Establishment and validation of a prediction model for intention to accept recombinant zoster vaccine among community residents based on reasoned action approach

  Objective  To construct a prediction model for the intention of accepting recombinant zoster vaccine(RZV) vaccination among community residents based on reasoned action approach for promoting RZV vaccination in the population.

  Methods  Totally 1 480 attendees of four community health service centers were recruited with accidental sampling in three districts of Shanghai city for an online survey conducted during October – November 2022. The vaccination intention scale developed by Vissor et al was translated into Chinese and modified to collect participants′ information relevant to accept RZV vaccination. Of the 1 471 individual data from eligible respondents, 981 and 490 were randomly assigned into a training set and a test set. Lasso regression analysis on the data of training set was performed to explore the determinants of RZV vaccination intention. A nomogram prediction model for RZV vaccination intention was constructed using multivariate logistic regression analysis on the data of training set and the performance of the established model was evaluated with the receiver operating characteristic curve (ROC).

  Results  The intention to have RZV vaccination within 6 months was reported by 8.36% (123) of the eligible respondents. The results of Lasso regression analysis screened out nine predictive factors of RZV vaccination intention, including perception ability (β = 0.149), perceived infection risk (β = 0.074), moral norm (β = 0.061), decisional uncertainty (β = – 0.060), general vaccination beliefs (β = – 0.007), never having herpes zoster infection (β = 0.812), perceived cost-benefit (β = – 0.014), social norm (β = 0.019), and attitude (β = 0.003) and the nine factors were used to construct the prediction model. Based on the nomogram of the established prediction model, the possibility to have RZV vaccination within next 6 months is 9.00% for the respondents never having herpes zoster infection and with following scale items′ scores: 7 for perception ability, 7 for perceived infection risk, 7 for moral norm, 7 for decisional uncertainty, 42 for general vaccination beliefs, 28 for perceived cost-benefit, 49 for social norm, and 42 for the attitude towards RZV vaccination. The results of internal and external validation analysis showed that the established model is of good differentiation and calibration（the area under the curve AUC_{training set} = 0.717, AUC_{test set} = 0.689; χ²_{training set} = 9.061, degree of freedom df_{training set} = 8, p _{training set} = 0.337, χ²_{test set} = 12.024, df_{test set} = 8, p_{test set} = 0.150）.

  Conclusion  Perception ability, perceived infection risk, moral norm, decisional uncertainty, general vaccination beliefs, whether or not having herpes zoster infection, perceived cost-benefit, social norm, and attitude towards RZV vaccination are main predictive factors for RZV vaccination intention among community residents of Shanghai city.