Abstract
The World Health Organization (WHO) defines adherence to treatment as "the extent to which a person’s behavior – taking medication, following a diet, and/or executing lifestyle changes, corresponds with agreed recommendations from a health care provider" [1, p 3]. It is a critical factor in determining health outcomes, but adherence rates are alarmingly low, averaging only 50% in developed countries and even lower in developing ones. The WHO believes that improving adherence to treatment could significantly increase clinical benefits, efficient use of healthcare resources, and quality of life. More than 300,000 mobile health apps are currently on the market, many of which focus on patient adherence. However, there is little evidence of their success. Known problems include high patient attrition and physician reluctance to rely on patient apps isolated from their EMR applications. This dissertation explores ways to overcome these problems and develop digital health solutions that improve adherence. Design science research methods are used to analyze the problems, define objectives for the digital solutions, design and develop prototypes, and evaluate the results. The contributions of this dissertation are presented in five publications. The first two analyze the problems based on an extensive literature review and in-depth interviews with 30 physicians and 32 patients. Their opinions on hypothetical digital health solutions led to the problem and solution scenarios in the first publication and nine requirements for the design of digital health agents in the second. The third publication explains the core concepts of my dissertation. It shows how the solution objectives are derived from the problems, how a solution architecture emerges from the definition of the solution objectives, and how the generically designed architecture found its way into a fully functional prototype customized for obese patients. The architecture connects the Digital Companion patient app with the physician EMR and consists of three software agents: the Health Literacy Agent, the Adherence Agent, and the Conversational Agent. These agents assist the patient in improving health literacy and adhering to treatment. The fourth publication describes a software framework that simplifies the development of complex chatbot interactions based on generative AI. These chatbot interactions serve as the main interface for patients to the Digital Companion, allowing them to report on their eating behaviors and well-being and receive adherence support. The fifth publication evaluates the first prototype of the Digital Companion as a technology-based boundary negotiating artifact. Twenty-seven patients and six physicians tested the Digital Companion solution in a real-world setting, which included two consultations and a therapy phase. Several in-depth interviews with all participants revealed the potential of the Digital Companion solution to improve patient adherence to treatment and quality of life.
Färber, Andri