Sai Raya, PhD is founder and CEO of ScImage of Los Altos, CA.
Tell me about yourself and the company.
I’ve been in medical imaging for a long time. For 30 years, from my university days at Hospital of the University of Pennsylvania. I started a little company for 3D imaging workstations and that kind of stuff.
After that company, I started this company with a different mindset, with zero investment from any external people. I wanted to take one customer at a time and build a good ecosystem with customers. That’s what we have done.
What effect have publicly traded conglomerates and startups had on the healthcare IT market?
The fundamental problem with quarter-to-quarter financial reporting is that middle managers are forced to sell whatever they can, say whatever they can, and show the numbers. In the process, they have to go out and acquire new companies and change the solutions and whatnot.
Over time, they are working with one hospital. Maybe in about 10 years, there may be a couple of forklift updates. They acquire Company A and they have a solution for that. Then they acquire a new company, so they take the Company A solution out and put the second company’s solution in. In the process, hospitals are paying more. They don’t have the continuity in terms of what the hospital would like to do with data mining and all that stuff.
Small companies, on the other hand, can’t survive without proper financial backing. It is a competitive world.
From the lessons that I learned in my first venture, it’s very clear to me that the only way to build a solid company with a good financial foundation and bring that equilibrium is to have customer loyalty, and then continue the same solution over and over again. That’s why we have customers still with us since 1996. We never did the forklift update. The programs that we rolled in 2000 still work fine. That’s why there’s a kind of loyalty and a relationship between vendor and the hospitals and the physicians. That’s what we’re trying to enjoy.
It’s tough being an early adopter, like those pioneers who wanted to get rid of film and paper and move to PACS and EMRs before those systems were ready. Did hospitals jump on board too early?
Absolutely. Some people jumped on board without much thought. Somebody came and said, if you go to digital, you will save 50 cents per film or something like that. But first generation is first generation. They chose certain solutions.
Now we may be in the third generation. But in the grand scheme of things, digitization of the enterprise is just the first phase of what is going to happen to this healthcare IT in general. Whatever digitization that we’re trying to do these days, it is not dead yet. We’re maybe 70 percent of the way there.
This becomes kind of a building block for the future healthcare IT, where information and imaging have to co-exist. There cannot be any boundaries between these two things. A patient record is a patient record. It has to have everything that patient has ever done.
How do you see the market shaking out as imaging systems and EMRs try to figure out that co-existence?
If you went to something like RSNA in 2007, everybody was a PACS vendor. Everybody was changing film. But if you went to the latest RSNA, some companies went away and some got merged. A lot of consolidation is going on. In the process, certain hospitals learned something and some did not.
Images are growing. The image pointer that’s in the EMR seems to be the buzzword right now. That will go on for some time.
What are the most important workflows that an imaging system needs to address?
When we started this product we call PICOM, the fundamental point that I was trying to make was, if you go to any department — doesn’t matter, radiology or other — you see images and information. You have images and then lots of requisition sheets and observations and tech notes and physician notes and all kinds of things.
We wanted to create a platform that combines images and information together. Of course, we’re talking data in terms of components in the departments. We’re not talking EMR kind of systems.
For some of them, we had standards like DICOM and document exchange kind of things. Others we did not. We started acquiring them in their native format and put that solution together.
These days, if you go to any kind of imaging system, you have people talking about not just images, but information. Once this total data package is available, it needs to be seamlessly available to the front-end portal that the physician is going to interact with. There’s a lot of work to be done there. In my opinion, that is going to be the key for people that are involved with imaging.
How often do physicians who aren’t radiologists want to see the original image versus the interpretation from the radiologist?
We are coming from a multi-departmental type of company. We treat both radiology and cardiology together. In fact, we deal with radiology, cardiology, OB-GYN, and orthopedics all together. That’s one product for us.
Any patient that goes to the hospital many times, you get the ECG done more than you get chest film done. Images and waveforms are all together. In the case of ECG-type studies, as soon as the physician gets some kind of test result for the patient and before the physician wants to consult with the patient, they do want to take a look at these waveforms to tell them exactly what’s going on.
Modalities like ECG where the waveforms and interpretation are together – they are bound to open those kinds of things. Similarly these days, mainly for the orthopedic things, they generally use the images, but if it is a big CT study, I don’t think they’ll be using it.
What’s the status and challenges of sharing images across organizations?
The basic problem is there’s no universal patient identifier. We have our own way of doing it, but fundamentally it’s exchanging information from one PACS system to another system or one ordering system.
We created what is known as a universal MPI translator. That’s what we do. Right now, 20 percent of our business is interoperability, where we have to pull and push information from disparate systems and consolidate the reading and that kind of stuff.
That seems to be the name of the game for the next two years before somebody has to come up with a standard. If that somebody is the government, it’s not going to happen any time soon. [laughs]
What are the most pressing issues you are seeing from providers?
On one side, it’s the image life cycle management. It’s well defined and many companies have good solutions, including us.
But in the whole process, the diagnostic physician contributes the most complex and important content. The diagnostic physician’s impressions need to be distributed everywhere, wherever it’s needed. In fact, even for Meaningful Use, we have to take certain key measurements or key statements that need to be delivered to the EMR in a separate channel.
These are all the challenges. We have doubled up a good set of tools to do those things. Of course other people have also done that. But in the process, we’re still learning.
I see the importance of driving the subset of information from the diagnostic report and making that information co-exist with the image pointers or images and making them travel across the enterprise or make them travel outside the enterprise. That is the challenge.
Is there anything that’s being discussed that would allow images to be searched on qualities that weren’t noted by the interpretation, like the content of the image itself that might interest a researcher?
We have a lot of metadata in these images. If you want to search by image type or study type, it is possible. But the quality of the image, still it’s a visual perception, and a trained eye is the only one that seems to be doing a good job in terms of image quality audit.
But in terms of searchable images using, for example, something like “mitral valve prolapse,” that is easy to search and get information. It depends on the system. Some systems can do it.
In our system, we maintain an outcomes database and analytics and other things that we take very seriously. Every data object that comes into our system has the metadata latched on. It’s embedded right there in the image itself. It becomes easy to share that information or maybe make it available as an API for other systems to search.
With the rise of the vendor-neutral archive, what data types are people wanting to store that you wouldn’t have expected five years ago?
That’s funny. In 2000, we started an online “PACS in the cloud” type of environment called PICOM Online. Those days it was not cloud — it was an America Online-type of company name, so we called ourselves PICOM Online. [laughs]
My fundamental thing there was exactly this. It’s not just images. You’ve got to get all your documents, your spreadsheets, your PowerPoint presentations, and your business documents or billing statements — whatever is needed. They all get packed up into one object. It’s called study object. That study gets archived. The intelligence on the back end of the archiving system should handle based on how the client is interacting with it.
That’s exactly what we have done. After 14 years, more than 100 hospitals are using our online cloud solution. It’s a complete PACS, including reporting, voice recognition, and all kinds of crazy things. Some of the big companies these days are now finally opening their eyes and looking at the importance of delivering the documents with the images.
But 80 percent of the industry is still DICOM in, DICOM out, DICOM in, DICOM out. That’s all they talk about.
What’s the future of the industry or the company or both over the next few years?
Interoperability and making the image pointers universal. That’s one thing.
Security seems to be the biggest factor now, in terms of how securely we can encrypt this data and make it available to the right people at the right place at the right time and have the complete audit trails going with it? That is the key technology that we as an imaging provider needs to provide to the EMR companies.
No matter what, the biggest companies like Epic, when it comes to imaging intensified activity, it’s going to be with PACS vendors and image workflow vendors. We collect the data and then we have to make this data properly available to these people. That is a growing opportunity for us and I think it’s going to be there for a long time to come.
There’s going to be major consolidation and all that stuff, but still lots of hospitals don’t like this cookie cutter type of an approach. They would like to have customizable solutions that works for their hospital. That’s the opportunity smaller companies like us have.
Do you have any final thoughts?
We like what we are doing and we’re having fun. Being a private company with a good balance is a nice thing to do. We’re enjoying our little company.