RES 6023 Module 5 Quantitative Design Proposal Example

A Two-County Stepped-Wedge Cluster Randomized Trial of Text-Message Reminders for Childhood Immunization Series Completion: A Quantitative Research Design Proposal

Student Name

American College of Education

RES6023: Quantitative Research Designs

Module 5 Assignment

Instructor Name

November 9, 2026

Problem and Purpose

Many children served by rural and small-city health departments do not complete the recommended vaccine series by 24 months of age. At the four clinics of our northern Nevada county health department, registry data show completion of 62% over the past three years. Reminder and recall interventions likely improve immunization rates in general (Jacobson Vann et al., 2018), but the evidence for text-message reminders comes mostly from urban settings and single vaccines, and no trial has tested series completion in a mixed rural population with many Spanish-speaking families. The purpose of this stepped-wedge cluster randomized trial is to determine whether automated text-message reminders increase series completion by 24 months among children served by eight health department clinics in two Nevada counties.

Research Questions and Hypotheses

RQ1: Does the introduction of text-message reminders increase the proportion of children who complete the recommended vaccine series by 24 months of age? H0: the odds of series completion do not differ between periods with and without reminders, after adjusting for time; H1: the odds of completion are higher in reminder periods.

RQ2: Does the effect of reminders differ by parents' preferred language or by baseline vaccine hesitancy? H0: the effect does not vary by language or hesitancy; H1: the effect varies by at least one.

RQ1 is the primary question, and the study is powered for it. RQ2 is secondary and will be interpreted cautiously, since the study is not powered to detect interactions.

Design

The study is a stepped-wedge cluster randomized trial, chosen in Module 1 because it keeps the protection of randomization while allowing every clinic to receive reminders. Eight clinics, four in each county, form the clusters. The study runs for five six-month periods. In the first, no clinic sends reminders. At the start of each later period, two clinics, chosen by random draw before the study begins, start sending reminders and continue to the end. By the fifth period all eight clinics take part. Randomization will be stratified by county so that each step includes one clinic from each, keeping the counties balanced over time.

Participants and Sample Size

Participants are children who have their first visit at a study clinic before 12 months of age during the study period, along with a parent or guardian. Children are assigned to the clinic of their first visit for the entire study, which limits contamination between clinics. Children with a documented medical exemption from one or more vaccines are excluded from the primary outcome.

Using the Hussey and Hughes (2007) method, Module 4 found that eight clinics with 85 children each per period give 80% power to detect an increase in completion from 62% to 70% at a two-sided α of .05, assuming an intraclass correlation of .02. Allowing for 8% loss to follow-up, the target is 93 children per clinic per period, about 3,700 children in all.

Intervention and Comparison

In reminder periods, parents receive automated text messages in English or Spanish seven days and one day before each scheduled vaccine visit, and a recall message if a dose becomes overdue by 30 days. Messages are short, name the clinic and give a number to call to reschedule. In comparison periods, clinics continue usual care, which includes a mailed postcard for overdue doses. Usual care continues in reminder periods as well, so the comparison isolates the added effect of text messages.

Measures

The primary outcome is series completion by 24 months, defined from the state immunization registry and checked against clinic charts in a pre-study sample, as set out in Module 3. Parent vaccine hesitancy is measured at enrollment with the Parent Attitudes About Childhood Vaccines survey, whose scores have predicted later immunization status (Opel et al., 2013); a Spanish version will be prepared through translation, back translation and cognitive interviews. Exposure is measured by message delivery receipts and a parent-reported recall item. Covariates from registration records include insurance type, preferred language, distance to clinic and the child's sex.

Procedures

Before the study begins, the clinic order will be randomized by a statistician outside both departments, and the registry-chart comparison and the Spanish survey testing will be completed. Clinic staff will enroll families at the first visit, explain the study, collect consent and offer the hesitancy survey. The messaging vendor will switch each clinic on at its assigned start date. Registry data will be extracted every six months, and final outcomes will be extracted when the last enrolled child reaches 24 months, which adds about 18 months of follow-up after the final period.

Analysis Plan

For the main test, a mixed-effects logistic regression will be fit with a fixed effect for each period, a fixed effect for the reminder condition and a random intercept for clinic, the model Hemming et al. (2015) described for stepped-wedge trials. The period effects adjust for secular trends and history, the threats Module 2 identified as most serious for this design. Child-level covariates will be added in a secondary adjusted model. The effect will be reported as an odds ratio and, because odds ratios are easily misread, also as a difference in predicted completion percentages, each with a 95% confidence interval. RQ2 will add interaction terms between condition and language and between condition and hesitancy score. Sensitivity analyses will count children lost to follow-up as unvaccinated in one run and as fully vaccinated in another and will repeat the primary model using only children whose families received at least one delivered message.

Validity Safeguards

Each threat from Module 2 has a response built into the design or analysis. Randomized and stratified clinic order addresses selection. Period effects address history and secular trends. Use of the registry, with a reporting-change notice from the state, addresses instrumentation. Assignment by first-visit clinic addresses contamination. Reporting of losses by condition, with sensitivity analyses, addresses attrition. Delivery receipts address the construct question of whether the intervention was received. External validity is supported by the two-county setting and a detailed description of the population.

Ethics

Text reminders carry minimal risk, and every clinic receives them by the end of the study. Parents will give consent for use of their child's registry data and for the survey, and families who decline will still receive whatever reminders their clinic sends, since the reminders are a service, not only a research procedure. Message content will be reviewed to avoid revealing a child's vaccination status to anyone who sees the phone. The protocol will be reviewed by the university's IRB and registered in a public trial registry before enrollment begins.

Timeline and Dissemination

Preparation, including IRB review, registry agreements and Spanish survey testing, takes six months. The five study periods run for two and a half years, and follow-up to 24 months for the last enrolled children adds about 18 months, so final results are expected roughly four and a half years after approval. Interim reports on enrollment and message delivery will go to both health departments every six months. Results will be shared with participating families in plain language, presented to the state immunization coalition and submitted to a peer-reviewed public health journal whatever their direction.

Limitations

With eight clusters, chance imbalance between clinics remains possible despite randomization. The study is powered for an 8-point effect and may miss a smaller one, though the confidence interval will show which effects remain plausible. The 24-month outcome means results will not be available for more than four years. And the findings will apply most directly to health departments serving similar populations.

References

Hemming, K., Haines, T. P., Chilton, P. J., Girling, A. J., & Lilford, R. J. (2015). The stepped wedge cluster randomised trial: Rationale, design, analysis, and reporting. BMJ, 350, Article h391. https://doi.org/10.1136/bmj.h391

Hussey, M. A., & Hughes, J. P. (2007). Design and analysis of stepped wedge cluster randomized trials. Contemporary Clinical Trials, 28(2), 182-191. https://doi.org/10.1016/j.cct.2006.05.007

Jacobson Vann, J. C., Jacobson, R. M., Coyne-Beasley, T., Asafu-Adjei, J. K., & Szilagyi, P. G. (2018). Patient reminder and recall interventions to improve immunization rates. Cochrane Database of Systematic Reviews, 2018(1), Article CD003941. https://doi.org/10.1002/14651858.CD003941.pub3

Opel, D. J., Taylor, J. A., Zhou, C., Catz, S., Myaing, M., & Mangione-Smith, R. (2013). The relationship between parent attitudes about childhood vaccines survey scores and future child immunization status: A validation study. JAMA Pediatrics, 167(11), 1065-1071. https://doi.org/10.1001/jamapediatrics.2013.2483