Author ORCID Identifier

0000-0001-7657-9184

Date of Award

7-31-2023

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Information Systems

First Advisor

Dr. Mark Keil

Second Advisor

Dr. Jong Seok Lee

Third Advisor

Dr. Aaron Baird

Fourth Advisor

Dr. JJ Po-An Hsieh

Abstract

Algorithms and AI are playing an ever-growing role in healthcare and health-related decision making. Algorithmic tools in healthcare have the potential to support preventive care and also have the potential to enable better access to healthcare. As these tools continue to shape and transform the healthcare landscape, it is important to understand how individuals interact with algorithms and the output or recommendations that they provide. Failure to anticipate human reactions to algorithms and their outputs may lead to unintended consequences, and as a result, promoting such algorithms as a means of improving health-related decisions could backfire. Essay 1 investigates how a risk assessment algorithm affects individuals’ health-protective behavior, showing that men and women respond differently. For men, the CRC risk score increased intentions to undergo CRC screening, while for women, it reduced intentions due to lowered perceived susceptibility. Essay 2 explores the link between algorithm literacy and algorithm aversion in medical decision-making. Contrary to expectations, higher algorithm literacy actually leads to greater aversion behavior. Essay 3 uses AI-generated age progression videos to encourage future health decisions. It increases willingness to engage in various future decisions but decreases the perception of connectedness to one’s future self. The three essays of this dissertation explores the outputs of algorithms in healthcare and the mechanisms by which the use of algorithms affects individuals’ health decisions. The essays collectively emphasize that encouraging people to embrace algorithmic tools to improve decision-making about the future may produce counter-intuitive results and operate through mechanisms that are, as yet, not well understood. This highlights the need for further research in the field of human-algorithm interaction, understanding how humans react to algorithms and the advice or outcomes they provide, and uncovering the underlying mechanisms behind these reactions.

DOI

https://doi.org/10.57709/35910397

File Upload Confirmation

1

Share

COinS