Implementation of Text-Messaging and Social Media Strategies in a Multilevel Childhood Obesity Prevention Intervention: Process Evaluation Results

Abstract

Social media and text messaging show promise as public health interventions, but little evaluation of implementation exists. The B'more Healthy Communities for Kids (BHCK) was a multilevel, multicomponent (wholesalers, food stores, recreation centers) childhood obesity prevention trial that included social media and text-messaging components. The BHCK was implemented in 28 low-income areas of Baltimore City, Maryland, in 2 waves. The texting intervention targeted 241 low-income African American caregivers (of 283), who received 3 texts/week reinforcing key messages, providing nutrition information, and weekly goals. Regular posting on social media platforms (Facebook, Instagram, Twitter) targeted community members and local stakeholders. High implementation standards were set a priori (57 for social media, 11 for texting), with low implementation defined as <50%, medium as 50% to 99%, high as ≥100% of the high standard for each measure. Reach, dose delivered, and fidelity were assessed via web-based analytic tools. Between waves, social media implementation improved from low-moderate to high reach, dose delivered, and fidelity. Text messaging increased from moderate to high in reach and dose delivered, fidelity decreased from high to moderate. Data were used to monitor and revise the BHCK intervention throughout implementation. Our model for evaluating text messaging-based and social media-based interventions may be applicable to other settings.

Keywords: African Americans; Baltimore; caregivers; evaluation studies; health promotion; obesity; process evaluation; social marketing; social media; text messaging.

Figure 1.

Social media and text-messaging intervention reach. Note. The data points were calculated by averaging the percentage of minimum high standard met for all reach standards. Implementation was considered low for <50% of the high standard. Medium is 50% to 99% of the high standard, high is 100% or above of the high standard. For texting, 2 out of 2 of wave 2 reach standards were analyzed for both wave 1 and wave 2. Based on wave 2 Facebook standards, 1 out of 6 reach standards were analyzed for wave 1; six out of 6 reach standards for wave 2. Based on wave 2 Instagram standards, 1 out of 2 reach standards were analyzed for wave 1, and 2/2 for wave 2. For Twitter, 1/6 wave 2 reach standards were analyzed for wave 1 and 6/6 for wave 2.

Figure 2.

Social media and text-messaging intervention dose delivered. Note. The data points were calculated by averaging the percentage of minimum high standard met for all dose delivered standards. Low dose delivery was defined as meeting <50% of high standard. Moderate dose delivered was 50% to 99%, and high dose delivered is ≥100%. For texting, 6 out of 7 wave 2 dose delivered standards were analyzed for wave 1 and 7 out of 7 for wave 2. For Twitter, 2/5 wave 2 dose delivered standards were analyzed for wave 1 and 5/5 for wave 2; for Instagram, 1/9 collected for wave 1 and 9/9 for wave 2; for Facebook, 2/7 collected for wave 1 and 7/7 for wave 2.

Figure 3.

Social media and text-messaging intervention fidelity. Note. The data points were calculated by averaging the percentage of minimum high standard met for all fidelity standards. Low fidelity was defined as meeting <50% of high standard, 50% to 99% (moderate), and ≥100% high. For texting, 4/4 wave 2 fidelity standards were analyzed for both waves 1 and 2. For all social media, no wave 2 fidelity standard was collected during wave 1. Six out of 6 wave 2 Twitter fidelity standards were collected for wave 2; nine out of nine for Instagram; five out of five for Facebook.