HLTH 6246 Week 7 Church Program Design Rationale Example

Design Rationale for a Church-Based Blood Pressure Education Study: Why a Cluster Randomized Trial of Congregations Fits the Question, and Why Randomizing Members Within Churches and Comparing Volunteer Churches Were Each Set Aside

Student Name

Walden University

HLTH 6246: Introduction to Research in Health and Wellness Education

Instructor Name

Month Day, Year

Design Rationale for a Church-Based Blood Pressure Education Study: Why a Cluster Randomized Trial of Congregations Fits the Question, and Why Randomizing Members Within Churches and Comparing Volunteer Churches Were Each Set Aside

The Research Question

Among adults with uncontrolled hypertension who attend participating Black churches in one composite city, does a ten-session blood pressure education program led by trained lay health advisors from the congregation reduce systolic blood pressure at six months more than a single education session with printed materials?

Every design choice below is judged against this sentence. Four of its features carry the argument: the comparison between two programs, the outcome measured at a fixed point after the program, the unit of delivery, which is the congregation, and the outcome itself, a clinical measure taken by staff rather than reported by participants.

The Selected Design: A Cluster Randomized Trial

Churches, not individuals, are randomly assigned. Roughly half of the participating congregations would offer the ten-session program to eligible members; the other half would offer the single session and print materials, with the full program offered afterward. Blood pressure is measured by trained staff at baseline and six months.

The fit comes from the question's features. Because the question asks whether one program produces more change than another, random assignment is needed to make the two arms comparable at the start. Because the program is delivered to a congregation, through group sessions, announcements from the pulpit and conversations after services, assignment has to happen at the level where delivery happens. Church-based health promotion works through the life of the congregation, with partnership and congregational involvement central to recruitment and participation (Campbell et al., 2007); a program designed that way cannot be given to some members and withheld from others in the same pews.

This design has been used for exactly this kind of question. The FAITH trial assigned 32 New York City churches to a lifestyle program with motivational interviewing or to health education, enrolled 373 Black adults with uncontrolled hypertension and found a greater systolic reduction of 5.79 mm Hg in the intervention arm at six months (Schoenthaler et al., 2018). The present study borrows the architecture, not the program.

The design has costs. Members of the same church resemble each other, so their outcomes are correlated, and each additional member adds less information than an independent participant would. The sample size must account for that correlation, and the analysis must model it, usually with mixed-effects methods; trials that ignore clustering overstate their precision (Murray et al., 2004). With a limited number of churches, chance imbalance between arms is also more likely than in an individually randomized trial, which argues for matching or stratifying churches by size before randomization.

Rejected Alternative 1: Randomizing Individuals Within Churches

What it would have offered: maximum statistical efficiency. Every participant would be an independent unit, so fewer people would be needed, and chance imbalance would be small.

Why it was set aside: contamination. If half the hypertensive members of a congregation attend ten sessions and the other half do not, the second half will hear the advice in the fellowship hall, see the recipes at church dinners and be encouraged by the same lay health advisors who are their fellow members. The comparison arm would receive part of the program informally, shrinking the observed difference toward zero and biasing the study against finding an effect. Withholding a program from some members of a congregation that has agreed to host it would also strain the partnership on which recruitment depends. The question names a congregational program, and individual randomization would test something else.

Rejected Alternative 2: Comparing Churches That Volunteer With Churches That Do Not

What it would have offered: feasibility and goodwill. Churches eager to host the program would receive it, comparison churches would be recruited separately, and no congregation would be asked to wait.

Why it was set aside: selection. Churches that volunteer for a health program are likely to differ from those that do not: they may already have a health ministry, a pastor who preaches on health or members who are more engaged. Those differences could produce a gap in blood pressure change by themselves. Statistical adjustment can address the differences that are measured, such as congregation size or members' baseline blood pressure, but not commitment or culture, which are hard to measure and likely to matter most. A further problem arises because participants are selected for uncontrolled blood pressure: readings that are high at one visit tend to be lower at the next for reasons unrelated to any program, so without a randomized comparison group that drifts the same way, a decline could be mistaken for an effect.

Why the Comparison Arm Receives Something

The comparison churches receive one education session and printed materials rather than nothing, and the full program after the study. That choice follows the question, which compares two programs, and it mirrors the FAITH design, in which the control arm received health education sessions (Schoenthaler et al., 2018). It also makes the study easier to accept for congregations assigned to wait.

How Blood Pressure Will Be Measured

Because the outcome is a clinical reading, the protocol matters as much as the design. Trained staff, not the lay health advisors who deliver the program, will measure blood pressure with the same validated automated device at every church, after five minutes of seated rest, taking three readings a minute apart and averaging the last two. Measurements at six months will be scheduled at the same time of day as baseline where possible. Staff will not be told which arm a church is in, although members may mention it, so the protocol relies on the device rather than on judgment. A standard protocol keeps a difference between arms from reflecting differences in how readings were taken.

What the Design Cannot Answer

The trial will estimate the program's effect among members who enroll at participating churches over six months. It cannot say whether effects last beyond six months without longer follow-up, whether the program would work in congregations that decline to participate or which of the program's components produced the effect. Those would require different questions and, in turn, different designs.

References

Campbell, M. K., Hudson, M. A., Resnicow, K., Blakeney, N., Paxton, A., & Baskin, M. (2007). Church-based health promotion interventions: Evidence and lessons learned. Annual Review of Public Health, 28, 213-234. https://doi.org/10.1146/annurev.publhealth.28.021406.144016

Murray, D. M., Varnell, S. P., & Blitstein, J. L. (2004). Design and analysis of group-randomized trials: A review of recent methodological developments. American Journal of Public Health, 94(3), 423-432. https://doi.org/10.2105/AJPH.94.3.423

Schoenthaler, A. M., Lancaster, K. J., Chaplin, W., Butler, M., Forsyth, J., & Ogedegbe, G. (2018). Cluster randomized clinical trial of FAITH (Faith-Based Approaches in the Treatment of Hypertension) in Blacks. Circulation: Cardiovascular Quality and Outcomes, 11(10), Article e004691. https://doi.org/10.1161/CIRCOUTCOMES.118.004691