Deepan Kamaraj, Director, Analytics & Informatics, UPMC Enterprises

Summary

Becker’s Healthcare Podcast — Episode Summary

Guest: Deepan Kamaraj, Director, Analytics & Informatics, UPMC Enterprises
Host: Gracelyn Keller
Date: October 20, 2025
Location: Recorded live at the 2025 Health IT Digital Health & RCM Conference

---

Episode Overview

This episode features Deepan Kamaraj discussing the role of analytics and informatics in healthcare innovation at UPMC Enterprises. The conversation centers around the current landscape of artificial intelligence (AI) in healthcare, adoption and governance challenges, real-world legislative impacts, and strategic advice for leaders facing technology’s rapid evolution.

---

Key Discussion Points & Insights

1. UPMC Enterprises: Innovation Through Analytics

[00:52]

• UPMC is a large health system based in Pittsburgh, comprising 40 hospitals and up to 1,000 clinics.
• UPMC Enterprises serves as the innovation and commercialization arm, responsible for both exporting internal innovations and bringing in promising external technologies.
• Deepan’s team evaluates opportunities from a data-driven, problem-centered perspective.

  "When a company comes to us and says, hey, this is a wonderful opportunity for you to either invest or partner, we go to the EMR and say, what is the problem here? How big of a problem are they solving for the system and how big is it we think it's anticipated in the market?" — Deepan Kamaraj [01:28]

2. The Current State & Approach to AI

[02:00 – 04:25]

• AI is not new to healthcare analytics at UPMC; the organization has engaged with AI for 10–20 years.
• The recent surge in generative AI is seen more as an advanced tool than a standalone solution.
• Emphasis on trust, clinical care improvement, and data governance.
• UPMC uses "closed container" environments to rapidly test AI tools on de-identified data before clinical deployment, improving safety and confidence.
• Evaluating ROI and tool performance post-piloting is integrated into the process.

  "Within UPMC, [we] view AI as more of a tool than a solution and we'd like to evaluate it as in totality. What is the solution offering to us?" — Deepan Kamaraj [02:37] "We have design[ed] closed container systems where we can isolate data...and test these tools before we put them in front of our end users." — Deepan Kamaraj [03:35]

3. Operationalizing Innovation: Governance & ROI

[04:25 – 05:51]

• UPMC is harmonizing IT, supply chain, and other infrastructures into a single, streamlined innovation vetting process.
• AHAVI, a novel data platform, provides a secure testing ground for innovative solutions, accelerating evaluation from several months to a few weeks.

  "We've also stood up a data platform called AHAVI...for novel companies that's approaching us and saying, hey, we have a tool that we would like to test with UPMC data." — Deepan Kamaraj [05:13] "We can replicate these systems and evaluate them at scale in a really quick pace...in a matter of weeks to probably a couple of months." — Deepan Kamaraj [05:23]

4. Adapting to Legislation: U.S. and International Trends

[05:51 – 07:23]

• UPMC monitors both global and local health policies, noting that U.S. AI regulations and guardrails are still forming.
• Active adoption of European policies for model evaluation and data quality.
• Cites the FDA’s Real World Evidence program as indicative of the direction of regulation — anticipating higher standards for data provenance and evidence in AI.

  "We have started adopting from European policies as well as local policies like char, which are particularly in the context of AI." — Deepan Kamaraj [06:25] "One example is FDA's real world evidence program...what kind of data provenance are you bringing into play so your evaluation methodology can be replicable?" — Deepan Kamaraj [06:46]

5. Strategic Guidance for Healthcare Leaders

[07:23 – 08:11]

• Deepan’s core advice highlights the need to integrate data governance, model transparency, and change management for clinicians and patients alike.
• Successful implementations require transparent narratives to promote trust and adoption.

  "What's worked well for us is tying data governance to transparency of what model performance is to change management. Meaning how do you get clinicians and providers comfortable as part of workflows, along with getting these in front of patients?" — Deepan Kamaraj [07:33] "Bringing all those together in one stream of work is the challenge, but it's also been the fun part of really making progress on from pilot to meaningful work." — Deepan Kamaraj [07:44]

---

Notable Quotes & Memorable Moments

• On AI in Healthcare:

  "AI is not new to us. We've been playing within the space for about 10, 20 years now...the generative AI hype has brought on what we can do with the tool." — Deepan Kamaraj [02:23]

• On Testing and Safety:

  "We have design[ed] closed container systems where we can isolate data, large volumes of de-identified data, and test these tools before we put them in front of our end users." — Deepan Kamaraj [03:33]

• On Innovation Process:

  "We've also stood up a data platform called AHAVI within UPMC Enterprises, particularly for novel companies that's approaching us...so we've approached it from can we produce a closed container environment where we can replicate these systems and evaluate them at scale." — Deepan Kamaraj [05:13]

• On Legislative Preparedness:

  "It's really dependent on individual systems to figure out what the roadmap is and what are we comfortable with in terms of playing there." — Deepan Kamaraj [06:17]

• On Transformation Advice:

  "That would be our suggestion — to look to tie those things together in telling a very cohesive narrative." — Deepan Kamaraj [07:55]

---

Timestamps for Key Segments

• UPMC Enterprises context & role — [00:52]
• AI in healthcare, approach and challenges — [02:21]
• Testing AI tools securely & at scale — [03:35]
• Balancing innovation & operations (AHAVI platform) — [04:44]
• Legislative trends and compliance — [06:05]
• Advice for leaders & the future — [07:32]

---

Concluding Thoughts

Deepan Kamaraj’s perspective aligns innovation with pragmatism: test fast, govern tightly, and build trust with both clinicians and patients. With AI poised to transform healthcare, UPMC Enterprises demonstrates that sustainable progress comes from strategic rigor and a commitment to transparency. This episode is a practical guide for industry leaders navigating the intersection of technology, regulation, and patient care.

Transcript

[A]

Philips is a health tech leader focused on innovation that improves the health and well being of people. Our healthcare technology and informatics solutions help care teams diagnose, treat and manage more patients with greater precision, speed and confidence. Across the care journey With Philips, clinicians are empowered with streamlined insights in the moments that matter for every patient. Better care for more people Philips.

[B]

This is Gracelyn Keller with the Becker's Healthcare Podcast and we are recording live at the 2025 Health IT Digital Health and RCM Conference. I'm currently joined by Deepan C. Kamaraj who is the Director of Analytics and Informatics at UPMC Enterprises. So Deepan, thanks so much for joining me today. Let's start off by having you share a little bit more about yourself and your work in healthcare.

[A]

Of course. Well, thanks for having me. Maybe we'll start a little bit about UPMC and what we do within UPMC Enterprises to just give folks a little bit of context of what we do. UPMC we're a health system based out of Pittsburgh, about 40 hospitals, over 800 to 1000 clinics depending on how we count it. We are an idn, so we're a payer provider as well as our division, the UPMC Enterprises Division, which is the Innovation and Commercialization division within upmc. So what we do is twofold work. We take innovation that's happening within the UPMC ecosystem and bring it out to the market. But we also go out to the market and partner with companies that are doing innovative work and bring it in within that ecosystem as person overseeing of analytics and informatics. What I do is when a company comes to us and says, hey, this is a wonderful opportunity for you to either invest or partner, we go to the EMR and say we what is the problem here? How big of a problem are they solving for the system and how big is it we think it's anticipated in the market. That's what I do as a core function within UPMC Enterprises. So go from there.

[B]

Wonderful. Well, let's start our conversation today with AI Hot topic today. So nearly half of medical practices reported using AI in some capacity in the last year and it remains a key topic for health IT leaders. So from your perspective, what are the use cases that are making the most difference right now and how are you leveraging them in your organization?

[A]

Sure, I think, I mean we can talk about how hot AI is, but maybe we can talk about it in the context of what we mean by AI. Because sitting with analytics and informatics, AI is not new to us. We've been playing within the space for about 10, 20 years now, depending on who you talk to and think the generative AI hype has brought on what we can do with the tool. So within UPMC view AI as more of a tool than a solution and we'd like to evaluate it as in totality. What is the solution offering to us? That's how we look at it today. And what does that mean is how does that affect clinical care? Patient interaction as well as provider trust is primary center and focus for us in order to do those things. What can we do prior to getting these tools, tools or solutions in front of our providers? What do we need to do? How comfortable can we get in terms of confidence of what the tool is telling us it's meant to do? Those are kinds of things that we're dabbling a lot more into. And the way we're trying to do that is provide closed container environments where we can test these tools in rapid iterations rather than pull them into full production systems and go live and then evaluate what happens. We have design closed container systems where we can isolate data, large volumes of de identified data, and test these tools before we put them in front of our end users. So that's how we are approaching AI. Combined with all of that is what the system is able to do in terms of data governance. So IT scores inherent with AI, given what we need to do and how much these tools need data. And then equally important is evaluating how well these tools are doing beyond the pilot, what's the potential? So evaluating the ROI of it, I think within the system, what we've been able to do is tie all of those threads together in different chunks of work and bringing multiple leaders onto the table to inform that entire process. So that's been very informative for us and it's been an exciting journey.

[B]

Absolutely. And as virtual care expands from AI enabled tools and remote monitoring to broader digital health platforms, introducing new technology brings challenges. So what advice do you have for leaders navigating everything from governance to patient engagement? And can you share an example of how your organization has balanced innovation with operational constraints?

[A]

Yeah, absolutely. I think that's a great question in terms of tying those, those loose ends, so to speak, which have always existed in different entities from what I mean by that is solution. We have always evaluated tools within our IT infrastructure, within our supply chain infrastructure. Now we're trying to bring all of that into one streamlined process where we're but putting some hypothesis on what an ROI for a solution could be beforehand before we go into the IT Evaluation and deploying them within our solution. So that's how we're tackling that. We've also stood up a data platform called AHAVI within UPMC Enterprises, particularly for novel companies that's approaching us and saying, hey, we have a tool that we would like to test with UPMC data. How does that go? So we've approached it from can we produce a closed container environment where we can replicate these systems and evaluate them at scale in a really quick pace. When I say really quick pace, in a matter of weeks to probably a couple of months, rather than putting them in full system production for like months, months on end. Hope that addresses the question.

[B]

Yeah, absolutely. Thank you. And switching gears a little bit, how are you seeing recent legislation, both state and federal, affect health care organizations and health IT specifically, And have you adjusted strategies in response to this?

[A]

Absolutely. I think we always keep an eye on both global as well as local legislations and policies. Even what's coming up. I think in the United States, what we notice is still guardrails are evolving in terms of how do you evaluate these systems. It's really dependent on individual systems to figure out what the roadmap is and what are we comfortable with in terms of playing there. So to that end, we have started adopting from European policies as well as local policies like char, which are particularly in the context of AI. How do you evaluate these models and what else do we need to do? Is something we're paying very keen attention on. The second portion is also just paying attention to data quality. And how does that play into what are the regulations around that? So One example is FDA's real world evidence program. And what are the governance policies of when you generate evidence, be it for digital technology or pharmaceutical use cases, what kind of data provenance are you bringing into play so your evaluation methodology can be replicable? That's something we're paying close attention to at this point. We think that's not a requirement for AI companies, but we see that coming more, more sooner than later, particularly from an FDA perspective. So we're paying close attention there.

[B]

And as we wrap our conversation up, I'd love to know your top piece of advice for healthcare leaders as they prepare for future advancements in technology and rising demands for care.

[A]

Well, what I can say is, from our experience, what's worked well for us is tying data governance to transparency of what model performance is to change management. Meaning how do you get clinicians and providers comfortable as part of workflows, along with getting these in front of patients? And how do you get patients comfortable. Bringing all those together in one stream of work is the challenge, but it's also been the fun part of really making progress on from pilot to meaningful work. So that that would be our suggestion is to look to tie those things together in telling a very cohesive narrative.

[B]

Wonderful. Well, Deepan, thanks so much for joining me today on the Backers Healthcare podcast and sharing these thoughts and insights again. We are recording live at the 2025 Health IT&RCM Conference.

[A]

Thank you. Thanks for having me.