Pilot evaluation of the text4baby mobile health program

Abstract

Background: Mobile phone technologies for health promotion and disease prevention have evolved rapidly, but few studies have tested the efficacy of mobile health in full-fledged programs. Text4baby is an example of mobile health based on behavioral theory, and it delivers text messages to traditionally underserved pregnant women and new mothers to change their health, health care beliefs, practices, and behaviors in order to improve clinical outcomes. The purpose of this pilot evaluation study is to assess the efficacy of this text messaging campaign.

Methods: We conducted a randomized pilot evaluation study. All participants were pregnant women first presenting for care at the Fairfax County, Virginia Health Department. We randomized participants to enroll in text4baby and receive usual health care (intervention), or continue simply to receive usual care (control). We then conducted a 24-item survey by telephone of attitudes and behaviors related to text4baby. We surveyed participants at baseline, before text4baby was delivered to the intervention group, and at follow-up at approximately 28 weeks of baby's gestational age.

Results: We completed 123 baseline interviews in English and in Spanish. Overall, the sample was predominantly of Hispanic origin (79.7%) with an average age of 27.6 years. We completed 90 follow-up interviews, and achieved a 73% retention rate. We used a logistic generalized estimating equation model to evaluate intervention effects on measured outcomes. We found a significant effect of text4baby intervention exposure on increased agreement with the attitude statement "I am prepared to be a new mother" (OR = 2.73, CI = 1.04, 7.18, p = 0.042) between baseline and follow-up. For those who had attained a high school education or greater, we observed a significantly higher overall agreement to attitudes against alcohol consumption during pregnancy (OR = 2.80, CI = 1.13, 6.90, p = 0.026). We also observed a significant improvement of attitudes toward alcohol consumption from baseline to follow-up (OR = 3.57, CI = 1.13 - 11.24, p = 0.029).

Conclusions: This pilot study is the first randomized evaluation of text4baby. It is a promising program in that exposure to the text messages was associated with changes in specific beliefs targeted by the messages.

Figure 1

mHealth conceptual model of behavior change for the Text4baby project. Social cognitive theory and the health belief model used to predict behavior change. Source: Evans et al, 2012 [13], reprinted by permission of Taylor & Francis (http://www.tandfonline.com).