SHI Fall Summit 2025 | Raajen Patel, Sickbay, HPE

Clips
More from SHI Fall Summit 2025

Raajen Patel

EVP Innovation

Sickbay, HPE

play_circle_outline Sickbay's role in collecting and delivering real-time physiologic data in hospitals.

play_circle_outline Traditional hospital informatics relies on analog data; Sickbay provides digitized solutions.

play_circle_outline Sickbay enables immediate access to previously unavailable data, improving clinician responsiveness.

Info
Transcript

Raajen Patel, Sickbay, HPE

Raajen Patel

EVP Innovation Sickbay, HPE

At the SHI Fall Summit 2025, theCUBE’s Paul Gillin sits down with Raajen Patel, executive vice president of innovation at Sickbay, to unpack how the company captures and delivers high-fidelity physiologic data inside hospitals. Patel explains how Sickbay aggregates vendor-neutral data streams (EKGs, arterial blood pressure, ventilators, NIRS, EEGs, ECMO and more), time-aligns them and provides live or near-real-time visualization for clinicians and care teams. Built to “play well with others,” Sickbay works across on-prem, hybrid and cloud environments so hos... Read more

explore Keep Exploring

What is the purpose and functionality of Sickbay software in hospitals? add

What happens to the data after it is collected and displayed on a device? add

How does your technology improve the quality of care provided to patients? add

bolt Powered by CUBE AI

Raajen Patel, Sickbay, HPE

search

Paul Gillin

>> This is theCUBE. I'm Paul Gillin. We're here in Somerset, New Jersey at the SHI Fall Summit 2025, sponsored in part by HPE Unleash AI. We're talking a lot about AI here at this conference, and the potential of AI to do good, to make life better for people on earth. One company that is doing that is Sickbay. We're joined by Raaj Patel, who is the EVP of innovation at Sickbay. You're in the medical informatics field. What specifically does Sickbay do?

Raajen Patel

>> Yeah. First, thanks for having me, Paul, and thanks to HPE and Unleash AI. We're really excited to be part of this today. Sickbay is software that lives inside hospitals, and it collects and delivers physiologic data at scale. So this is like EKGs or arterial blood pressures, all of the lines and the numerics that are streaming across the cardiac monitor device, also vents, NIRS, anything that might be hooked up to the patient. The idea is that this data is frequently lost. So instead, Sickbay is software that takes this data, time aligns it, and delivers it live or in near real time to clinicians, and it builds up a dataset over time so that the researchers and quality folks at that institution can learn more about their patient populations, develop new analytics and advanced care.

Paul Gillin

>> Now, I was surprised to learn that in this digital age, most hospital informatics, the type of physiological informatics you talk about, is actually analog. Once it's collected, that data disappears. Is that the case?

Raajen Patel

>> Yeah, yeah. So traditionally, it was a device that's a box that's kind of like a little bit like a TV. Once the data's off the screen, it's gone. That's advanced somewhat, but not uniformly across devices and not across types of devices. So Sickbay is vendor-neutral. We've written quite a few drivers, and we work with various middleware companies to draw out that data, to aggregate it, and to provide it to hospital IT, to biomed, to nurses, to doctors, whoever might need those kinds of streams of data.

Paul Gillin

>> Do you require hospitals to upgrade their infrastructure to feed the data, to allow you to capture their data? Or do you work with the existing equipment they have?

Raajen Patel

>> Yeah. So we play well with others, and so we work within the hospital infrastructure. If they're entirely on-prem, if they're hybrid, if they're in cloud, we can work with whatever they've got. The intention here is to help advance care with the resources that they have. Now, of course, as bigger and better compute engines come online, especially with the AI boom and the data boom, infrastructure that's supported by HPE, this becomes easier for hospitals to install and implement and maintain. But the core idea here, as always, is to make sure that these caregivers are getting the data they need when they need it.

Paul Gillin

>> So make this real for us. A nurse's station in the hospital has data coming in from all sorts of sensors. They want to know if something goes wrong, but they aren't necessarily interested in historical data. Why would they want have that information?

Raajen Patel

>> Yeah, it's a great question. A lot of times, if you're at the central and you see something going on live, the first thing is to take care of that patient. Then the second thing is, is this alarm real? Is this reflective of an actual trend that's going on? What happened in the last six hours? What kinds of medications or labs were drawn? How does this align with the physiologic trends that we're seeing? So that level of review during handoff, during night shift, when you're looking at multiple patients at once, when you're catching up on work, when you're documenting, when you're trying to keep everything from all of your patients in mind. This is where being able to get that visualization quickly to be able to share those views securely with your team, promotes that sort of level of team care and communication.

Paul Gillin

>> Now, what types of data do you gather? You mentioned a few earlier, such as EKG, blood pressure. What other types of data might you want to gather?

Raajen Patel

>> Yeah, data from ventilators, data from NIRS devices, from CRT machines. I think we've integrated with some ECMO devices, which these are starting to get in the healthcare space known, but fairly esoteric, but then also EEGs. What we like to say is, anything with a timestamp and a number, we can collect and we can display.

Paul Gillin

>> How fast do you sample this data?

Raajen Patel

>> We sample at native resolution. So when a lot of EKGs are like 250 to 500 hertz, and then EEGs get into the thousands, the numerics are more like once a second, once every other second. But all this data is collected and delivered at native resolution. So we don't decimate, we don't remove any of the data points. It's presented, it's collected at its measurement's frequency.

Paul Gillin

>> How much data can a patient generate?

Raajen Patel

>> That's a great question. So we were looking into this a little bit. If you're, let's say, 60 beats a minute, and each one of those beats is one per second, that's 240 samples every single second across an entire day, but then across 1,000 beds or 2,000 beds. So this starts to become quite a lot of information in the time series area. Now, of course, it's not the biggest dataset in hospitals, but when you start recording at that fidelity, the ways that you format this information, the ways that you store it start becoming a consideration because this can get, for lack of a better phrase, out of control pretty quickly with respect to your storage needs.

Paul Gillin

>> What kind of results have your patients or have your customers told you they're having? How is your technology making the quality of care better?

Raajen Patel

>> Yeah. The first thing that we're doing is, this is a dataset, or this is a data stream that's previously unavailable at scale. So clinicians could go to the EMR. If they happen to be at the bedside, they could go to the monitor, but otherwise it was whatever might've been documented or not. And maybe that was in live, maybe that was 12 hours ago. Not sure. So the first benefit is to be able to show this data immediately, not just to the clinician, but to the care team in charge of that patient. So maybe a specialist is at home or off-site, and they can take a look at it, those kinds of things. So that's the first benefit, is just be able to surface this new data to folks in a way they can immediately act upon. The next benefit is that over time, with respect to we're seeing all these patients that have a particular physiology, or we're seeing all this particular type of patients in this geographic area, and these patients have these specific conditions, you can start looking into, well, does our policy and practice reflect the best care that we can provide? Are there tweaks that we can make to our policy or practice that improve that care over time? The first thing is, how do we improve patient care immediately with less operational burden? And the next is like, how do we improve our practices over time so that more patients can be seen, more patients can be cared for, patients to be cared for better, with better outcomes?

Paul Gillin

>> So geek out with me a little bit here, what's the underlying technology? Are you using Kafka to capture the streaming data? You have a time series database? What's the backend?

Raajen Patel

>> Yeah, that's a great question. Sickbay has been around for quite a long time, and Kafka is a great tech. Time series databases are really good. We started out before those were in wide use. So our backend is a combination of our own file format that is open to our clients, and then actually a Postgres database, which turns out-

Paul Gillin

>> Postgres.

Raajen Patel

>> Yeah, turns to be quite performant. But in that sense, we are working with clients. We partner well with other device manufacturers. They have APIs, things like that. Some of them are using these kinds of backends, especially some of the latest and greatest. And we're happy to work with them to integrate effectively with our system, because ultimately the goal is to provide our users with this data so that they can make the best decisions for their patients. That's our goal ultimately, is to get this to those folks.

Paul Gillin

>> You're collecting a huge amount of data, and a lot of hospitals are resource-constrained right now. What do you do to minimize the storage and processing costs?

Raajen Patel

>> Yeah, it's a great question. So part of us working well with others is to make sure that we're fitting the needs of the IS department, the IT department, as well as the clinicians. Some clients, they're big research clients, and they want to make sure that they maintain all of this data indefinitely for various projects. In that case, maybe they decide that some of it goes into cold storage, some of it stays in hot storage. Some clients, they are just using it clinically, and they're not as interested in the research part, which is totally fine. And then they can scale up or down their storage as they need. One of the things we're really excited about is, over time storage is becoming cheaper. It's becoming easier to keep this data and easier to access this data, thanks to advances in infrastructure. So our general stance is to maintain this dataset, or our general recommendation rather to our clients, is to maintain this dataset. But again, we work with them, and we work with the kind of resourcing that they've got to make sure that they're getting the most out of what they bought, frankly.

Paul Gillin

>> Well, one of the interesting things about what you're doing is, you are really breaking new ground. You're capturing data that has not been captured before. What is the research community doing with this? What has their response been to having these datasets finally available?

Raajen Patel

>> Yeah, they're really excited about it, right? Because traditionally, there's clinical notes and there's what's in the EMR, EHR, there's images, and there's genomics. The time series data has been largely missing from these datasets. So the idea that you can get this sub-second granularity of understanding of the patient's physiology leads to greater and greater insight. In fact, the research tools that we provide and support have supported, I think, somewhere over a hundred papers, presentations, and posters, and has won around, I want to say, and I use these round numbers, but it's over $10 million in grant funding. So we've supported quite a few clients doing lots of interesting things in pediatrics, in adult, in cardiac, in neuro, wherever the specialty might be in the interest of the researcher. The wide-scale, sort of at-scale collection and access and frictionless delivery of this data enables them to get the subset of data that they want, run the analytics that they're interested in, and find interesting things about their patient population.

Paul Gillin

>> That must be very exciting to you fueling all this new innovation in research.

Raajen Patel

>> I come from a math background, but I also want whatever I build to be useful. So whenever I see another paper, I'm very excited that we've built something that folks are using to advance the state of the art in medicine, to advance our understanding of the human body, and to ultimately take better care of people.

Paul Gillin

>> You're part of the HPE Unleash AI ecosystem. What is your relationship with HPE?

Raajen Patel

>> Yeah. We're really excited to be part of HPE. It's a fairly new relationship, but I think one that really draws to the strengths of HPE as infrastructure and then our ability to collect a dataset that's new and that can be applied to build more trusted AI. AI, of course, is one of these things that's dominating the tech conversation, but for sure, AI is only going to be as good as the data that feeds it. And this is one of those datasets that, because it's being measured at the device level, because it's really measuring the actual physiology of the patient, can help support that kind of trust in whatever recommendations the AI may make at the end of it. So it's that nice joining, I think, of technology and a use and an infrastructure that will allow these things to come to fruition.

Paul Gillin

>> The company is Sickbay. Where can our viewers find you online?

Raajen Patel

>> Yeah, sickbay.com. It's spelled like Sickbay. Our founders and many of us at the company, I would say, all of us are big Star Trek fans. We're fans of the type of promise of technology that that particular show likes to posit. So, sickbay.com, you can find out a lot more, and you can get in touch with us.

Paul Gillin

>> Dr. Scott would be proud. Raaj Patel, thanks very much for joining us today.

Raajen Patel

>> Thanks for having us, Paul.

Paul Gillin

>> I'm Paul Gillin. This is theCUBE, the source for news, tech news, and analysis.