Introduction
- Speaker's Background: Medical student in Boston (2008-2010) with a background in software development and experience during the transition from paper to electronic health records (EHR).
- Themes: Access to health information, better data collection, and making use of existing healthcare data.
Key Points
[00:00:00] Transition to Digital Health Records
- Hospitals in Boston were transitioning from paper to EHR systems, which initially had inefficiencies.
- Speaker automated surgical ward summaries using scripts to retrieve patient data, demonstrating a concept of "permissionless innovation."
[00:03:44] Permissionless Innovation in Healthcare
- Advocated for authorized users of health systems to access data without restrictive layers, emphasizing "permissionless innovation."
- Started working on solutions to improve data integration across EHR systems.
[00:05:59] Federal Investment in EHRs
- U.S. government invested $35 billion to incentivize the adoption of EHRs (2010-2014) to make "meaningful use" of health records.
- Speaker participated in a $15 million Harvard project to create technologies for next-generation EHRs.
[00:07:33] Data Integration Challenges
- EHR systems (e.g., Epic, Cerner) function differently, making app integration difficult. The project aimed to standardize core healthcare data using APIs for broader EHR compatibility.
- Example: Pediatric blood pressure percentile app created to streamline pediatric data tracking.
[00:12:32] Policy and Regulation Changes (2016-2024)
- The 2016 CURES Act mandated that health data should be accessible "without special effort." Regulations were established by 2020.
- Information blocking was prohibited, encouraging smoother data sharing and greater app integration into EHR systems.
[00:16:32] Impact of AI on Healthcare Data
- AI’s role in healthcare: Automating data structuring and making data access easier. The speaker highlighted the growing importance of access policies over structured data management.
- Example: AI’s potential to extract structured data from images, like prescriptions, with minimal human effort.
[00:19:17] Continuous Glucose Monitoring (CGM)
- The advent of CGM technology allows for minute-by-minute glucose data collection, enabling patients and clinicians to make informed decisions.
- Speaker worked on projects sharing continuous glucose data, demonstrating the potential for better patient outcomes through high-velocity data streams.
[00:27:29] Integrating Continuous Data into EHRs
- Continuous data from devices like CGMs is not yet widely stored in EHRs, but there is potential to include this granular data in the future.
- Discussion on the gradual shift toward storing and using raw data in healthcare for better insights.
[00:30:11] AI in Clinical Workflows
- AI has potential in summarizing clinical notes, creating timelines from EHR data, and assisting in administrative tasks like charting.
- Example: AI summarizing the speaker’s concussion history from EHR data and producing accurate timelines.
[00:34:13] Improving Clinical Encounters with AI
- AI tools can record patient visits and generate summaries to help patients remember key points. This can also aid doctors by generating empathetic, detailed responses to patient queries.
[00:41:14] AI for Charting and Reducing Administrative Burden
- AI's potential in reducing clinician burnout by automating charting tasks, helping doctors spend more time with patients rather than on administrative work.
- Pilot programs show AI can create drafts for clinical notes, streamlining workflows.
[01:00:00] Closing Thoughts
- Speaker emphasized the slow pace of change in healthcare technology, particularly in adopting AI. However, rapid developments in AI could transform healthcare data management and patient care in the near future.
Conclusion
Digital health is evolving rapidly, with AI offering new possibilities for data access and automation. However, significant challenges remain in terms of system adoption, clinician engagement, and balancing patient care with administrative tasks. The future holds promise, but careful integration and continuous innovation are essential.
Digital Health: The Good, Bad & Ugly