Sample Health Plan · Breast Cancer Screening
One journey. Three people. Every message rewritten.
Pick a person. Then tap the replies on the phone, as they would. Watch the journey branch. “STOP” always works right away.
Sample Health Plan✓ Verified business sender (RCS)
Journey map
“Opted out” can be reached from any step: STOP ends the journey at once. Pink lines show the path taken.
Step: Reminder
Persona test for this message
SIMULATED — ILLUSTRATIVEGuardrails in this journey: max 3 nudges in 30 days · quiet hours 8 PM–8 AM local (Janet opted in to evenings until 8 PM) · STOP / ALTO / BAJA honored instantly · no health details beyond “mammogram is due” in a text.
Why this works
| Evidence basis | Meta-analyses of text-message interventions show small-to-moderate improvements in health behaviors, including screening uptake, with tailored messages tending to outperform generic ones. Tailoring to the person’s barriers and characteristics is supported by Kreuter and by Noar et al.’s meta-analysis of tailored print. Just-in-time adaptive interventions (Nahum-Shani et al.) frame timing and context as part of the intervention. USPSTF (2024): screening mammography every 2 years for women aged 40–74. |
|---|---|
| BCTs used | 7.1 Prompts/cues · 1.2 Problem solving · 1.4 Action planning · 3.2 Social support (practical) · 5.1 Information about health consequences (kept light) · 9.1 Credible source · 12.1 Restructuring the physical environment (ride, nearby site) · 2.2 Feedback/confirmation. |
| COM-B target | Mostly Opportunity (time, transport, language, cost) and reflective Motivation (trust, self-care). Not fear. |
| Compliance design | Consent required before texting; opt-out in the first message and honored instantly; quiet hours; frequency cap; minimal PHI in SMS; Spanish opt-out keywords. TCPA consent and plan marketing rules must be confirmed by compliance for each program. |
| How HealthLiteracyCopilot would make it | A single approved journey (nodes, facts, and branch logic) is the source of truth. HealthLiteracyCopilot would rewrite each node for each persona, keep locked facts identical ($0 in-network screening, 20 minutes, prep tips), check readability and Spanish quality, run simulated persona tests (simulated participants, not patients), and export the flow to the organization’s messaging platform. |
| What a pilot would measure | Proximal: reply rate, booking rate, opt-out rate, time-to-book. Distal: completed screening (HEDIS BCS-E), disparities by language and age. Measures a pilot would track, not results or promised effects. |