The Effect of a Longitudinal Intervention to Influence Participants’ Beliefs, Intention, and Actual Behavior Performed within a Health Context

Abstract

Millions of people across the globe live with chronic conditions such as obesity (1.9 billion) (World Health Organization, 2018) and diabetes (420 million) (World Health Organization, 2016). This leads to burdening costs on the economy of at least $190 billion (Economic Costs of Obesity | Healthy Communities for a Healthy Future, n.d.), and $327 billion respectively (Centers for Disease Control and Prevention, 2018). An unhealthy lifestyle that includes unhealthy food choices and behaviors such as low physical activity may be contribute to the development of these conditions (Healthdirect Australia, 2018; Lin et al., 2012). Emerging technologies such as mobile health (mHealth) have shown promise in providing a platform to interact with and motivate patients to adhere to a particular self-care regimen, or to change their behavior. However, limited long-term engagement and attrition remain common bottlenecks limiting the success of related interventions. To improve our understanding of what contributes to adherence and engagement with positive health-related behaviors, several approaches have been proposed. Among them, understanding behavioral characteristics and beliefs has shown promise in predicting intention to perform such behaviors. However, this emerging body of literature is limited and requires a deeper understanding of the efficacy of using behavioral constructs to investigate intentions and actual engagement with interventions aimed at forming healthy habits or improving self-care. Therefore, my research efforts were to investigate the behavioral models and constructs that may help understand intention to use a technology to help promote a healthy behavior. To do so, I performed a review of the health beliefs models and theories used to assess usage of technology within the health context, and showed that the Technology Acceptance Model (TAM) is the most commonly used model to understand usage of technology. Additionally, a few studies recommended the Health Belief Model (HBM) to be used in conjunction with TAM in order to improve prediction of user intention. Next, the review identified important gaps in the literature starting with the fact that that beliefs are assessed at one point in time only, and that there is limited comparison of intention and actual behaviors. Following the review, four studies were conducted to address these gaps. Study #1 conducted a nationwide survey of patients with Type 1 Diabetes to assess the relationships between belief constructs and intentions to use a diabetes management technology. The next three studies were conducted to assess changes in behavioral constructs and their efficacy in predicting behavior over time. Study #2 assessed night-shift nurses’ beliefs and attitude towards an intervention to support them in avoiding drowsy driving. Study #2 revealed significant changes in nurses’ beliefs and attitudes towards an intervention that utilized education and a drowsy driving detection technology to alert nurses when they were driving drowsy. Study #3 evaluated changes in belief constructs in the context of assessing an mHealth app that helped participants self-manage their hypertension. Study #3 highlighted significant relationships between beliefs and adherence to BP measurements as well as clinical outcomes. Finally, Study #4 utilized similar methods to provide mHealth coaching to college students to enable them to self-manage their mental health. The study found significant relationships between participants’ self-efficacy and mental health outcomes, as well as significant differences in belief means across different levels of engagement with the app. The findings from this dissertation suggest that (1) the synthesized model borrowing behavioral constructs from the HBM and TAM helps significantly predict intention, (2) certain beliefs (e.g., self-efficacy, cues to action) may significantly change as a result of an intervention, (3) some of the beliefs (e.g., attitude, perceived health threat) are recurrently identified as significant predictors of intention across several studies and health contexts, and (4) behaviors and outcomes may be influenced by behavioral constructs such as self-efficacy, perceived health threat, and perceived ease of use.