Ensuring AI Safety in Healthcare: Evaluating Vendors through Governance Frameworks
The Future of Patient-Care Technology
In the evolving landscape of healthcare, evaluating vendors’ AI technologies has emerged as a pivotal task within an AI governance framework. Experts emphasize that any tools and systems provided by vendors must not only be effective but also safe and ethical—crucial elements for achieving optimal outcomes for all patients.
Understanding AI Risk in Clinical Settings
The establishment of a robust governance structure is essential for effectively managing the risks associated with AI systems used in clinical settings. A well-defined framework will categorize these risks and grade them accordingly, enabling stakeholders to safeguard patient care while integrating advanced technologies.
Four Key Categories of Risk Assessment
As part of this governance framework, experts have identified four key categories of risks that healthcare organizations must address:
- Correctness and Transparency
- Fairness and Equity
- Integrated Workflow
- Safety and Privacy
By focusing on these categories, healthcare institutions can make well-informed decisions about which AI tools to adopt, tailoring their risk mitigations to the respective technologies and workflows involved.
Harnessing AI’s Promise Safely
Glenn Wasson, the administrator of analytics at UVA Health, believes that a robust governance framework will enable healthcare organizations to leverage the advantages of commercial AI products while minimizing potential risks. Wasson, who will be speaking at the upcoming HIMSS25 conference, is an advocate for transparency in healthcare technology deployment.
Join Wasson at HIMSS25
Wasson will delve deeper into this subject during his educational session titled "Dear AI Vendors: This Is What We Need." His insights will focus on the critical need for healthcare organizations to establish effective partnerships with vendors to ensure that AI systems align with patient safety and ethical standards.
A Leader in Data Analytics
At UVA Health, Wasson oversees the curation and analysis of data essential for patient care and research. His areas of responsibility encompass everything from data operations and analytics to data science and visualization. His multifaceted approach means he is engaged with various applications of AI, from bedside predictive care to population health modeling.
The Impacts of AI Beyond Traditional Software
In a recent interview, Wasson was asked about the implications of AI in healthcare compared to older software solutions. He stated, “AI is infiltrating every aspect of healthcare and holds the potential to transform diagnoses, treatment planning, billing, and resource management.”
However, he cautioned that AI systems introduce new risks that healthcare organizations must navigate. Understanding these risks involves a complex interplay of access, bandwidth, and talent—elements often lacking in provider organizations.
Encouraging Open Dialogues with Vendors
Wasson’s session at HIMSS25 seeks to promote an open dialogue between healthcare providers and vendors. This collaboration is vital for identifying risk sources and fostering increased transparency regarding data, algorithms, and operational workflows. Such discussions can lead to enhanced confidence in the tools vendors provide.
Exploring Use Cases for AI Integration
Wasson will also highlight various use cases for AI, emphasizing how it can complement human decision-making. Applications range from diagnosis predictions and treatment selections to remote monitoring and improvements in billing procedures. By understanding the outcomes of AI-driven decisions, healthcare organizations can ensure the technology functions accurately and effectively.
Key Takeaways for HIMSS25 Attendees
Attendees at Wasson’s session will gain valuable insights into the novel aspects of evaluating AI systems—elements that traditional technology assessments may overlook. A key focus will be the framework for discussion that enables providers to critically assess risks and necessary mitigations.
The dialog will not only revolve around technical specifics but will also engage leaders and operators familiar with the intended workflows. This inclusive approach ensures that AI assessments are thorough and meaningful.
An Ongoing Commitment to Safety
Lastly, Wasson underscores that because of AI’s dynamic learning capabilities, the conversations regarding risk assessments and safety must be continuous. As technology evolves, so too must the frameworks that govern its use in healthcare settings.
Closely Monitoring the Future of Healthcare AI
Wasson’s session, “Dear AI Vendors: This Is What We Need,” is scheduled for March 4, 2024, from 10:15 to 11:15 a.m. at HIMSS25 in Las Vegas.
In a world increasingly dominated by AI, establishing a governance framework that emphasizes safety and ethical considerations is vital for ensuring that healthcare remains focused on patient outcomes and public trust. By engaging in meaningful dialogues with vendors and clarifying potential risks, healthcare organizations can confidently harness the power of AI for transformative benefits.
