TY - JOUR
T1 - Predictive modeling and explainable AI in addressing disparities in adolescent vaccination uptake
T2 - Insights from the US national immunization survey-Teen
AU - Mathkor, Darin Mansor
AU - Sufyan, Muhammad
AU - Haque, Shafiul
N1 - Publisher Copyright:
© 2025 Elsevier Ltd
PY - 2025/9
Y1 - 2025/9
N2 - Purpose: Vaccination uptake refers to the proportion of individuals within a specific population who receive recommended vaccinations within a given timeframe. While previous studies have primarily focused on describing trends in adolescent vaccination coverage, this study introduces advanced predictive modeling and explainable AI to enhance immunization strategies. Methods: The research team explored the NIS-Teen (National Immunization Survey-Teen) with 27,565 records. After much preprocessing, exploratory data analysis (EDA), and model development, Linear Regression, Balanced Random Forest, XGBoost, and Support Vector Regression were performed on the data. Outputs of models were explained employing SHAP analysis to discern key predictors, such as vaccine type, dose history, geography, and the survey year. Results: Our analysis showed that geographic location, including state and HHS region, is the biggest predictor of adolescent vaccination coverage. This finding points to significant differences between states and regions. After geography, the type of vaccine and the survey year were also important factors. This highlights the specific challenges related to certain vaccines, like HPV, and how coverage trends change over time. Other key factors included poverty level and race or ethnicity. This underscores how social and demographic inequalities affect vaccination rates. Conclusions: Adolescent vaccination coverage in the U.S. varies based on different influences, mainly geographic location and vaccine type. These findings underscore the urgent need to move beyond one-size-fits-all strategies and toward geographically-targeted, vaccine-specific public health campaigns that address the socioeconomic and demographic barriers identified in this study. By focusing on these key drivers, public health officials can better allocate resources to close critical vaccination gaps and protect adolescent health.
AB - Purpose: Vaccination uptake refers to the proportion of individuals within a specific population who receive recommended vaccinations within a given timeframe. While previous studies have primarily focused on describing trends in adolescent vaccination coverage, this study introduces advanced predictive modeling and explainable AI to enhance immunization strategies. Methods: The research team explored the NIS-Teen (National Immunization Survey-Teen) with 27,565 records. After much preprocessing, exploratory data analysis (EDA), and model development, Linear Regression, Balanced Random Forest, XGBoost, and Support Vector Regression were performed on the data. Outputs of models were explained employing SHAP analysis to discern key predictors, such as vaccine type, dose history, geography, and the survey year. Results: Our analysis showed that geographic location, including state and HHS region, is the biggest predictor of adolescent vaccination coverage. This finding points to significant differences between states and regions. After geography, the type of vaccine and the survey year were also important factors. This highlights the specific challenges related to certain vaccines, like HPV, and how coverage trends change over time. Other key factors included poverty level and race or ethnicity. This underscores how social and demographic inequalities affect vaccination rates. Conclusions: Adolescent vaccination coverage in the U.S. varies based on different influences, mainly geographic location and vaccine type. These findings underscore the urgent need to move beyond one-size-fits-all strategies and toward geographically-targeted, vaccine-specific public health campaigns that address the socioeconomic and demographic barriers identified in this study. By focusing on these key drivers, public health officials can better allocate resources to close critical vaccination gaps and protect adolescent health.
KW - Geographic disparities
KW - Immune estimation
KW - Machine learning
KW - Predictive modeling
KW - SHAP analysis
KW - Vaccination
UR - https://www.scopus.com/pages/publications/105013380701
U2 - 10.1016/j.compbiomed.2025.110953
DO - 10.1016/j.compbiomed.2025.110953
M3 - Artículo
AN - SCOPUS:105013380701
SN - 0010-4825
VL - 196
JO - Computers in Biology and Medicine
JF - Computers in Biology and Medicine
M1 - 110953
ER -