Personalized Medication Prompting via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional medication adherence systems fail to provide optimal prompts due to their generic nature, not considering individual user preferences, environmental contexts, and medical conditions, leading to low effectiveness in reminding patients to take their medication.
Innovation Solution
A medication prompting artificial intelligence (MPAI) system that uses machine learning to train personalized prompting models based on user-specific activities and environmental data, adjusting notifications to maximize adherence by considering current and future states of the user and their environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If rules-based algorithms are used to provide medication reminders, then the system is simple to implement, but the effectiveness of reminders is low because they do not consider individual user contexts
Solution Approach 1:
The system changes parameters by transitioning from fixed rules-based reminders to dynamic machine learning models that adjust reminder timing and content based on user state parameters (activity level, location, environmental context). This allows the system to optimize adherence effectiveness while managing complexity through automated parameter adjustment.
Solution Approach 2:
The machine learning models continuously learn from user responses and environmental data to automatically optimize reminder strategies without requiring manual intervention. The system serves itself by autonomously improving its prompting effectiveness based on accumulated data and user behavior patterns.
2Ease of operation
If generic prompts are sent to all users, then the system is easy to operate, but user engagement decreases as prompts become annoying and are ignored
Solution Approach 1:
The system applies local quality by customizing reminder characteristics (timing, frequency, modality, content) to match individual user contexts and preferences. Each user receives locally optimized prompts based on their specific behavior patterns, environmental factors, and stated preferences rather than generic universal prompts.
Solution Approach 2:
The system transitions from static generic prompts to dynamic personalized reminders that adapt in real-time based on user state changes, environmental conditions, and learned preferences. The prompting strategy dynamically adjusts to maintain optimal engagement while preserving ease of operation through automated adaptation.
3Reliability
If machine learning models are trained with user state and environmental data, then personalized prompting effectiveness increases, but data processing requirements and computational complexity increase
Solution Approach 1:
The system performs preliminary actions by pre-processing and storing user state and environmental data as it becomes available, preparing features in advance for model training. This reduces computational complexity during inference by having data ready in structured formats, allowing personalized prompting to be generated efficiently when needed.
Solution Approach 2:
The system introduces intermediaries such as feature extraction layers, data preprocessing pipelines, and model caching mechanisms that bridge raw data and the machine learning models. These intermediaries reduce computational complexity by transforming raw data into model-ready features and managing computational resources efficiently.
4Measurement precision
If continuous monitoring of user activity and environment is performed, then the accuracy of adherence prediction improves, but energy consumption and system resource usage increase
Solution Approach 1:
The system implements periodic monitoring instead of continuous monitoring, sampling user activity and environmental data at strategically determined intervals. This reduces energy consumption while maintaining adequate prediction accuracy by updating models periodically with fresh data rather than constantly processing every change.
Solution Approach 2:
The system applies partial monitoring by selectively observing only the most relevant user states and environmental factors for prediction accuracy. Instead of continuously monitoring all possible parameters, it focuses on key indicators that have the greatest impact on adherence prediction, reducing energy consumption while maintaining effectiveness.
Data Source
AI summary
Without timely reminders, certain individuals may be highly unlikely to complete various tasks scheduled for completion. For example, medicinal therapy regimen patients often have difficulty in remembering to take their prescribed medications in accordance with care provider specified instructions. By providing a reminder prompt generation system that utilizes machine-learning and/or artificial intelligence to determine appropriate reminder times and modalities for providing automated-prompting notifications to users, the users' adherence to scheduled task completions can be significantly increased.