Abstract
OBJECTIVES: To describe the development, validation and reliability of the Oxford Vaccine Hesitancy Scale (OVHS), a new instrument to assess vaccine hesitancy in the general population. DESIGN: Cross-sectional validation study. SETTING: Internet-based study with participants in the UK and USA. PARTICIPANTS: Demographically representative (stratified by age, sex and race) samples from the UK and USA recruited through the Prolific Academic platform. MAIN OUTCOME MEASURES: To demonstrate OVHS development, exploratory factor analysis with categorical variables and a polychoric correlation matrix followed by promax oblique rotation on the UK sample was performed. Confirmatory factor analysis with a Satorra-Bentler scaled test statistic evaluating goodness of fit statistics including the root mean squared error of approximation (RMSEA), standardised root mean squared residual (SRMR) and comparative fit index (CFI) was performed on the US sample. Reliability as internal consistency was assessed using McDonald's omega. Evidence in support of the predictive, convergent and discriminant validity of the scale was assessed using logistic regression ORs of association (OR) or Pearson correlation coefficients. RESULTS: Data for factor analysis were obtained from 1004 respondents, 504 in the UK and 500 in the USA. A scree plot, minimum average partial correlation analysis and parallel analysis suggested a three-factor 13-item scale with domains of vaccine beliefs (seven items), pain (three items) and personal deliberation (three items). Responses were recorded on a Likert scale ranging from disagree completely to agree completely, with higher score reflecting greater hesitancy. Potential total scores ranged from 13 to 65. Goodness of fit was excellent, with RMSEA=0.044, SRMR=0.041 and CFI=0.977. Predictive validity for COVID-19 vaccination status was excellent, with logistic regression ORs of association (95% CI) of 0.07 (0.04, 0.13), p<0.0001 for the UK sample for each SD increase in OVHS score, suggesting a 93% decrease in the odds of being vaccinated against COVID-19 for each SD increase in OVHS score. Convergent validity between the OVHS score and the 5C short version scale demonstrated a correlation coefficient of 0.32 (p<0.0001). Discriminant validity with an unrelated desire to perform outdoor activities demonstrated an OR (95% CI) of 1.06 (0.88, 1.29), p=0.523 for the UK sample for each SD increase in OVHS score. McDonald's omega was 0.86 and 0.87 in the UK and US samples, respectively. CONCLUSIONS: The OVHS is a feasible, valid and reliable scale for assessing vaccine hesitancy; further testing is warranted.