Paper short abstract:

Personalized learning digital adherence tools make use of data from individuals taking medication. The question is, how do individuals feel about use of data and how is security of data in the global south. We would like to develop a theoretical framework to understand acceptability of such DATs.

Paper long abstract:

Adherence to treatment for individuals with chronic or longterm conditions, such as people living with HIV, multi-drug resistant tuberculosis patients, diabetes patients and people with cardiovascular diseases, is important. Digital adherence tools have been developed for monitoring adherence and intervening when necessary. Currently, such tools are not designed to intervene in a completely personalized manner. Artificial intelligence could help to intervene when adherence is low through, for example, medication reminders or triggering health care workers to provide extra counseling.

Using machine learning for development and continuous learning through medication intake algorithms, requires continuous use of medication intake data. However, end-users may not accept such interventions as it may lead to the feeling of 'big brother is watching you'. In addition, in the global south, there may be issues related to data security and end-users may also feel threatened by that. We propose to develop a theoretical framework investigating factors that may hamper acceptability of digital adherence tools that make use of machine learning principles among people living with chronic conditions in Tanzania. We will develop an algorithm for learning adherence patterns through existing or proposed studies in Tanzania. Based on existing theoretical frameworks for acceptability of interventions and other factors that play a role in these kind of interventions, we will collect data from end-users to inform development of a new theoretical framework for acceptability of digital adherence tools including artificial intelligence. The framework will, in future, contribute to development of better accepted interventions.