OBJECTIVES: To analyze spatio-temporal patterns and rotavirus (RV) vaccination effect on the acute gastroenteritis (AGE)-hospitalization risk among Valencia Region’s population aged less than 3 years.

METHODS: A retrospective population-based study using real-world data from the Valencia Region. Data sources: population information system (SIP), vaccine information system (SIV) and Minimum Basic Data Set (MBDS) (AGE hospitalizations, CIE-9: 001-009, 558.9, 787.91 and their respective CIE-10 in 2016). Study population: Children aged less than 3 years old residing in the Valencia Region (Spain) between 2005-2016. Spatio-temporal modelling: A Bayesian binomial model adjusted by vaccination status, age, sex, health department (as random effect), hospital attraction (health center hospitalization probability as an offset variable), and the spatio-temporal component (health care district was considered as geographical unit and biennial periods as time unit). For the period effect we considered a correlation between adjacent periods. The spatio-temporal effect takes into account an order one autoregressive temporal dependence. The spatial effect in each period assumed a neighbourhood relationship between adjacent geographical zones following a Besag-York-Mollié model, which also includes the heterogeneity among zones.

RESULTS: AGE hospitalization rate was 47% (95%CI: 45-50) lower in RV vaccinees and 15% (95%CI: 13-18) lower in girls. AGE-hospitalization risk was 15% (95%CI: 12-18) and 58% (95%CI: 56-60) lower in children one and two years old compared to children aged less than one year, respectively. Hospitalization rates were strongly associated with health departments. The spatio-temporal pattern was non-random.

CONCLUSIONS: RV vaccination prevents one out of two AGE-hospitalizations. Significant variability in AGE admission was found among health departments. The observed Spatio-temporal pattern could be explained by climatic, environmental, socio-demographic or economic differences.