At Vega Health, we've assembled a network of collaborators including team members, advisors, investors, and allies with decades of experience making healthcare technology actually work for clinicians and operators. Vega Conversations is a series highlighting their perspectives: what brought this group together, what we're doing differently, and how we believe AI can structurally improve healthcare.
Every year, clinicians and health system leaders solve hard problems: reducing sepsis mortality, cutting clinician burnout, and improving care transitions. And, despite the value those solutions could provide for other care delivery organizations, the knowledge often stays inside the building.
As a physician and health system executive, Dr. Chris DeRienzo knows this problem first-hand. DeRienzo trained as a neonatologist, then built his administrative career in leadership at the intersection of patient safety, quality improvement, and clinical informatics. He served as Chief Quality Officer at Mission Health before stepping outside traditional health system roles to serve as Chief Medical Officer at a startup — an experience that sharpened his thinking about the consistent problem of healthcare technology overpromising and underdelivering. He returned to North Carolina as CMO at WakeMed, leading the organization through the pandemic, before his work in health policy positioned him for his current role as Chief Physician Executive at the American Hospital Association.
“Chris was someone I crossed paths with early in my career as an MD, MPP candidate at Duke,” Mark Sendak, CEO and Co-Founder of Vega Health, said. “During my time at the Duke Institute for Health Innovation (DIHI), I saw him grow in medical and quality leadership roles at multiple community hospital systems in North Carolina. He then became the leading physician at the American Hospital Association, where I sought his advice on how to broaden the reach of Health AI Partnership, a multistakeholder AI practice network. Chris cares deeply about rural hospitals and was one of the first people to help me understand that education and technical assistance aren’t enough to support rural hospitals. To truly support the safe, effective, and ethical use of AI across ALL hospitals, you have torelieve the burden from community hospital systems to figure this out on their own. Chris shaped the concept around Vega Health and is helping us better serve community hospital systems.”
DeRienzo's career runs through exactly the kinds of institutions Vega Health aims to serve, and he’s seen patterns that repeat across the field firsthand: what works in academic medical centers and what doesn't translate to community hospitals, where implementation stalls and why, and what it actually takes to build the internal infrastructure for sustainable AI adoption.
Operational pressure prevents knowledge sharing
The healthcare industry would benefit from significantly scaling shared learnings across care delivery organizations. Especially as health systems try to best determine how to navigate a new technology like AI, each organization shouldn’t have to reinvent the wheel when determining how to procure, assess, and evaluate AI.
Health systems are not, by and large, protecting their best practices from competing health systems. Few people are intentionally keeping clinical outcomes improvements, safety protocols, or workflow redesigns a secret.
In DeRienzo’s experience, clinicians and operators who have developed something that genuinely works are almost universally willing to talk about it.
The biggest barriers remaining are time and opportunity. Health system leaders are running institutions under relentless operational pressure. Unless knowledge sharing is a literal part of the job description, the knowledgeoften stays within the walled garden out of sheer inertia. When key insights do surface through a presentationor a published paper, the audience catches a glimpse of something new and then returns to their day jobs. What they rarely get is the implementation context that would allow them to replicate it.
There's also a harder truth about transferability: knowing what worked somewhere else is useful, but the real work is in the how. Even a well-documented approach, DeRienzo estimates, may only translate about sixty percent of the way to a different health system operating with different people, different constraints, and a different institutional culture.
Part of the problem is structural. Health systems are delivery organizations first. The constant work of actually running a hospital — keeping the lights on, managing staff shortages, hitting financial targets — leaves little room for the kind of reflection and documentation that knowledge transfer requires. And unlike aviation or manufacturing, where industry bodies actively collect and redistribute safety learnings, healthcare has no reliable equivalent. There are conferences, journals, and trade associations, but none of them have crackedthe problem of getting operational knowledge out of one institution and into another in a form that's usable.
Competitive pressure certainly doesn't help. While it may not be the main barrier to sharing knowledge, health systems are competing for physicians and patients. Meanwhile, the backdrop of legal and regulatory concerns can make the effort of knowledge sharing appear riskier.
Where AI can help in the healthcare journey: the ambient scribe example
DeRienzo says that four things need to happen to complete a successful implementation in healthcare. First isa multi-stakeholder team that is involved in the discovery phase from the start. Second is a clinical championwith organizational influence who can lead peers through a workflow change.
Third is a solution that genuinely improves the lives of end users, and, finally, a way to measure whether it'sworking. He pointed to ambient scribes as a clear case for AI implementation that’s worth the lift.
His reasoning is simple: it takes something physicians actively dislike, documentation, and automates it at the push of a button. Ambient scribes also make progress in restoring the personal dimension of the provider-patient relationship.
When considering whether a workflow could benefit from AI, DeRienzo says that the question isn’t if AI could automate the task, but whether end users want it to be. Involving the end user early on is a key practice both for product design and for product implementation, he explained.
When designed and implemented effectively, AI can meaningfully improve clinician experience and replace inefficient processes.
Why knowledge sharing matters for Vega Health
Every dollar spent on an AI implementation is a dollar that doesn’t go towards other worthy investments in people, infrastructure or equipment. That’s why Vega Health exists to help health systems sustainably integrate AI into their organizations.
At Vega Health, our practical experience in the field drives one of our core beliefs: that true AI integration requires knowledge transfer between what has worked at other healthcare organizations.
Chris believes that getting AI right in healthcare is less about the technology than about the team, processes, and environment (both physical and cultural) surrounding it. The knowledge-sharing problem with workflow and technology that Chris has spent his career engaging with and working to improve is the problem Vega was built to solve.
Knowledge sharing is a core component of Vega Health’s model. We are both diffusing innovative models beyond the environments in which they were developed and proactively connecting the developers with our customers. Vega Health also makes a conscious effort to facilitate shared learnings between institutions and people addressing the same problems, even when it doesn’t directly translate into financial gain for us.
Healthcare organizations must have foresight into what it will take to effectively integrate a model before committing constrained capital towards it based on vague promises.
