(PS6-C51) Development of Health Belief Model in Context of a Global Pandemic

Background: The Health Belief Model was developed to explain decisions regarding engagement in prevention and disease detection programs (Champion & Skinner, 2008).This original model has been adapted to a variety of health concerns (Bishop et al., 2015; Champion & Skinner, 2008; Saunders et al., 2013). As governments and individuals began to respond to COVID-19 in 2020, this model was adapted to understand how individuals understood and made sense of the pandemic threat. The aim of this study was to develop and evaluate this model in a cohort of individuals recruited very early in the pandemic, and to understand whether mental health and other variables are associated with facets of this model. 

Methods: Participants (N = 334) were recruited via MTurk, on March 17, 2020, which is within a day of the first major stay-at-home orders in the United States.  The COVID HBM items were based on prior HBM measures (e.g., Bishop et al., 2014; Saunders et al., 2013). Participants also responded to a variety of other measures (e.g., Perceived Stress Scale-10 (PSS; Cohen et al., 1983), Generalized Anxiety Disorder-7 (GAD-7; Spitzer et al., 2006), and Patient Health Questionnaire-9 (PHQ-9; Kroenke et al., 1999) and other questions. HBM subscales (Perceived Susceptibility, Perceived Severity, Perceived Benefit of Action, Perceived Barriers, Cues to Action, and Self-Efficacy) were used as dependent variables in six ANCOVAs (one for each subscale), with gender, level of education, year of birth, having children under 18, political affiliation, subjective social status (SSS), and an overall mental health difficulties score (combined variable from normalized PHQ9, GAD7, and PSS scores) as independent variables.

Results: Each HBM subscale had good internal consistency (α = .70 to .90), with the exception of Cues to Action (α = .57). Mental health difficulties were significantly associated with all HBM subscales (all ps < .001) except Cues to Action (p = .09). Women reported higher Self-efficacy and Perceived Benefit (ps < .05). Higher scores on SSS were associated with lower scores on  Perceived Susceptibility and higher scores on Perceived Benefit and Self-efficacy (p < .05). Individuals identifying as Republican reported lower Cues to Action and higher Perceived Benefits (ps < .05). Older individuals reported greater Perceived Severity and lower Barriers (ps < .05). To better understand associations of mental health difficulties with HBM subscales, partial correlation analyses (with the same controls as above) were conducted. Higher scores on PSS, GAD-7, and PHQ-9 correlated with higher Susceptibility and Severity but with lower Self-efficacy and lower Benefit (ps < .01). Perceived Barriers was related to higher PHQ-9 (p < .001) but not to PSS or GAD-7. Cues to Action were related to higher PSS and GAD-7 (ps < .05) but not to PHQ-9. 

Conclusions: HBM as modified for COVID-19 can be useful for understanding both individuals’ reactions in a pandemic context and factors that might influence their reactions. Results may inform public health messaging during future significant health events, to incentivize individuals to engage in productive and protective health behaviors, especially individuals who may be vulnerable due to their mental health concerns.