Increasing Adolescent Vaccination Rates via AI-Powered Communication with Patients


Adolescent Vaccination Rates: A Pressing Issue


Research shows that even before the pandemic adolescence is a time when parents have a hard time following recommended well-visits and children start to have different views with their parents about healthcare choices which can lead to lower vaccination rates for this population. However, adolescent vaccines (HPV, Meningococcal, Tdap, flu) are crucial. We do not know how adolescents and parents feel about getting vaccinated against COVID-19. Public health officials worry that the COVID-19 pandemic is negatively impacting already stagnant uptake rates of the HPV vaccine. In addition, it is unclear how parents, many of whom are hesitant towards childhood immunizations, will react to their adolescent children getting vaccinated against COVID-19. 

This Project’s Goal


This project aims to develop an innovative deep-learning algorithm and multimethod approach for modeling large amounts of electronic health records and patients’ vaccine uptake data to answer these questions. Specifically, our innovative machine-learning algorithm will alleviate the current challenges in identifying predictors of vaccine denials and refusals. 


Our overall objectives are to develop highly accurate deep-learning algorithms to predict adolescent patients’ uptakes of Human Papillomavirus (HPV) and COVID-19 vaccines to aid in:


(1) clinicians’ communication with patients and 
(2) patients’ vaccination decision-making. 


Our central hypothesis is that the use of our deep-learning enabled clinician system will increase the uptake rates of HPV and COVID-19 immunizations among adolescents aged 13-17.

How will we achieve this Goal?


Aim 1: Develop a machine-learning algorithm to predict patients’ predispositions for vaccine delay/refusal. 
Aim 2: Generate interactive and conversational scripts embedded in clinician information systems to aid in clinicians’ communications with patients. 
Aim 3: Conduct a randomized clinical trial to test our deep-learning-based interactive clinician system (Aims 1 & 2) for increasing HPV and COVID-19 vaccination rates.

Approach


Our approach includes the development of a machine-learning algorithm, installing it within a clinical information system to provide conversational assistance to physicians, and testing its efficacy for increasing patients’ immunization rates. We will conduct these research activities at MSU Health Team. 

Overall Health Impact:


Our innovative machine-learning algorithm will alleviate the current challenges in identifying predictors of vaccine denials and refusals. Our algorithm to predict patients’ vaccination predispositions will enhance a clinical information system, thereby facilitating communications about vaccines between healthcare providers and patients. Ultimately, this project will help increase adolescents’ immunization rates, reduce the costs of managing public health, and aid the prevention of epidemics/pandemics and HPV-associated cancers. 
 

Collaborators

Lee-Dongwon.jpeg

Dongwon Lee,

Ph.D.

Aiping Xiong

Aiping Xiong,

Ph.D.

Boyce Keith English

Boyce Keith English,

MD

Sigal_Yakov.jpg

Yakov Sigal,

MD

截屏2021-08-20 16.41_edited.jpg

Victoria Nelson,

JD