| Launching a New National Strategy

| Shalini Kathuria Narang

| The National Academy of Medicine (NAM) recently launched a national initiative, Patient Safety in the Era of AI, to harness the promise of AI to improve patient safety and prevent harm across all U.S. healthcare delivery systems to develop a national strategy reaffirming safety as an inviolable precondition of care. 

According to Devarashetty Shreya et al.’s Navigating the Future of Elderly Healthcare Analysis, the elderly population presents a challenge not only in numbers but also due to the specific health needs that increase with age. Diseases such as Alzheimer’s, heart disease, stroke, and diabetes increase substantially with age. According to CDC data, older adults are more likely to have multiple comorbidities that complicate treatment and care, increasing hospital admissions and healthcare costs. These factors, and expected increases in life expectancy, have profound implications for the future of healthcare, necessitating urgent reforms in care models, medical innovation, and policy.

Laura Adams, Senior Advisor, NAM, shares that although safety in applying AI is an important focus, the initiative extends beyond that to making healthcare safer for all patients, wherever and whenever they seek care. The focus will be on a limited set of high-leverage actions that can unlock cascading opportunities across the broader health ecosystem. 

Support independent journalism on aging

In 2000, the Institute of Medicine (now NAM) released To Err Is Human to elevate patient safety as a national priority in the U.S. As healthcare grows more complex, Patient Safety in the Era of AI builds on that legacy by moving closer to the vision of zero harm. “Medicine and healthcare have gotten more complex. There are more challenges than ever. With AI, we have some new tools and new hope,” says Adams, but clarifies, “AI is not a savior. We can’t simply bolt on a new technology and have it fix all things. Our initiative is focused on using AI to make care safe. It brings many new capabilities that historically have been unachievable and out of reach for us.”

Transformational Opportunities in Patient Safety

Since patient safety efforts have yielded insufficient progress, new approaches are needed for a national strategy. “What we’d like to see is that a patient could walk into any hospital in this country and have the same chance of receiving error-free care,” says Adams.

Patient safety emanates from the dynamic interplay among multiple actors, incentives, and contexts, creating feedback loops that can amplify improvements or trigger unintended consequences. AI tools can aid by discovering complex interactions within observed data, monitoring feedback loops in real time, and facilitating adaptive interventions at scale to deliver scalable, sustainable safety improvements.

”We’ve never had a national strategy for patient safety,” says Adams. Referring to equity and the well-being of the healthcare workforce, she adds, “We missed that during EHR (electronic health records) development, and paid the price.” 

The initiative goals include employing AI to identify and close safety gaps on a national level—for example, a standard national measurement system that detects safety patterns across health systems, supporting shared learning and improvement and connecting currently fragmented spaces between patients, caregivers, and the broader health system, improving transparency, communication, and downstream outcomes consistent with patient goals.

Patients Using AI

Patients are using commercially available AI tools to query health symptoms and medical diagnosis and treatment information translated to their desired health literacy and have also been aggregating their medical records and wearables data to receive personalized health advice. While this capability is transforming the way patients can direct their healthcare, it is also presenting new safety risks and requires providers in the traditional healthcare context to engage, communicate, and partner with patients in innovative ways. 

According to Adams, a major contribution to patient safety is patient agency. “If they believe there’s been a misdiagnosis, they could approach the health system with what they have learned, and it’ll have to be a conversation with their providers. No LLM is a substitute for their doctor. It can make these conversations richer, stronger, and may help. We think of AI as a transfer of power, and there’s a great deal of power being transferred to patients right now to be real partners in improving the safety and quality of their care.”

“Artificial intelligence is already reshaping care delivery and improving diagnosis,” says Sue Sheridan, CEO, Patients for Patient Safety, who serves as co-chair of the NAM initiative.   “It is also democratizing information, empowering patients, and leveraging the autonomy of patients, all of which contribute to greater patient safety. If we are intentional about how we wield these tools, AI can enable the most powerful strategy and policy levers possible to reduce preventable harm and strengthen trust in the health system.” 

Essential System Conditions for Scaling AI-Enabled Patient Safety

Some proposed system conditions for AI-Enabled patient safety at scale include:

• Aligned Governance: driven by the needs of patients and caregivers.
• Adaptive Capacity: securing flexible resources including skilled staff, advanced data infrastructure, and funding to adapt to changing threats.
• Co-created Solutions: directly with patients, communities, and frontline staff to identify safety issues and design improvements.
• Automated Visibility: use automated, transparent tracking of safety metrics to drive learning, accountability, and system-wide improvement.
• Cohesive Signaling: ensuring all external pressures, like financial incentives and reputational risks, strictly reinforce patient safety goals.

Next Steps

Ethical deployment of AI requires proactive harm-reduction strategies for maximizing its positive effects. AI Code of Conduct framework by NAM is designed as a benchmark for entities developing their own approaches to consider for inclusion and alignment when assessing internal guidance for completeness in their specific context, thereby advancing trust and minimizing the likelihood of actors across the field working at cross-purposes. Thinking of AI capabilities as a means of transfer of power to patients to be real partners in improving safety and quality of their care, Adams says it is being closely scrutinized as a candidate for what might go into a national strategy to be released next year. 

“There are ways in which we need to address persistent barriers that have been around,” says Adams. “AI may be able to help us address those barriers. What we want is AI to act as a solvent, to dissolve old dysfunctional processes, and allow them to be replaced with more efficient, more effective processes.” “Leaders of health systems respond to the system in which they’re in. There’s real work to be done in how we structure and incentivize safety.”

Shalini Kathuria Narang is a Bay Area-based independent writer and software professional. She writes about health, wellness, education and technology.