in preventing disease, achieving high coverage remains challenging due to uncertainty, forgetfulness, and procrastination. Behavioural nudges—low-cost interventions that alter choices without restricting options—have successfully influenced various health behaviors. The authors hypothesize that text-message reminders that make vaccination salient and easy, and that leverage psychological ownership, can increase COVID-19 vaccination uptake. They also explore whether pairing reminders with an informational video addressing hesitancy further improves outcomes. The purpose is to evaluate, in field settings, which behavioural designs effectively convert intentions into completed vaccinations.

Prior work has emphasized two complementary strategies: increasing intentions among vaccine-hesitant individuals and facilitating follow-through among those already inclined to vaccinate. Reminders have proven effective across domains such as savings, benefit uptake, exercise, healthy eating, and influenza vaccination. Behavioural science suggests that reducing friction (e.g., forgetfulness, hassle) and using ownership framing can increase compliance. However, much prior COVID-19 work examined hypothetical responses rather than real-world uptake, and evidence on combining reminders with informational content to address hesitancy is mixed. This study extends the literature by testing reminder designs and ownership framing on actual vaccination behavior and by comparing field results with online intention measures.

Design: Two preregistered, sequential randomized controlled trials (RCTs) conducted within the UCLA Health system. Participants were drawn from UCLA Health primary and specialty care patient lists who became eligible for COVID-19 vaccination and received an initial invitation to schedule. Timing: First RCT enrolled eligible patients on the first weekday after the initial invitation (the first reminder date). Second RCT enrolled those still eligible and unscheduled on the eighth day after the initial invitation (the second reminder date). Interventions: In both RCTs, participants were randomized to receive text-message reminders or to a holdout (no reminder) condition. Messages made vaccination top-of-mind and reduced friction by including a direct scheduling link. First RCT factorial: Within the reminder arm, a 2×2 factorial tested: (a) ownership language framing (“a dose has just been made available for you” and “claim your dose”) intended to induce psychological ownership, and (b) an informational video designed to address vaccine hesitancy. Thus, reminder variants included: standard reminder, ownership reminder, standard reminder + video, ownership reminder + video. Outcomes: Primary—scheduled a first-dose appointment at UCLA Health within six days of the reminder. Secondary—received the first dose at UCLA Health within four weeks. Additional analysis assessed vaccinations anywhere (not limited to UCLA Health) and persistence over time (Kaplan–Meier analysis). Subgroup analyses examined heterogeneity by prior influenza vaccination, age, and race/ethnicity. Samples: First RCT n ≈ 93,334 included in analyses per preregistration. Second RCT n ≈ 67,092 initially, with analytic Ns varying by outcome windows. Online intention studies (n = 3,181) recruited via Amazon Mechanical Turk and Prolific in January and April 2021 to mirror intervention arms from the first RCT. Analysis: Ordinary least squares (OLS) regressions estimated effects on appointment scheduling and vaccination outcomes, reporting percentage-point (pp) differences, standard errors, P-values, and, where relevant, multiple-comparison corrections (Holm–Bonferroni). Kaplan–Meier curves assessed time-to-vaccination differences between reminder and holdout arms.

• First reminder (day ~1): Increased appointment scheduling at UCLA Health by 6.07 percentage points and vaccinations at UCLA Health by 3.57 percentage points relative to holdout; described as 84% and 26% relative increases, respectively.
• Ownership framing: Adding ownership language further increased vaccination uptake. For vaccinations anywhere, ownership framing added 0.9 pp (B = 0.009, s.e. = 0.003, P = 0.010 unadjusted; P = 0.020 Holm–Bonferroni; n = 74,605).
• Informational video: No evidence that adding a video addressing hesitancy improved appointment or vaccination outcomes beyond reminders without the video.
• Persistence: Kaplan–Meier analyses showed a persistent gap in vaccination between reminder and holdout arms over the 55-day observation window, with 18.38% in the holdout arm vaccinated at UCLA Health by day 55; effects indicate reminders increased the number vaccinated rather than merely accelerating vaccinations.
• Vaccination anywhere (first RCT): One reminder increased vaccinations anywhere by 2.1 pp relative to a 31.5% baseline in the holdout arm (B = 0.021, s.e. = 0.004, P < 0.001; n = 93,354).
• Heterogeneity: Effects were larger among participants who had received an influenza vaccine in either of the two recent seasons: +4.4 pp for appointments (B = 0.044, s.e. = 0.004, P < 0.01 for interaction) and +2.3 pp for vaccinations at UCLA Health (B = 0.023, s.e. = 0.006, P < 0.01). Effects were broadly comparable across racial/ethnic groups (white, Hispanic, Black, Asian) and across age groups (≥65 and <65 years).
• Second reminder (day ~8): Increased vaccinations anywhere by about 1.0–1.2 pp at two weeks post-reminder (e.g., B = 0.012, s.e. = 0.004, P = 0.008; baseline 12.04% in holdout), but the effect was not statistically significant at four weeks (B = 0.007, s.e. = 0.004, P = 0.127). Effects likely reflect acceleration and might have been larger had reminders facilitated vaccination across more locations beyond UCLA Health.
• Online intention studies (n = 3,181): Patterns diverged from field results; small increases in stated interest but no evidence that the ownership framing or video produced the same intention effects as the field outcomes, underscoring the limitation of hypothetical measures.

The findings demonstrate that simple, low-cost behavioral nudges—timely text reminders that make action salient and easy—significantly increase COVID-19 vaccination appointments and completions. Inducing psychological ownership over a vaccine dose further boosts impact, suggesting that motivational framing can complement friction reduction. Conversely, pairing reminders with an informational video addressing hesitancy did not improve outcomes, indicating that information-focused strategies may be less effective when added to reminders or may require different designs or delivery to influence behavior. Effects were robust across key demographic subgroups and stronger among individuals with a prior influenza vaccination history, implying reminders are particularly effective for those already favorable to vaccination but facing follow-through barriers. The persistent time-to-vaccination differences suggest reminders increase total vaccinations, not merely timing. Discrepancies between field behavior and online intentions highlight the importance of testing interventions on real-world outcomes rather than relying solely on self-reports. Overall, the results support deploying behavioral reminders at scale to improve vaccine uptake and potentially other preventive health behaviors.

This study shows that behaviorally informed text reminders substantially increase COVID-19 vaccination uptake, with ownership framing providing additional benefit, while an added informational video showed no incremental effect. By reducing friction and leveraging psychological ownership, reminders can help translate intentions into action across diverse populations. At scale, such nudges could generate millions of additional vaccinations within weeks. Future research should identify when and how to effectively integrate information-based interventions, optimize message timing and content across multiple vaccination sites, and tailor strategies for populations with high hesitancy or limited healthcare access. Field experiments should continue to validate hypothetical findings and assess long-term health outcomes and cost-effectiveness.

• Generalizability: The sample was predominantly older and white UCLA Health patients, which may limit external validity to broader, more diverse populations or health systems.
• Setting constraint: Interventions primarily facilitated scheduling within UCLA Health; effects on vaccinations anywhere may be underestimated and may differ if reminders enabled broader location options.
• Second reminder durability: The second reminder’s effect was not statistically significant at four weeks, suggesting limited sustained impact.
• Video engagement: In the field, fewer than about 21% of participants watched the informational video, limiting its potential effect and complicating comparisons with online studies where viewing could be enforced.
• Intentions vs behavior: Online intention results diverged from field behavior, highlighting measurement and context differences between hypothetical responses and real-world actions.
• Potential data/reporting noise: Some text indicates minor inconsistencies/typos in descriptive passages; however, preregistered analyses and main effects are clearly reported.