TLDR: This study analyzed 116 patient-centric e-health apps from the Google Play Store to understand the integration and maturity of AI, particularly Foundational Models. It found that over 86% of these applications are still in the early stages of AI adoption, primarily offering basic AI functionalities like chatbots or rule-based systems. Only a small fraction demonstrated advanced AI integration. While there’s a weak positive correlation between higher AI maturity and app popularity, the full potential of AI and Foundational Models in patient-centric digital health remains largely untapped, highlighting opportunities for more strategic AI adoption and further research into practical, ethical implementations.
Artificial Intelligence (AI) is increasingly becoming a part of our daily lives, and its presence in healthcare technology is no exception. However, a recent study delves into how maturely AI is integrated into applications designed specifically for patients. Understanding this maturity is crucial for assessing how trustworthy, transparent, and impactful these AI-powered health tools truly are in the real world.
The research, titled “The Impact of Foundational Models on Patient-Centric e-Health Systems”, was conducted by Elmira Onagh, Alireza Davoodi, and Maleknaz Nayebi. Their work sheds light on the current landscape of AI integration in patient-focused digital health solutions.
Exploring AI in Patient-Centric Apps
The study investigated 116 patient-centric healthcare applications available on the Google Play Store. The researchers used advanced AI models, specifically Large Language Models (LLMs), to extract key functional features from these apps. These features were then categorized according to the Gartner AI maturity model, a widely recognized framework for evaluating AI integration levels.
The findings revealed a significant insight: a vast majority of the applications, over 86%, are still in the early stages of AI integration. This means their AI capabilities are often basic, such as simple rule-based systems or static information delivery. Only a small percentage, about 13.79%, showed more advanced AI integration.
Common Features and AI Maturity Levels
The study identified 942 unique features across the analyzed applications. On average, each app offered about 11 functionalities. Some of the most common features included “Provide Health Data,” “Access Medical Record,” “View Medication,” “Symptom Tracking,” and “Schedule Upcoming Appointment.” Interestingly, many of these popular features do not require sophisticated AI, suggesting that the current value of many e-health apps lies more in their usability and accessibility than in complex AI-driven functions.
When categorizing the apps by AI maturity, the researchers found that 70 applications were in the “Awareness” stage, meaning AI discussions were informal with no pilot projects. Another 30 were in the “Active” stage, where pilot projects were underway but limited in scope. Only 11 apps reached the “Operational” stage, indicating consistent and interactive AI features. A mere five applications were classified as “Systematic,” showing a deeper, more integrated use of AI. Notably, no applications were found to be in the “Transformational” stage, which would imply a complete reshaping of operations driven by AI, such as personalized treatment recommendations.
The Link Between AI Maturity and Popularity
The study also explored whether higher AI maturity correlated with app popularity, measured by user ratings and download counts. While the correlation was weak, it was positive, suggesting that applications with more advanced AI integration tend to receive slightly better average ratings. This hints at a potential market preference for more intelligent and responsive applications, though further research is needed to establish a direct causal link.
Also Read:
- Beyond the Buzz: Understanding Large Language Models in Medicine
- Mastering AI Interaction: How Prompt Engineering Drives Productivity
Implications for the Future of Digital Health
The research highlights that while AI, especially Foundational Models, holds immense potential for patient-centric e-health, its full capabilities are yet to be realized in real-world applications. Most current AI implementations are superficial, relying on basic automation or chatbots rather than advanced predictive analytics or intelligent recommendations.
For developers and healthcare providers, this presents a clear opportunity to strategically adopt more sophisticated AI capabilities. The focus should be on functionalities that genuinely improve patient engagement, care coordination, and long-term health outcomes, rather than just technological novelty. For researchers, the study underscores the need to bridge the gap between AI innovation and practical implementation, ensuring that future AI models in healthcare are not only high-performing but also usable, safe, interpretable, and compliant with ethical and regulatory standards.
This study provides a foundational analysis of AI maturity in patient-centric e-health, offering a baseline for tracking the real-world impact of Foundational Models as the field continues to evolve. You can find more details about this research paper here.
