Reply to: Comment on: The Halo Effect: Perceptions of Information Privacy Among Healthcare Chatbot Users

We thank Drs. Daungsupawong and Wiwanikit for their comments [1] on our paper [2], which found sociodemographic differences in how patients perceive information privacy related to information they shared with a patient-facing chatbot that was linked to the electronic health record (EHR).

Drs. Daungsupawong and Wiwanikit express concern that our study's sampling method could have affected our results. We agree and appreciate the opportunity to expand on this point. Our study enrolled existing chatbot users within a single large, multi-site healthcare system. The perspectives of those who have never used the chatbot, or who might not want to use the chatbot because of privacy concerns, are necessarily not included. This is why we were particularly cautious in interpreting our findings regarding race or ethnicity. Other studies—for example, in biobanking—have found that race or ethnicity and mistrust may result in greater privacy concerns among certain people, such as Black or African-American individuals in the United States [3]. The fact that our findings contradict other research led us to conclude that this finding should be the subject of future study within more generalizable samples or with specific attention to recruiting individuals who may avoid chatbot interactions altogether.

As a result, we agree with Drs. Daungsupawong and Wiwanikit about the need for additional research in more generalizable settings. Nevertheless, our findings also suggest caution in interpreting findings from surveys of the general public abstracted from the health systems where they receive medical care. If the “halo effect” is a true phenomenon, it would be incomplete to ask patients about generic, patient-facing chatbots delinked from their own health system; or, at the very least, it would require accounting for where patients receive care and for mistrust in health systems [4].

Regarding our study's representativeness of older adults, we also emphasize that older adults' experiences with chatbots and their preferences for privacy are not homogenous. Another group recently reported on older adults' preferences for a caregiving chatbot and stressed that older adults are not homogenously unfamiliar with or resistant to technology [5]. Rather than estimate a definitive preference for privacy among such a large and diverse group, our study describes an explanation for older adults who associate these chatbots with a particular health system.

One point deserves clarification. Our study found that individuals who reported higher educational attainment were more concerned about privacy, not less—this was incorrectly reported in the abstract but was correct in the manuscript body. The abstract was corrected online at the Journal of the American Geriatrics Society on April 4, 2025. Nevertheless, we hope that our study's findings motivate more generalizable and more inclusive research as Drs. Daungsupawong and Wiwanikit recommend.

All listed authors contributed significantly to the work. Matthew DeCamp: Concept and design, preparation of manuscript. Jessica R. Ellis: Concept and design, preparation of manuscript.

This research was funded by a “Making a Difference Grant” from the Greenwall Foundation. The funder had no involvement in study design; collection, management, analysis and interpretation of data; or the decision to submit for publication.

The authors declare no conflicts of interest.

This publication is linked to a related comment by Daungsupawong and Wiwanitkit. To view this article, visit https://doi.org/10.1111/jgs.19542.