Aim: This ecological study aimed to examine the geographic patterns in Human Papillomavirus (HPV) vaccination rates among boys and girls aged 15 years across locations in Australia, in addition to assessing contextual area-level factors that may explain the variations in HPV vaccination coverage.
Methods: Aggregate HPV vaccination data for Australian girls and boys aged 15 years from 2015 to 16 was obtained from the Australian Institute of Health and Welfare for each Statistical Area level 4 (SA4). A Gradient Boosting Machine learning model was applied to assess the predictors' importance for the study outcomes. Geographically weighted regression (GWR) models were run to assess whether substantially different relationships between predictors and outcomes occur at different locations in space.
Results: Completed HPV vaccination across the 88 SA4 regions ranged from 57.6% to 90.6% among girls, and from 53.6% to 85.5% among boys. The 2016 SEIFA Index of Economic Resources was the variable with the highest contribution to the predictions of both girls' and boys' HPV vaccination rates. Selected predictors explained 45% and 72% of the geographic variance in vaccination rates among boys and girls, respectively. Normalised coefficients for both GWR models showed a high variation in the associations between predictors and HPV vaccination rates across regions.
Conclusion: Socioeconomic and education factors were important predictors for HPV vaccination rates among Australian boys and girls aged 15 years, although no variable presented a uniform effect on HPV vaccination across SA4 regions. Important spatial heterogeneity in the effect of predictors was identified across the study area.
Keywords: Geographical maps; Geospatial analysis; HPV vaccination; HPV vaccination disparities; Machine learning; School-based vaccination.
Copyright © 2022 Elsevier Ltd. All rights reserved.