Contextually tailored text messages to augment cardiac rehabilitation: the virtual application-supported environment to increase exercise (valentine) study

We previously found text messages (TM) delivered as part of a comprehensive mobile health intervention may augment physical activity levels in patients enrolled in cardiac rehabilitation (CR) during its initiation phase, although optimal TM tailoring is unknown. Determine optimal time and environment to deliver TMs to CR enrollees after initiating CR. We conducted a micro-randomized trial that delivered TMs to low and moderate-risk patients, ages 18-75 enrolled in CR. Participants received a compatible smartwatch (Apple Watch or FitBit Versa 2) and were randomized to receive a TM or no TM at 4 user-selected time points over 6-months. TMs were meant to be actionable in real-time and were tailored on time of day, day of week, weather, and phase of CR. TMs were walking or anti-sedentary (e.g., stretching) and half were personalized with participants’ names. Our primary outcome was step count 60-minutes following a TM. To account for measurement differences between devices, we assessed results separately by device and focused on the initiation phase of the study (i.e., 0-4 weeks after enrollment). 108 participants were included with mean age 59.5 (10.7) years, 36 (32.4%) female, and 68 (63%) an Apple Watch. During the initiation phase of the study, use of TMs increased step counts by 17% in the 60-minutes after a TM for Fitbit users and by a non-significant 10% for Apple Watch users. For Apple Watch users, walking TMs significantly increased step count vs anti-sedentary TMs, and the impact of TMs was greatest for younger participants as compared to older participants (Table 1). For FitBit users, TMs were more effective at lunch vs afternoon and for CR participants who underwent valve repair (as compared to PCI or CABG) (Table 2). Personalization did not impact TM efficacy. TMs can augment the benefits of CR early after initiation with the potential for differential effects across varying contexts and scope for optimization.Table 1Moderator analysis of contextually tailored text messages to promote physical activity amongst Apple Watch users in VALENTINE study.ModeratorOverallInitiation PhaseRRRCI, p-valueRRRCI, p-valueTime of dayMorning vs Lunch0.930.84-1.02, 0.10.880.63-1.21, 0.4Morning vs Afternoon0.950.85-1.06, 0.40.970.73-1.29, 0.8Morning vs Evening0.930.84-1.03, 0.20.860.63-1.19, 0.4Lunch vs Afternoon1.020.93-1.12, 0.61.010.86-1.42, 0.4Lunch vs Evening1.000.91-1.10, 0.90.990.78-1.24, 0.9Afternoon vs Evening0.980.90-1.06, 0.60.890.70-1.11, 0.4Time of weekWeekend vs Weekday0.970.90-1.05, 0.50.960.78-1.18, 0.7Message typeAnti-sedentary vs Walking0.920.85-0.99, 0.040.950.76-1.19, 0.7PersonalizationName vs no name0.990.95-1.04, 0.80.910.79-1.06, 0.2Age categoryYoung (<65) vs Old (>65)1.040.98-1.11, 0.21.271.07-1.51, 0.006Activity statusLess active vs Active0.980.92-1.04, 0.60.910.75-1.1, 0.33Heart FailureHeart failure vs None1.030.97-1.10, 0.31.020.81-1.13, 0.9Indication for CRPCI vs CABG1.040.95-1.14, 0.40.840.56-1.26, 0.4Valve repair vs CABG1.030.93-1.15, 0.60.720.47-1.09, 0.1Valve repair vs PCI0.990.91-1.08, 0.90.850.70-1.05, 0.1Abbreviations: RRR, relative risk ratio; CI, confidence interval; CR, cardiac rehabilitation; PCI, percutaneous coronary intervention; CABG, coronary artery bypass graft. Highlighted cells indicate significant values. Open table in a new tab Table 2Moderator analysis of contextually tailored text messages to promote physical activity amongst Fitbit Versa 2 users in VALENTINE study.ModeratorOverallInitiation PhaseRRRCI, p-valueRRRCI, p-valueTime of dayMorning vs Lunch0.940.82-1.07, 0.31.020.72-1.44, 0.9Morning vs Afternoon1.040.91-1.19, 0.61.380.97-1.97, 0.08Morning vs Evening1.040.89-1.22, 0.61.180.82-1.69, 0.4Lunch vs Afternoon1.110.99-1.24, 0.071.351-1.82, 0.05Lunch vs Evening1.120.97-1.28, 0.11.150.82-1.63, 0.4Afternoon vs Evening1.010.89-1.14, 0.90.850.585-1.24, 0.4Time of weekWeekend vs Weekday0.950.87-1.03, 0.20.9360.70-1.26, 0.7Message typeAnti-sedentary vs Walking1.000.93-1.09, 0.91.070.83-1.38, 0.6PersonalizationName vs no name0.990.95-1.04, 0.80.830.66-1.04, 0.1Age categoryYoung (<65) vs Old (≥65)0.990.91-1.07, 0.70.940.78-1.13, 0.5Activity statusLess active vs Active0.950.88-1.02, 0.10.890.76-1.04, 0.14Heart FailureHeart failure vs None0.990.89-1.1, 0.90.9140.77-1.09, 0.3Indication for CRPCI vs CABG0.990.93-1.06, 0.80.8850.74-1.06, 0.2Valve Repair vs CABG1.010.92-1.11, 0.91.311.03-1.67, 0.03Valve Repair vs PCI1.020.92-1.13, 0.81.481.17-1.87, 0.001Abbreviations: RRR, relative risk ratio; CI, confidence interval; CR, cardiac rehabilitation; PCI, percutaneous coronary intervention; CABG, coronary artery bypass graft. Highlighted cells indicate significant values. Open table in a new tab