AI Meets Attitudes: Decoding COVID-19 Vaccine Hesitancy in Alaska’s Diverse Communities (Preprint)
Journal of Medical Internet Research · 2025
DOI: 10.2196/81099 ↗At a glance
Journal article in Journal of Medical Internet Research, 2025. Written by Pritom Kumar Saha and 4 co-authors. Institution: Department of Computer Science, University of Texas at Dallas, Dallas, TX, United States. Major MEDLINE topics: COVID-19 Vaccines, Vaccination Hesitancy, COVID-19. 43 references. The full text is freely available in PubMed Central.
Abstract (original language)
Background
The global COVID-19 vaccine rollout faces challenges from persistent hesitancy, especially in rural and underserved regions. Alaska's unique geographic, cultural, and infrastructural challenges create complex dynamics for vaccine uptake.
Objective
This study uses machine learning on survey data to identify key sociodemographic and attitudinal predictors of hesitancy, informing targeted public health strategies.
Methods
This study surveyed 720 Alaska adults, selected via targeted sampling to capture diverse COVID-19 vaccine attitudes across demographics and regions. A structured questionnaire assessed hesitancy through 17 indicators. We applied extreme gradient boosting, random forest, and K-nearest neighbors models for both regression and classification, and interpreted classification results via Shapley Additive Explanations values.
Results
Analysis of 720 respondents showed that in Alaska, 1.8% (13/720) of surveyed individuals completed the full primary vaccination series (doses 1-3) and received all 3 booster doses. A vaccination rate of 63.47% (at least 1 dose), with Pfizer preferred over Moderna. A total of 34% (238/720) of participants reported receiving the first dose of the COVID-19 vaccine, 43% (310/720) received the second dose, 18% (130/720) received a third dose, 22% (158/720) received the first booster, 13% (94/720) received the second booster, and only 4% (29/720) received a third booster. Geographic data revealed higher uptake in urban centers and variability in rural areas. Young adult males exhibited the highest hesitancy, while lesbian, gay, bisexual, and transgender individuals showed the lowest. Trust in the health care system was the strongest predictor, confirmed by machine learning analyses.
Conclusions
Focusing on a geographically and demographically distinct US population, this study advances the scientific understanding of vaccine hesitancy while informing context-sensitive public health strategies. The findings offer actionable evidence to guide targeted communication, equitable outreach, and data-driven policy in Alaska and similarly underserved regions across the United States, underscoring the importance of culturally tailored, trust-centered interventions to promote vaccine uptake and health equity.
Subjects
MeSH terms (MEDLINE)
Author keywords
Substances
Authors and institutions
- Department of Computer Science, University of Texas at Dallas, Dallas, TX, United States.
- See Acknowledgments.
- Rutgers Global Health Institute, Rutgers University, 112 Patterson Street, New Brunswick, NJ, 08901, United States, 1 848-932-5827.
- SFI MediaFutures - Research Centre for Responsible Media Technology & Innovation, Department of Information Science and Media Studies, University of Bergen, Bergen, Norway.
Cite
APA
Saha, P. K., Guo, J., Khan, L., Robinson, R. F., & Haque, U. (2025). AI Meets Attitudes: Decoding COVID-19 Vaccine Hesitancy in Alaska’s Diverse Communities (Preprint). Journal of Medical Internet Research. https://doi.org/10.2196/81099
Vancouver
Saha PK, Guo J, Khan L, Robinson RF, Haque U. AI Meets Attitudes: Decoding COVID-19 Vaccine Hesitancy in Alaska’s Diverse Communities (Preprint). Journal of Medical Internet Research. 2025. doi:10.2196/81099
BibTeX
@article{Saha2025,
author = {Saha, Pritom Kumar and Guo, Jinghui and Khan, Latifur and Robinson, Renee F and Haque, Ubydul},
title = {AI Meets Attitudes: Decoding COVID-19 Vaccine Hesitancy in Alaska’s Diverse Communities (Preprint)},
journal = {Journal of Medical Internet Research},
year = {2025},
doi = {10.2196/81099}
}
EUVIMED