Characters of Local Governments as Predictors for Improving the HPV Vaccination Rate in Japan

We conducted a nationwide questionnaire survey of 500 randomly sampled municipalities to calculate the public-funded human papilloma virus (HPV) vaccination rates in fiscal years 2011-2013. The characteristics of different municipalities were determined from the national statistical data, including those for all the municipalities in Japan. Based on the questionnaire results, we divided the fiscal 2012 HPV vaccination rate into two groups using the median value as the threshold: objective variable and municipality characteristics (29 items) as explanatory variables. Univariate and multivariate regression analyses were performed using the Akaike Information Criterion (AIC) scoring system to examine the valid explanatory variables. Univariate analysis of data from 150 municipalities indicated that the model was good because among 29 explanatory variables, AIC values were the lowest for "number of kindergartens" (AIC: -22.35), followed by "population" (AIC: -11.32), and the two-dimensional cross tables suggested that these explanatory variables were negatively correlated with HPV vaccination rates. Multivariate analysis showed that the combination of "number of kindergartens" and "number of households" (AIC: -30.29) had the smallest AIC values and were the optimal model for explaining HPV vaccination rates in local municipalities. The predictors for improved HPV vaccination rates suggested that these rates were high in municipalities characterized by a small population, large proportion of the elderly, and increased financial stability.