Innovating in Community Hospitals: AI Adoption and Digital Transformation with Sam Ash, M.D.

Sam Ash, M.D., is one of a rare breed of health system chief information officers who remains a practicing physician. Besides leading the IT team at South Shore Health in Weymouth, Mass., he also is a pulmonary and critical care medicine doctor. Ash recently sat down with Healthcare Innovation to discuss the challenges and opportunities community health systems like South Shore face in keeping up with larger health systems on the adoption of AI and other digital health innovations. 

Healthcare Innovation: Are there some benefits to taking on the CIO position while still being a practicing physician? Can you see and feel the impact of IT decisions right away?

Ash: Exactly as you just said, it is really helpful to be able to understand what is the immediate impact and what are the needs. I will give you an example. We recently implemented a new alert in Epic to alert people to changes that might happen in the insurance landscape that fortunately didn't end up happening. It was really well intentioned, and people were really glad that it was done, but I had a clinical shift that night and noticed that every 30 seconds I was getting one of these alerts, and it was incredibly disruptive. So immediately being able to make a change and decrease that impact is one small example, but one that resonates with our providers, because they often feel put upon by technology instead of helped by technology. So being able to say to them, “Look, I'm one of you. I feel your pain” — and feel the benefit other times really helps. 

HCI: We regularly interview CMIOs and chief AI officers about AI implementations, but mostly at large academic medical centers. Some people express concerns that community hospitals aren't going to have the resources and could be left behind. Could you talk about that from your perspective at South Shore? Are you concerned about whether you will be able to keep up with where AI is going in healthcare?

Ash: I think it takes a concerted effort. I do think it's a challenge for us in community health systems to keep up. We have a relatively small IT division compared to a place like MGB [Mass General Brigham], and we're trying to solve nearly all the same problems that they're trying to solve. If we think about implementing a new ambient dictation system, we have one or two analysts instead of a team of analysts who are working on that, so it's a challenge. We can certainly buy the same things that they can buy, but on the internal level, it's a little bit harder to make sure that we have the resources to do those things. 

That said, we're fully committed to trying to do everything possible, and there are benefits to being smaller, too, in terms of being more nimble. We don't have to corral multiple hospitals and multiple groups to get them on the same page about what we might want to adopt. Using ambient as an example, we have a pilot with Microsoft DAX that's been going really well. That product is an expensive product, and when we think about trying to scale it to the entire health system, cost is certainly something that we think about. That said, because we're smaller and a little bit more nimble, we're able to go out to other competitors and meet with them. We actually have another vendor coming in later this week to do some demos with us, and we didn't need to coordinate across five hospitals to get everyone on the same page. I think that's where the opportunity is — to find these ways of being more nimble. It makes it a little bit easier to adopt new technologies.

HCI: Do you have to weigh the pros and cons of working with startups that may have interesting new technology vs. the more established vendors? 

Ash: There are risk with both approaches, and for us it's about balancing that risk. Many of these startups have been around for five years or less. I do ask myself: are they going to be around for another five years? I hope so, because I think competition is what's going to bring down prices and improve the product. We're weighing how heavy the lift is for implementation. If it's something that is a relatively light implementation, then maybe we can go with something that is a little bit earlier stage. Then if we need to pivot, we can. But, for example, we’re in conversations right now to think about revamping our unified communication strategy. That’s obviously a very heavy lift, and we wouldn't want to do that with someone who is not going to be with us for the long term.

HCI: I understand you implemented a new AI-enabled MRI scanner. Can you talk about that?

Ash: If you've ever watched a spy movie and they are looking at CCTV footage, and someone in the room says “enhance” and all of a sudden you can see the newspaper print better, that is what this does. It has been a real win for everyone, because it allows for scans to be acquired faster. So the patients really enjoy it, because they can have their MRI in considerably less time. The technologists enjoy it because they can move more patients through faster, which also obviously helps the return on investment. The radiologists like it because it make the image clearer and crisper for them to read. The AI is not providing any sort of read or anything. It's really just making it easier for the radiologist to see. From an ROI standpoint, it's been incredibly beneficial to us as a community hospital that is really the only place to go for about a million people on the South Shore.

HCI: Imaging is an area where there's been a lot of early AI adoption...

Ash: Yes. There are some that I think of as sort of sneaky AI. So just using an example from my own clinical work, ultrasound has become a huge part of what we do in the ICU, whether it's bedside echos or during procedures. We recently upgraded our bedside ultrasound equipment, so that it includes a couple of artificial intelligence tools, one to automatically measure ejection fraction and another to measure the cardiac output of the heart. So that's one that is not super obvious to the whole health system, but for us in the ICU it is really beneficial, and the ER uses that as well. 

HCI: Do you have an AI governance committee to vet new tools? 

Ash: We have an AI governance committee that I chair. It also includes our chief medical officer, our chief nursing officer, and folks from compliance, risk and legal, as well as operations. Any new AI tool needs to come through that committee and be reviewed. We think a lot about the patient safety and privacy. 

In a prior life, I was part of a early stage software development company that worked on an AI tool for lung cancer detection. Those companies need data in order to build their tools. And obviously there's a tension there between that and between us wanting to make sure that we protect our patients’ privacy, so we we do a lot of work with our compliance team to make sure that our patients’ data are not leaking into the vendors, and making sure that all of our data stays in our system.

HCI: What about cybersecurity? Do you have a CISO on staff? Or does that fall under your responsibilities?

Ash: We do have a CISO who does roll up to me. He's fantastic. His name is Bob Sanderson. He's been with us for quite some time, and he has an incredible depth of experience in the area. He's an integral part of project meetings to make sure that everything that we're doing and planning is safe and secure.

HCI: What about data and analytics in support of population health and value-based care arrangements. Is that something South Shore has invested in quite a bit already, or is that an area of focus going forward?

Ash: I'd say it's more of an area of focus growing forward. We are a mandatory participant in the new Medicare TEAM program. That and a few other initiatives have really raised the recognition of how important this is. We've spent a fair amount of time building out our analytics platforms and are actively recruiting for several new business intelligence analysts to help with that process.

One of the new generative AI tools that Epic has is their SlicerDicer Sidekick, where you can query data in Epic just by asking a question, and it will provide the report for you. We plan to implement that in March 2026 to help our end users more effectively use data.

Circling back to your initial question around being a community hospital, one thing that we're thinking about over the next couple years is how we enable more effective self-service — whether that's through a new ticketing system or through tools like the one I just mentioned in Epic. We recognize that we're a small team, so it's hard for us to always be able to provide white glove support to the end user. We aspire to do that, but we recognize our limitations in terms of bandwidth. We are trying to make sure that we provide tools that enable the end user to do what they need to do, and do it effectively with our support, but a little bit more self-sufficiently.

HCI: What are some top of your priorities going into 2026? 

Ash: We are excited about implementing Epic’s new generative AI suite, which we are going to kick off in February of 2026. We’re going to start by trying to roll out a tool for each of our user bases. So there'll be a nursing-focused tool with the end-of-shift summary, an inpatient provider-focused tool with our hospital core summary, and the ambulatory provider-focused tool with a pre-visit summary. Then there'll be a rev cycle-focused tool for the automatic denial letters and coding, so we are really trying to spread the love and make sure that we're thinking about all of our our different constituents.

HCI: So is all of that going to happen in February?

Ash: No, we will kick off the project in February. We think it'll probably be a six- to nine-month implementation. 

HCI: I was going to say, it didn’t sound like you could just turn those all on at once and be done with it. 

Ash: No. Going back to being a physician in the CIO role, one of the things that I think is helpful is being able to communicate to the providers, to business operations, to everyone, that it's not just that we just turn it on, right? I think there's often this thought that you just flip a switch, or there's one line of code or what have you. We need to think about what the clinical workflow is or the business workflow that it impacts, and how we're going to adapt or change that workflow in order to make sure that the technology works the way that we intend it to.

I think that's one of many constant struggles is making sure that people understand we can't just turn things on or turn things off, because they all have unintended consequences —  even things as simple as an operating system, which, we and others are having to deal with right now, with the sunsetting of Windows 10. Making the change to Windows 11 has numerous ramifications across the system and requires extensive testing. Many people have experiences with their own home PC, and they think you just do the upgrade and it's done and you're good to go. That's clearly not the case in any big organization, but especially in healthcare, where we have so many dependencies and things that need to interact, but we'll get there.