Monthly Writings

Evaluations and reviews of the latest in the field.

The Future of AI in Chronic Care Management

After reading this article, you will be able to discuss the value of combining AI with multiple data streams in chronic care management and current challenges.

SUMMARY:

  • Artificial Intelligence (AI) will transform chronic care disease management (CCM).

  • The value of AI is combining multiple data streams to provide timely risk assessments and early warning prompts to better coordinate care plans between patients and clinicians.

  • The key concerns center on patient engagement for self-care and workflow integration. 

  • Validation of clinical effectiveness remains missing.


CHRONIC CARE CONSIDERATIONS:

  • In 2023, approximately 194 million (76.4%) of U.S. adults reported at least 1 chronic disease.

    • 130 million (51.4%) reported multiple chronic conditions

  • The number of chronic conditions increases with age.

  • Use of healthcare services increases with increased number of chronic conditions.

  • Self-management by patients is a valuable task to improve health, quality of life, and reduce costs.

    • Self-management efforts focus on:

      • Medication adherence

      • Lifestyle modifications

      • Diet

      • Exercise

      • Symptom identification and management

      • Active involvement in treatment decisions

ROLE OF AI:

  • The rapidly evolving advances in digital health and AI have provided tools to identify patterns and predict individual responses.

  • Unlike previously passive clinical decision support tools (CDSS), AI-CDSS can couple continuous multiple data sources (i.e., physiologic data from wearables, home devices, symptom diaries, medication history, EHR data to provide timely risk assessment and stratification and early warnings.

  • The goals of AI-CDSS in CCM are to:

    • Improve coordination of information between patient and clinician

    • Share jointly developed care plans

    • Use digital health solutions (secure messaging, teleconsultations, etc) for triage, escalations, and follow-up

    • Improve between face-to-face visits gaps

  • The strongest components of AI-CDSS in CCM are: monitoring and predictive analytics, producing:

    • Near real-time risk assessment

    • Early actionable warning prompts

  • Success of AI-CDSS for CCM depends on:

    • Patient engagement

    • Workflow integration

    • Data quality

    • Interpretability

    • Seamless integration into routine care

AI-CDSS CHALLENGES IN CCM:

  • TECHNICAL:

    • Data privacy

    • Security

    • Ignored alerts

    • Overreliance on automation

    • Unclear escalation pathways

    • Limited & heterogeneous clinical outcomes


    • Solutions:

      • Interdisciplinary collaboration & training

      • Data standardization and integration

      • Improved data quality

      • Collaborative healthcare models

      • Standardized data formats and outcome measures

      • External robust validation across patient populations, establishing clinical effectiveness


  • INCORPORATING AI-CDSS INTO CLINICAL WORKFLOWS:

    • Algorithm bias

    • Algorithm transparency

    • Informed consent

     

    • Solutions:

      • Rigorous external validation is necessary

      • Dataset composition – ensure patients are similar and grouped together

      • Clearly define how data inputs are used in model predictions

      • Ongoing de-biasing techniques, even after the system is in use


  • DIGITAL HEALTH LITERACY AND ACCESSIBILITY

    • Inequitable access to devices and/or network

    • Lack of knowledge in digital technology (patient and provider)

    • Ability to use technology


    • Solutions:

      • Voice of the patient included in designing the care model

      • Mobile options

      • Connectivity

      • Appropriate selection and training of the technology used

FUTURE DIRECTIONS & OPPORTUNITIES FOR AI-CDSS IN CCM:

  • Integration of AI-CDSS into the current healthcare system care models

    • Interoperable platforms coupling patient data from wearables and digital tools within the EHR and clinical workflows.

  • AI models with improved accuracy, reliability, and transparency

  • Attention to ethically designed AI-CDSS to minimize bias and ensure equitable access

  • Evolution of AI-CDSS to enhance patient engagement and education.

    • Integrate data from devices, diet, and daily activities

    • Provide real-time coaching

    • Reminders to encourage sustained lifestyle changes

  • Robust evaluation of long-term effectiveness, with standardized outcome measures across various patient populations.

  • Implementation of AI-CDSS fully integrated into clinical workflows

CONCLUSIONS:

  • AI-CDSS for CCM will play an increasing role in the future by coupling collected data from various devices with clinical notes, medication history, and social determinants of health.

  • Rigorous evaluations of long-term clinical effectiveness in various patient populations are needed.

  • Clinical oversight will be needed to combine the AI-CDSS CCM results with risk stratification, early warnings, diagnostic and screening to appropriate patient support.

  • Chronic diseases place a large burden on the global healthcare system.

  • Supporting patient self-management of multiple chronic diseases over a lifetime is a significant challenge.

  • The rapidly evolving digital health technologies and AI offer substantial potential to address these challenges.

  • AI in chronic care can enable personalized interventions, greater monitoring, predictive notifications, and improve patient engagement in self-care.

  • Initial results are encouraging, but challenges remain with data privacy; overreliance on automation; limited clinical outcome data; poor integration into clinical workflows; and data literacy and access.

The optimal approach is to develop an AI strategy prior to accumulating technological tools.

Let’s have a brief chat to discuss your unique situation

Erkan Hassan