Developing a Knowledge Base for Artificial Intelligence to Answer Questions Following Median Sternotomy
Author
Noboa, James AlejandroIssue Date
2024Advisor
Bartlett, Courtney
Metadata
Show full item recordPublisher
The University of Arizona.Rights
Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.Abstract
Purpose: The purpose of this project was to design an AI application that could answer commonquestions that patients have about self-care, after being discharged from the hospital post- sternotomy. The hope is that having these answers would reduce the potential post-surgical complications that cause readmission to the hospital. Background: Over 12% of CABG patients and more than 14% of SAVR patients will be rehospitalized in less than 30 days. Many of these readmissions could have been avoided (Hirji et al., 2020). To reduce these complications, discharge education is utilized. In addition to standard discharge practices, AI applications can provide recommendations, and the personalization that AI brings may convince patients to follow an approved recovery protocol (Davenport & Kalakota, 2019). Methods: A Beta version of a new AI application was created to increase understanding of sternotomy discharge. The author entered 75 potential patient questions individually into the program. After each inquiry, the application provided an answer gathered from 50 scholarly articles. With each AI response, the author was able to rate the response with a “thumbs up,” “thumbs down,” or “neither.” Results: Out of 75 questions inputted, the AI application generated 75 responses (100%) that correctly or incorrectly answered the posed question. Out of 75 questions inputted into the AI application, 65 responses (86.7%) located the correct answer from the knowledge base and answered appropriately. Conclusions: Patient education should be individualized so that the material is understood. This can lead to a reduction in complications after discharge from the hospital. A way to individualize education is to utilize an AI application that answers patient questions. This application should initially be evaluated by a cardiac surgery team at a local hospital to identify any inaccuracies and weaknesses in the program. Once the necessary modifications are made, the program should be tested by a sample of real sternotomy patients, at the same hospital, to discover the usability of the application and the appropriateness of the literacy level of the AI’s responses. After final revisions of the AI application are made, it can be utilized by all sternotomy patients at the hospital.Type
textElectronic Dissertation
Degree Name
D.N.P.Degree Level
doctoralDegree Program
Graduate CollegeNursing