VEText Message Framing and Covid-19 Vaccine Uptake Among at Risk Veterans

Overview

This is a prospective, randomized controlled trial that will evaluate the effectiveness of two different targeted outreach interventions aimed at increasing Covid-19 vaccine uptake. Veterans who are eligible for COVID-19 vaccine receipt by priority status (based on institutional guidelines) will be identified. Veterans must be enrolled into Veterans Health Administration (VHA) care and meet age or illness institutional priority guidelines to be eligible. Veterans eligible for enrollment into the trial will be randomized in a 1:1:1 allocation at an individual level, using permuted block randomization (with random block sizes of 3 and 6) to the following interventions: 1. Control arm - receives a text message with standard messaging 2. Arm 2 - receives a text message with a message framed for scarcity 3. Arm 3 - receives a text message with a message framed for improving the social good

OBJECTIVE: To compare the effectiveness of different text message reminders sent through an interactive mobile system (VEText) at increasing Covid-19 vaccine uptake among at risk Veterans. OVERVIEW: This is a prospective, randomized controlled trial that will evaluate the effectiveness of two different targeted outreach interventions at increasing Covid-19 vaccine uptake. Veterans who are eligible for Covid-19 vaccine receipt by priority status (based on institutional guidelines) will be identified. Veterans must be enrolled into VHA care and meet age or illness institutional priority guidelines to be eligible. Veterans eligible for enrollment into the trial will be randomized in a 1:1:1 allocation at an individual level, using permuted block randomization (with random block sizes of 3 and 6) to the following interventions: 1. Control arm - receives a text message with standard messaging 2. Arm 2 - receives a text message with a behavioral scarcity message 3. Arm 3 - receives a text message with a behavioral social good message Randomization will be stratified by VA center (VA Puget Sound or American Lake). Due to operational constraints, only one type of text message can be sent per day/batch. In order to account for any potential differences with the day of the week the text message is received, the three arm messages will be randomized by day of the week each week. For example, week 1 will have the control group message sent on Monday, the arm 2 message sent on Tuesday, and the arm 3 message sent on Wednesday. Week 2 will have the arm 2 message sent on Monday, the control group message sent on Tuesday, and the arm 3 message sent on Wednesday. Text messages will be sent based on the randomized order generated and will continue on those days of the week that the center has vaccine available. Our primary outcome of interest is the vaccine scheduling and/or completion rate within 7 days of receipt of the text message. Enrollment in the trial will occur between March 15, 2021 and May 1, 2021. Primary aims: Aim 1a: Investigate differences in vaccine scheduling/completion rates among different outreach strategies. Will test the hypothesis (HA1) that the proportion of Veterans either scheduling a vaccine appointment for a first dose or receiving a first vaccine dose varies between targeted behavioral messaging and standard messaging (arms 2 + 3 vs. control) at 7 days post randomization. Aim 1b: Investigate differences in vaccine scheduling/completion rates among different outreach strategies. Will test the hypothesis (HA1) that the proportion of Veterans either scheduling a vaccine appointment for a first dose or receiving a first vaccine dose varies between different types of behavioral messaging (arm 2 vs. arm 3) at 7 days post randomization. Secondary aims: None Sub-analyses: None STUDY POPULATION: All Veterans who are eligible for a first dose of Covid-19 vaccine by priority status (based on institutional guidelines). STUDY TIME PERIOD The enrollment period for the study is March 15, 2021 and May 1, 2021. Data collection and analysis will continue through at least 6 months post enrollment for the last Veteran evaluated. DATA SOURCES: Table: CDWWork.SVeteran.SMVIPerson Time period: Screening/enrollment Description: Contact information for kit/letter mailing and subsequent reminder calls Variable of interest: Name, address, zip code, phone number, SSN. Table: CDWWork.SVeteran.SMVIPerson Time period: Screening/enrollment Description: Demographics Analytic variable(s) of interest: DOB, age, gender, marital status, DOD (if applicable), service connectedness, copay. Table: Communicator report (not CDW table) Time period: Post randomization Description: Provides automated call information Analytic variable(s) of interest: Call date/time, Patient response, Result from non-response. Table: CDWWork.Appt.Appointment, CDWWork.Dim.Location Time period: Post randomization Description: Scheduled vaccination appointments Analytic variable(s) of interest: Appointment date/time, location. Table: CDWWork.Immun.Immunization, CDWWork.Dim.ImmunizationName Time period: Post randomization Description: Detailed vaccination receipt records Analytic variable(s) of interest: Vaccine receipt date/time, location. STATISTICAL ANALYSES AND DESCRIPTION OF MAIN TABLES Sample size analysis assuming a baseline vaccination proportion of approximately 33% among Veterans aged 65 and older (estimates based on vaccination data pulled 03/10/21) the estimated total sample size required to provide at least 80% power to detect a 5% difference between Arms 2 + 3 and the control arm is 4311 total (1437 per group). Descriptive patient-level statistics will be presented using the Pearson chi-square test for dichotomous variables and the Student's t test for continuous variables. The primary intention-to-treat analysis will use logistic regression to test the association between randomization group and vaccine scheduling/completion 7 days post randomization, adjusting for VHA center as a stratification variable. Odds ratios and predicted probabilities from this model will be reported. All descriptive and main analyses will be performed using R version 3.6.1. P-values \< 0.05 will be considered statistically significant. 1. Aim 1a Statistical Analyses Will test the hypothesis (HA1) that the proportion of Veterans scheduling an appointment for the first vaccine dose or receiving a first vaccine dose varies between targeted behavioral messaging and standard messaging (arms 2 + 3 vs. control) at 7 days post randomization. 2. Aim 1b Statistical Analyses Will test the hypothesis (HA1) that different types of behavioral messaging will have differing effects on vaccine scheduling/vaccine completion at 7 days post randomization (arm 2 vs. arm 3) STUDY OUTCOMES Primary analyses: % of first-dose vaccine appointments scheduled/vaccines completed at 7 days post randomization STUDY COVARIATES Primary analyses: Covariate of interest: intervention group indicator