Personalized Digital Health Communications to Increase COVID-19 Vaccination in Underserved Populations: A Double Diamond Approach to Behavioral Design

Abstract

The COVID-19 pandemic exacerbated pre-existing health disparities. People of historically underserved communities, including racial and ethnic minority groups and people with lower incomes and educational attainments, experienced disproportionate premature mortality, access to healthcare, and vaccination acceptance and adoption. At the same time, the pandemic increased reliance on digital devices, offering a unique opportunity to leverage digital communication channels to address health inequities, particularly related to COVID-19 vaccination. We offer a real-world, systematic approach to designing personalized behavior change email and text messaging interventions that address individual barriers with evidence-based behavioral science inclusive of underserved populations. Integrating design processes such as the Double Diamond model with evidence-based behavioral science intervention development offers a unique opportunity to create equitable interventions. Further, leveraging behavior change artificial intelligence (AI) capabilities allows for both personalizing and automating that personalization to address barriers to COVID-19 vaccination at scale. The result is an intervention whose broad component library meets the needs of a diverse population and whose technology can deliver the right components for each individual.

Feasibility of a Reinforcement Learning–Enabled Digital Health Intervention to Promote Mammograms: Retrospective, Single-Arm, Observational Study

“…The intervention used in this study was specifically designed to address a broad range of behavioral determinants, including those more common among underserved groups, like racial and ethnic minorities, and those experiencing poor social determinants of health [ 24 ]. A test of whether that approach was successful is assessing whether engagement and behavioral responses from members of those groups are at levels equal to or greater than responses from people who are Caucasian or of higher socioeconomic status.…”

“…Although the feasibility study shows promising reach and engagement across patient subgroups, future research should focus on ensuring equitable access and support for preventive care among groups with historical experience of structural inequalities [ 44 ]. This will require engaging members of those groups to understand their barriers to action and partnering to co-design solutions [ 24 ]. Ensuring equity will also require that the data used to train interventions like the one under study, which is driven by reinforcement learning, are representative of the populations at large and that the benefits conferred are available to all [ 45 ].…”

“…Those determinants are then organized using an intervention mapping process [ 22 ] that links barriers and facilitators to evidence-based behavior change techniques (BCTs) [ 23 ]. The BCTs are operationalized into a set of message components, such as subject lines, body copy paragraphs, and visual illustrations, that form a content library [ 24 ]. Interrater agreement of the content (ie, subject lines and body content) was assessed by two trained coders [ 25 ] to ensure that each component accurately operationalized the intended BCT; agreement exceeded the acceptability threshold of κ=0.80.…”

Applications of Social Media and Digital Technologies in COVID-19 Vaccination: Scoping Review

“…Machine learning models can help quickly assess the potential output of and design a smart parking system for drive-through vaccination clinics [ 170 , 172 ]. Digital technologies were also used to examine the accessibility of vaccine registration websites to ensure that the disabled can independently schedule vaccination appointments [ 179 ], to develop a multilingual app to facilize people with limited local language skills [ 174 ], to schedule people at more suitable vaccination centers [ 171 ], to remind about the next vaccination date [ 176 ], and to provide personalized emails/messages for vaccination promotion [ 178 ].…”

“…Eleven articles [170][171][172][173][174][175][176][177][178][179][180] aimed to improve the efficiency and quality of COVID-19 vaccination services through algorithms and mHealth apps, as a support or complement to current vaccination services. Technical issues, such as operator difficulty in adopting these technologies, hindered the promotion of these technologies.…”