Tailored text messages may improve physical activity in cardiac rehabilitation, but personalization details matter less than expected

Texts were tailored by time of day, weekday vs. weekend, weather, and rehabilitation phase, and randomized to differ in content (walking promotion vs. antisedentary reminders) and personalization (with or without the participant's name). The primary outcome was step count in the 60 minutes following each text prompt. Surprisingly, no text characteristics—including personalization with the participant's name—or participant-level factors (age, sex, baseline activity, device type) consistently predicted responsiveness across both device types and trial phases. A trend emerged toward increased response to walking-focused texts (vs. antisedentary messages) and in younger (<65 years) and male participants, but these effects were not statistically consistent. Tailored texts did improve physical activity in the initiation phase of cardiac rehabilitation, but the authors note the effect was not explained by the specific moderators they tested, suggesting the benefit may come from the general act of tailoring rather than its particular components.

The study used smartwatch delivery and strict randomization to isolate the effect of message features: texts sent at random times eliminated selection bias, and the analysis stratified by device type (Apple Watch vs. Fitbit Versa) and rehabilitation phase because preliminary data suggested these might affect response. Interestingly, although walking-focused texts showed a trend toward higher engagement than antisedentary ones, and younger and male participants showed greater responsiveness in subgroup analyses, none of these effects reached statistical significance across all subgroups or remained consistent when comparing device types. The authors note that behavioral response to text prompts may depend on factors not measured in the study—such as social support, work schedule flexibility, or personal goal-setting—rather than on the demographic and message-design variables they tested. The effect size in the initiation phase (first ~8 weeks) was meaningful, but it's unclear whether this reflects novelty effects or sustained behavior change, since responsiveness to tailored messaging is known to decline over time in digital health interventions. The null findings for personalization (using names) are notable because this is a common and low-cost tactic in mobile health design; the study suggests it may not be necessary for engaging cardiac patients.