Tuesday, March 26, 2019

Image system monitoring that tells you what you really want to know

Kyle Henson, CEO of Intelligent
Kyle Henson’s career spans than evolution of radiology from PACS to VNAs and enterprise imaging. He started as an IT project manager at Humana, and AGFA, before becoming a consultant for a number of IDNs (Integrated Delivery Networks) including installing the first virtual McKesson servers at a site in Boise, ID. He has been a partner at Secant Healthcare managing data migrations and more recently served as director of enterprise imaging for Tenet Healthcare, a Dallas-based healthcare system operating more than 500 healthcare facilities nationwide. Last July, he and his partner Mike Dolan recently left institutional imaging to form a new company, Intelligent Imaging. Together they have developed an end-to-end proactive DICOM diagnostic and monitoring tool they call Heartbeat. It is an advanced web-based platform designed to support the imaging ecosystem; from imaging center, to hospital department, to healthcare enterprise. The beta-version was well received at RSNA last fall the company launched the product at HiMSS 2019 to much acclaim. The company will exhibit at SIIM 2019 at booth 607.

Q. After several years of hype about VNA's and enterprise imaging we now seem to have entered into the “VNA is dead” phase of the hype cycle, as someone who has been working in enterprise imaging for a long time where are we going with medical imaging?

A. I think like all technology we have entered into this phase that I would call the froth of disappointment. There was a lot of hype, a lot of over-promising that VNA would be this panacea that was going to solve all the problems in the world and no technology does.

So now there's a little bit of a backlash and right now we see a trend of what they call enterprise PACS, which is a different way of saying there's either a single vendor or a best-of-breed. It's really the same old conundrum. Do you just buy one vendor and accept that it may be a little less than optimal for every single thing, but offers a single point of contact, or you choose to buy the best-of-breed for each different department?

There are actually really good reasons to go either direction. I don't want to say one is right or one is wrong. It's what's right for you and your organization. There's a lot of reasons why you might choose one or the other depending on the internal resources, your different integrations.

I have observed from my experience in healthcare, however, that with the speed with which we go through acquisitions and divestitures the idea of being on a single vendor is very difficult to achieve because you know, as soon as you buy another organization or another facility you have to then do a complete rip and replace and do a forklift upgrade to bring them onto that single vendor solution.

So coming from a provider position, it is very difficult to stay with a single vendor even when you want to standardize on a particular system. Consequently, I tend to lean a little bit more towards the best-of-breed because there's no one solution that is ever perfect for every provider and for every situation, so I guess in my experience I expect that even if you want to standardize it will never be perfect. And so we have to be accepting and understanding that there will always be integration and that's just how it is, we'll always be in some state of flux.

Q. One of the challenges or the disappointments between the promise and delivery is that VNA's are a lot more complex. You have a lot more components to monitor and take care of. What are the problems that administrators face when they're managing a VNA?

A. The VNA itself is not a simple three-button system. It's not a (computer) mouse. It's not three buttons and you move it around. It is its own thing; I liken it to a Swiss army knife. You can buy one that only has three or four tools, or you can buy one that's got 20 or 30 tools to it. Most people when they buy a VNA they buy the one that has the most tools possible. As a result, you get a Swiss army knife that is this big. With that level of features and that level of configuration, you add to your own complexity. So, it's a self-fulfilling promise you get these features and these abilities and then you start creating your own environment, which happens to be very complex.

It's a self-made problem and in some ways, it's a good problem. We had the same thing, when you think back when routers were being implemented. Everybody said, ‘wow neat toy and we can do so many things with it’, but we create very complex environments for ourselves and usually for good reason, but that adds to the complexity which means we then have to manage that.

Q. Is part of the problem of monitoring being able to ensure timely identification of problems and have the ability to perform efficient troubleshooting and fixes. What are some of the day-to-day problems administrators work with?

A. Well, the first thing is most systems have some kind of self-monitoring. In fact, almost every time you buy a system. The vendor will tell you that they self-monitor and they all do but they all do monitor themselves from a very centric point of view. Typically what is those systems do is what I call environmental monitoring. They report on the CPU, the free space on the C drive, the memory utilization, and number of threads on the cue. All that's important information but looking at the CPU of an application is like telling me that a car is running because the engine is warm.

That is information, but it could be warm because it's sitting in the sun or could be warm because you drove it and parked it an hour ago. It doesn't actually tell you the most important thing and what we should all really focus on, which is, are images moving and are doctors viewing images?’ Everything else is a distraction.

The other thing with a lot of the monitoring tools is that being computers and being automated they do things automatically very well. On a normal day with no problems, my systems would send me between 140 and 200 emails. So that all becomes noise. That's a distraction. And guess what? It doesn’t take very long before all humans just tune that out. Now I will tell you that on a bad day, I got over 1800 emails. Guess what? Didn't read them.

In the end, when there was a problem, we might be able to go back and find relevant information, but it was a search for key relevant information. It wasn't in any way predictive. It was all backwards, so that's another problem with monitoring.

Another way of looking at this is do you know who the testers are in any environment? Who are actually testing your system? It is the doctors. Think about that. Those are the most highly paid testers in the history of man, because you have these doctors that use the system and when there's a problem, they call IT and that's often the first notification of a problem. Consequently, I feel like, as an IT person, I have failed my users. If I need a doctor to tell me my system's down, then I'm not doing my job. And so that that led me on a quest to find a better way, there has to be a better way than waiting for a doctor to tell me that there's a problem.

And of course, like I said, what was out there was monitoring in terms of CPU and C drive space, that information doesn't tell me what I need to know. Those were some of the problems that I found out when I went out to the market to find better tools, so I launched a program and a company to basically to come up with a solution to monitoring that would be more useful to administrators.

Q. What did you come up with and how does it work?

A. Aha, you want to secret sauce? So, I went to the market and found that there was nothing out there that was forward-looking and was functional. Nothing was looking at images and moving images, all the solutions were backwards looking, either they provided analytics, or again, environmental monitors.

So, seeing that there was nothing out there doing what I wanted to do, I took a leap of faith and left the position I loved and really enjoyed, and said, there's a got to be a way to solve this problem. And so we went off and we built this company and called it Intelligent Imaging, and we called our product Heartbeat because it's just an IT term for that type of test.

As to how it works, Heartbeat essentially sends something out and brings it back. Does the system do what it's supposed to do? We do what we call synthetic testing, which means I don't ever see PHI (personal health information) I don't want to see PHI because I'm creating a closed loop test.

A simple example is a study that goes to PACS, then goes to the VNA or the archive, or whatever it is. It should go from PACS to the archive and the archive should let you retrieve it. It is a closed loop test and we set a service level, we say it should take this long tell me if it takes longer.

It's really that simple we're using the system the way it was designed. I guess I skipped a step there. The first step is let's check PACS. Are you running? We put a study in the PACS, and we retrieve it back; if you can do that PACS is running. We just tested 80 percent of its functional use. You give it an image. It gives it back, wonderful. Then we looked at the VNA or archive. Hey archive, here's a study give it back. That boomerang has then told me that PACS is running and the VNA is running. So that's great information. Now, I know that two components are good.

The number one problem that I always had in my environment, however, was PACS would stop sending anything. It would be great in responding back to the doctors, but it wouldn't connect to other devices. So, let's start checking integrations. I put it here. It should go there. I'm going to check for it there and now we have three distinct transactions are providing three different pieces of information for each device and the integration and we've automated that entire solution.

Q. How does an administrator view that information? How do they monitor when they get the information and tells them that the PACS is running the VNA is running the integrations working?

A. Right, so I never would want to spam somebody with continually reporting, ‘I'm running, I'm good, I'm good.’ Heartbeat only notifies you when there's a problem. There's a simple dashboard green is good, red is bad, yellow is slow. So slow is a notification on the dashboard, we don't text you. Now when we find an issue, say when PACS is down, that is an all hands-on deck situation.

When a system is down seconds count and so when there's an actual outage the administrator will get an email and a text message if you choose, you can decide but it will text your cell phone. It will say this system is down and more importantly, it will tell you what specific application is down. Is it the router, is it the PACS, is it the archive, is it the integration between the router and the PACS? Then it actually has a link in the text that you can click on your phone. It'll take you to the staff dashboard which will give you more information about not only which one is down, but what are some of the things around it. This way your people know where to focus their attention.

By contrast in typical troubleshooting, you know, a doctor notifies you that the system is down and then you have to work backwards and say well is it viewer, no? Is it the archive? no, is the PACS? no. You have to work backwards every single time until you find where the issue is.

What we're trying to do is focus a spotlight and say we found a problem it's here. That way your PACS IT team and your vendor can focus all their efforts on one place. We know that the problem is either on server A or an application here. Spend your time there and then you have a plethora of tools already that let you do the deep dive.

Heartbeat is trying to focus attention and I think avoid what is death to any IT person when you have vendor A on one phone, vendor B on the other phone and you can only say y'all figure it out; that just never works. You have two people arguing and get very emotional.

Heartbeat aims to take all of that out of the situation. No more finger pointing; It simply says there is a problem here without any blame or judgment. Let's go find out what the problem is because it could be the C drive, it could be antivirus, it could be any one of these things. It's not about blame. It's about focusing everybody's attention in the right spot at the right time so that we can fix the problem before the doctors even know, and that's what we're really here for.

Q. Your system gives the notification to the administrator before he gets the call from the doctor?

A. Exactly.

Q. What precisely is Heartbeat? Is it a piece of software that the customer installs on their network or do they subscribe to your system like an ADT security monitoring kind of service?

A. We could run it in the cloud. In fact, I have a server that runs in the cloud and mobile checks cloud archives and all that, but ideally, it is intended to be sitting at the data center next to the archive and another one possibly at the hospital next to the PACS because I want to take the network out of the loop. I don't want to have to go through a firewall. I want to remove as many variables as possible.

It is a very small footprint. You won't believe me, but takes one core and two gigs of RAM, that's all it is. It is a super small footprint. It takes just a tiny transaction to check the plumbing. Does the plumbing go from A to B. I don't need a big pipe for that. We can run in the cloud, but I'd rather be on premises in a very small VM (virtual machine).

Then we look at it as a monthly service. It’s a small monthly transaction fee to just continually check all the different devices and notify the team.

Q. So it is a subscription proposition then where you subscribe to it and you install a piece of software on the network inside the system and then provide the round-the-clock monitoring?
A. Yes.

Q. What is the pricing structure? How do people contact you.

A. If you're interested and you want to talk to me you can find me find me anywhere. I'm pretty active on LinkedIn. You can email Kyle.hanson@Imagingheartbeat.com or sales@Imagingheartbeat.com, and you will find us at our website imagingheartbeat.com.

In terms of cost and the value proposition. It's important to look at what is the cost of downtime? According to the American College of Radiology, the average down time event is three-and-a-half hours. According to the Journal of General Digital Imaging, facilities have between three and seven of those three-and-a-half hour down times per year.

We all have downtime and there's a massive amount of revenue loss due to downtime.

You can go to Aunt Minnie and search for the data, but for a 200-bed hospital it comes to about three hundred thousand dollars in revenue per year due to down times. That's not soft numbers. That's not satisfaction, that's revenue. That's the value proposition, we're looking to decrease the downtime to get that revenue back into the hospital.

My goal is to take that three-and-a-half hour down time and make that two-and-a-half hours, or an hour, or 30 minutes. That's what our software is all about. Instead of looking for where the problem is. We will tell you before your users know. There's a problem here and it could be on VNA server 3 and we will get as specific as down to an individual server. Now the customer and vendor can get together and fix that problem. It's usually really simple once you get to that level it doesn't take long to fix the problem or remove that server. That's what we're all about, we reduce that down time.

Q. I assume you have a demo that people could come in and see, or can get access, or do you go to the hospital and show them or how do you how do you sell it?

A. We have a Powerpoint and it's a beautiful PowerPoint let me tell you, it is a work of art, but you know, what I like to do is just talk to people because it's not always immediately obvious how it works or why it works. So we talk a little bit about the problems maybe five minutes and then we talked about how it works.

I know that like myself, every IT person is going say, okay Kyle, show me how it works? So we show them. There is no PHI ever in my system because I'm creating the synthetic transaction and so once we understand a baseline of how it works, then we go through and show them the application. I mean, it's not vaporware. Having been a customer for so long I'm very distrustful of vendors even though I am one now. I love it when I get to jump onto my demo survey. Let me show you this. I have an actual VNA running in the cloud that one of the one of the major vendors gave me because they were so excited about our product.

What I love to do is just go in and shut down the VNA and actually send text messages to everybody on the phone. It's like they look at this and go. Wow, it worked. Of course, it works. I wouldn't be here. The demo takes maybe 30 to 45 minutes, pretty quick, but it helps people understand the story. They can do it remotely. Once they understand what we're doing and why it works the most common response I get is, ‘wow, why doesn't this exist already?’

It really is for geeks by geeks. I was a president and imaging director. I know the problems. I built the tool that I wanted. And so now it's my joy to share that with my peer group. How can I help you manage your environment, serve your customers, make patient care better? I don't know if you saw it on our website, but our tagline is “Saving patients’ lives one server at the time.” I truly believe that. That’s what we're here for especially as an IT professional I don't see patients. I'm not a doctor, never will be, but my part is making sure that the systems are running so doctors can do what they do. I truly believe that that's why we're all here.

Q. How long is Imaging Heartbeat been on the market? And how's it going?

A. It is going really well. We had our beta release version 1 out at RSNA so we had a beta product and I was blown away by the reception. I was expecting to grow really slowly and get a couple customers but no. It was a case of people knocking down my door. We took what we learned at RNSA, took the feedback, did some demos at beta sites out there and got their feedback and then went back and said, all right, we need to be ready for prime time.

We completely revamped the application added a GUI user interface that really reflects who we are, simple, clean, elegant and that's how I like to talk about our product. It is elegant, it’s not complex. It is the antithesis of complex. It is straightforward and elegant and that's why it's bulletproof and easy to install.

We released the full GUI version at HiMSS this year, and now I just I don't have enough time to meet with everybody that's interested. We are we are startup and we’re bootstrapping there's no venture capital. There is no big company it is just us because we have just that kind of passion.

No comments:

Post a Comment