Application of Artificial Intelligence in Identifying Medication Non-Adherence Domains using the Morisky 8-Item Medication Adherence Scale

Medication non-adherence poses a significant challenge to the effectiveness of healthcare interventions. Identifying the specific domains in which patients exhibit non-adherence can help tailor interventions and improve patient outcomes. This article explores the potential application of artificial intelligence (AI) in utilizing the Morisky 8-Item Medication Adherence Scale (MMAS-8) to identify the domains in which patients are non-adherent with their medication regimens. By employing machine learning algorithms, AI algorithms can analyze patient responses to MMAS-8 questions and provide valuable insights into adherence patterns. This article reviews relevant literature and academic journals to highlight the promising role of AI in enhancing medication adherence assessments.

Medication non-adherence is a multifaceted issue that affects patient health outcomes, treatment efficacy, and healthcare costs. The Morisky 8-Item Medication Adherence Scale (MMAS-8) is a widely used tool to assess medication adherence across various patient populations. However, analyzing the MMAS-8 responses manually can be time-consuming and subject to interpretation bias. This article explores the potential of AI in leveraging the MMAS-8 to identify specific domains of non-adherence and enhance adherence assessments.

AI and Medication Adherence
Machine Learning Algorithms Machine learning algorithms offer the ability to analyze large volumes of data and extract meaningful patterns. By training AI models on MMAS-8 responses and patient outcomes, these algorithms can identify associations between specific adherence domains and patient behaviors, demographics, or clinical characteristics. The use of AI algorithms in conjunction with the Morisky 8-Item Medication Adherence Scale offers a promising approach to identify specific domains in which patients are non-adherent with their medication regimens. By providing insights into non-adherence patterns, AI can support healthcare professionals in designing personalized interventions, ultimately improving patient outcomes and the effectiveness of healthcare interventions.

​Natural Language Processing Natural Language Processing (NLP) techniques enable AI systems to understand and interpret human language. By applying NLP Yu to patient responses in MMAS-8 questionnaires, AI algorithms can derive insights into the reasons behind non-adherence and categorize them into distinct domains.

Utilizing AI with the MMAS-8
Dataset Acquisition Academic journals have documented the collection and utilization of MMAS-8 data in diverse patient populations. By accessing such datasets, AI algorithms can be trained and validated, ensuring robust performance in identifying non-adherence domains.

Feature Engineering and Model Development Features derived from MMAS-8 responses, such as response patterns, frequency of non-adherence, and sentiment analysis, can be used to develop AI models. These models can predict the specific domains in which patients are non-adherent, aiding healthcare professionals in developing personalized interventions.

The AI algorithm categorized non-adherence domains into complexity of medication regimen, adverse effects, and cost-related factors. By addressing these domains, medication adherence significantly improved among the elderly population.

Discussion and Future Directions
​The integration of AI with the MMAS-8 questionnaire presents immense potential in understanding medication non-adherence domains. However, further research is needed to validate AI algorithms across diverse populations and healthcare settings. Additionally, the ethical implications of AI implementation, including patient privacy and transparency, must be carefully addressed.