Discipline: Technology and Engineering
Subcategory: Physiology and Health
Kailyn Cage - University of Maryland, College Park
Co-Author(s): Luis Santos, Sahaana Arumugam, Dania Morris, and Monifa Vaughn-Cooke
Introduction: Facilitating treatment adherence among patients with chronic diseases poses significant challenges to health care providers. Health Information Technology (HIT) is a critical component of chronic disease patient self-management and has the ability to support patient self-management through Personal Health Records (PHR). However, current PHR systems are not designed for the highest-risk segments of the chronic disease population (minorities, elderly, low socio-economic status). Hypothesis: This study was conducted to understand the knowledge base of the high-risk populations and utilize this information to develop and validate multiple modular software designs in the population. Methods: One-hundred and sixty-one subjects participated in the data collection segment of this research. Subjects completed a series of surveys assessing their demographics, medical history, health literacy, and technological competency. Results: Clustering algorithms for categorical data were applied and validated on the dataset specifically for medical history, health literacy, and technological competency. This analysis resulted in the selection of the Diana (Divisive Analysis) algorithm with four clusters. Each cluster represents a different software modular design persona which was developed in this research. The modular software design for each of the four personas varied between high and low levels of health complexity, health literacy, and technological competency, which influenced the information load and design orientation for each persona. Conclusion: The evaluation of the high risk patient populations is critical in the design of PHR systems. The present research identifies distinct differences in the design needs for different personas through the application of data analytics and cluster analysis. Future research is required to validate the designed personas in the same population of the initial data collection.
Funder Acknowledgement(s): FDA Office of Minority Health
Faculty Advisor: Monifa Vaughn-Cooke, mvc@umd.edu
Role: I assisted in the design and validation of the survey tool and data collection. I also assisted in the overall design and analysis method of this research. Finally, I assisted in the design of the modular software designs based on the cluster results, developing the method for segmenting for assigning design characteristics based on persona results.