June 10, 2019
In a recent podcast, Matthew Michela, CEO and president of Life Image, a global network for sharing clinical and imaging data powered by industry leading interoperability standards, discusses the barriers of innovating interoperability and the value of giving patients control over their own health data.
I'm Matthew Michela. I'm the president and CEO of Life Image.
My background, over the last three decades I've worked and I'm experienced in the payer world, focusing on insurance and risk management, reimbursement, and financing. In the provider world, really supporting the delivery of healthcare services at the physician practice level.
In technology, focused on automating solutions and accelerating the application of healthcare decisions by physicians and patients in chronic care and disease management, organizing healthcare services for very ill people and populations.
And then in broader population health, trying to provide systematic public policy and population‑based solutions in healthcare that really affect healthcare quality.
Now, I'm leading Life Image, which is working hard to solve the problems with accessing healthcare data, and the challenges which folks in the industry would reference as interoperability.
What patients would normally see and hear as breaking down these healthcare data silos that patients can receive the best care and the healthcare system itself can reduce waste in the medical spend.
At Life Image, we're a company that has spent really more than the last decade focused on improving interoperability. We began with the field of medical imaging itself, which is a chronically difficult area to access information other than on CDs and what used to be, unfortunately, actual film itself.
Now today we really orchestrate the flow of all types of clinical information to help care teams across really the broad healthcare ecosystem to make better clinical decisions, from the point of view of providing better care for the patients themselves.
At Life Image, we've got the largest global network for clinical information sharing. Today, it's about 10,000 hospitals and medical facilities that use Life Image to move medical information around in the United States, and another roughly 60,000 hospitals, clinics, and facilities around the globe.
We're today providing service to on average between 10 and 12 million unique patients a month throughout Life Image, with a wide variety of services we're providing them for their clinical needs.
We've got a very, very large network, as I had mentioned, just within the US itself of top healthcare companies, systems, and hospitals here. We have roughly 80 plus percent market penetration into academic medical centers and tertiary care facilities, which is where most of the most complex medical decisions and clinical occur.
But within our healthcare platform itself at Life Image, we provide data for a wide variety of clinical uses and clinical use cases.
For example, we provide information to telehealth companies. It might be a telehealth company that will receive inbound telephones for patients looking for second opinions or referrals in this new industry of telehealth that need medical information in order to be able to help a patient. We'll be able to access and provide that.
It might be telehealth in the sense that we move information around to connect radiologists to read imaging exams for providers and facilities in remote facilities, Montana or Idaho or Kansas, where they might not have the resources to staff full fledged radiology teams 24 hours a day, seven days a week, where they need someone else that's remotely staff to be able to get that information or send it in that regard.
We support physicians for receiving medical information of a wide variety of types inside and outside the hospital. Roughly 20,000 physicians use our platform in the US every single day.
We've got a wide variety of customers. I had mentioned, more than 10,000 facilities in the US themselves, which is moving data around for a wide variety of care cases.
Whether it is surgical cases or it's stroke, where someone might have imaging done at a community center and have to be transported by ambulance or helicopter to a level one trauma center, that want to be able to understand that healthcare data that was captured in the local community.
And can be transferred from the helicopter, from the ambulance, from the community center to the level one trauma center, so when the patient gets there they already have assessed the patient, determined what the appropriate treatment is, and they don't lose time doing that, etc.
We work with consumers. We've got a consumer application we're quite proud of, which we call Mammosphere, which in essence really breaks the silos down completely by putting the patient in the center of the world, and uses the Life Image infrastructure technology platform in order to acquire their medical records and their imaging, irrespective of where it was received ‑‑ California, Nevada, Kansas, Tokyo.
And collect their full medical records so that they own it, they control it, and then they can use it, transport it, provide it to doctors for referral, for diagnosis, treatment, etc. And avoid this conversation of running around and trying to collect partial records.
We work with medical device companies, helping them design, custom design medical devices for surgical implantation into patients. So knees, hips, etc. that you don't just plug in a standard device.
You have to understand the specific physicality of an individual patient, and that requires imaging and medical assessment of the patient, which then we can facilitate and help the device companies design that.
We work with payers and vendors to vendors to payers that need information to decide, should I pay for this claim or not? Should I review this case concurrently or retrospectively to determine whether it was medically appropriate or not? They need medical information and records to do that, and we facilitate that.
We're working extensively with this expanding industry around augmented intelligence and companies that are using machine learning in order to create algorithms that drive productivity, drive efficiency, low costs, and better precision medicine in virtually any kind of therapeutic area, from oncology to neurology to orthopedics, etc. that need lots of information order to work.
But also need the ability for that technology and algorithms to get into the practice level computer systems so that patients can use that and have that technology available to them. We're working with life sciences companies, and trial development for research and drugs and devices, etc, where they need large sets of data in order to run those clinical trials.
Across our ecosystem at Life Image, we really use our broad network connections and this ecosystem of constituents for whatever type and kind of medical information you need, so that you can make a decision for a patient.
You can make a decision to help improve a drug, to create a device, to decide whether payment is appropriate, etc. We help facilitate access to that information inside the healthcare system itself, because healthcare data is very, very challenging to get for lots of reasons that I can touch on in a minute. We really help solve those really complex technical problems here.
When you can't access medical information, skipping the question for a moment of why it's so hard, what happens to patients, ultimately if they can't get that?
Well, the way I would explain this is, I'll create a comparison perhaps that's less technical. But think of healthcare data as ultimately the blood that flows through our own systems today.
If you don't have blood to deliver oxygen, then you're certainly not going to live very long. In healthcare you need that data in order to actually make medical decisions that impact where a patient goes. It's about having really the right information and the correct information in the right time. That determines, ultimately, what is the quality of care that a patient ultimately receives.
Think about a mammogram as an example. We have this experience all the time. Whereas our patient may have this experience all the time, where they take a mammogram and perhaps that mammogram may not be entirely optimal and entirely pure.
Now that might be because the machine isn't exactly right, or it hasn't been set up correctly. The technician isn't there, the software isn't there, or even a woman's breast tissue might be slightly denser than normal. So the imaging itself might be cloudy or not as precise as it needs to be. So, the data isn't great here.
Well, if that occurs and it moves on for diagnosis, you could miss a mass. You could miss a growth. You could miss something along the line. The consequence of that for patients can be lots of unnecessary testing to try to straighten it out, or God forbid something gets missed and actually develops into cancer, and is a life and death situation.
Having quality of data at the right time matches up to determine what a patient outcome ultimately would be.
A different kind of example, in layman's terms, might be blood pressure. If I go to the doctor, and in my medical record there's a history of high blood pressure, and I might not have that because it's a mistake, and it's a bad data in the record or file.
When I go in they might prescribe me statins or other blood pressure medications, as an example. That actually could have a bad outcome and can actually hurt me, because they have side effects that weren't required here.
So, having the right data that's available to that clinician at the time that they're diagnosing the treatment, and it's at the time of diagnosing the treatment that it actually makes a difference here.
Getting a mammogram four weeks after a radiologist reads it and decides that there's going to be a mass there, that time difference of a month, if they miss a growth, could be very someone with triple negative breast cancer the absolute difference between life and death, since that disease progresses so rapidly.
You can't spend a month. That's literally the time when you have life and death. So, getting those exams at the right place in the right quality is what determines it.
If you don't have that, then ultimately patients are going to have suboptimal outcomes. There's going to be more cost in the system, and that's not obviously what we want.
At Life Image, our business is trying to get that right information with the right quality data into the hands of the people who need it to make medical decisions across this broad constituency.
As a technology company, that is how we help solve this problem here. Because technology is fundamentally really the only answer to this problem in healthcare.
Before there was technology, and we had paper records, and they were sitting in paper files, and they're influenced by the human data collection process, which means they're incredibly error‑ridden, they're really frankly inaccessible.
Without technology, we turn the dial back 100 years, and customer quality would be dramatically affected. Because, again, paper is just fundamentally a mess.
But the application of healthcare technology also creates substantial other barriers to accessing data, because just because you take a picture of it and put it in a database doesn't mean that you can get it where it needs to go at the right time and the right way and right quality that people can understand and accept that.
It can eliminate the physical transportation time burdens here. But in healthcare, data resides and is accessed in many, many different kinds of technical healthcare standards.
Without getting too technical around that, if you think about a word processing system from 10 years ago that files you created in a Microsoft Outlook system of 10 years ago is not going to be brought up and accessible today because they use different technical standards to create that document.
Now that gets organized. Today, outside of healthcare, much of data is ultimately standardized, and you can share it, talk with systems, and computers can talk with each other.
I can email a document that's created on a Microsoft platform and read it on my Apple iPhone. That's been standardized. But in healthcare, we have so many different standards for so many different data types, it just creates massive complexity here.
Then we have the science itself that changes around healthcare. We have an implicit bias in healthcare that says we want to improve healthcare quality all the time. So, we're going to focus on that science.
If I think about an imaging example, my focus if I'm creating an MRI machine to take pictures is to create better science so that imaging gets to be more precise, more accurate, better digitization for it, better ability to understand tumors and other things in the imaging.
The imaging we have today in this current generation is just light years ahead of imaging that existed even 10 years ago, let alone 20 years ago.
So, the science of creating magical, in almost regards, better science for diagnosis and treatment is fantastic. But if you're in the business of creating that better science and improvement in quality that has to be clinically tested, then that means sometimes you're creating new healthcare standards and formats for the data, because it never existed before.
In healthcare, we're creating all of these new standards in many regards because the science is driving that. It's a secondary use to say, I'm not going to put this wonderful healthcare technology in the market for five more years, because we're going to focus on the technical issue of standardization.
So, we have an underlying arms race in healthcare improvement that creates technical barriers. While we say technology solves all problems, let's standardize it, that runs counter in some regards to the improvement we desire in science, in some regards. That's just really one of the those issues here.
It also takes a long time to create healthcare technology, and then to validate it. Sometimes software has to be approved by the FDA. Sometimes trials have to be conducted around it. That extensive process adds a lot of cost and complexity, and then creates an inflexible system that's even more costly to change and modify and validate.
Then you have a really high cost of implementation of healthcare technology, and to train healthcare technology for people who actually use it themselves in the practice area.
Since healthcare data is generated from so many different kinds of healthcare sources...a type of data comes out of a blood pressure machine measuring your blood pressure. A different kind comes from your pharmacy as to which pharmaceutical you had be prescribed.
A different kind comes out of your imaging for head CTs. A different kind comes for your imaging for breasts. A different kind comes out for claims, healthcare data and what was paid for.
Since those systems, solutions, companies, and technology are created for very specific uses, and in some cases in non‑standard ways, or in some cases with technology that takes years of investment to fix, creates a very, very slow technical environment here to create change.
While I can create an application that's not a healthcare application, put it together, put it on to an iPhone exchange if it works to provide one use case, then change that 72 times in any given year. In healthcare that cycle time might be, maybe I can change it once in a year. Maybe I only change it once in three years.
That slowness of modification, change, and adaptation also creates technical challenges in healthcare to get information from A to B.
Again, that comes back to this inherit theme that then that makes it more challenging to get the correct data in the correct time frame to the correct people, so that patient outcomes and costs can be managed here, and optimized, across the board.
In Life Image itself and how we think about healthcare data, and one of the use cases that we like to talk about, is our Mammosphere solution and product, and putting women's healthcare information and breast health information and imaging itself into their own hands is, how can we help provide better access to that, and ultimately what does that mean for a patient's experience along their healthcare journey.
What we know from our work in imaging itself and in medical information and also working with large populations of women that are concerned about their breast health is, as you think about mammography and you think about breast imaging, breast imaging is actually very, very unique.
The way I try to explain that to folks is, I can take a picture, an image of my knee, and we could, with that image determine if I have physical abnormalities, if I have some tendon damage, if I've got ligament damage, if there's bone deterioration. I can take that from an image. I can look at it, and I understand what's going on.
But breast tissue sits outside of a woman's body, and is frankly unique to her. You can't make a comparison between one woman's breast and another woman's breast, and say I understand whether one has a problem or not. You have to have prior exams.
Protocols for mammography determine that at a certain age you get a baseline exam. From then on, you make comparisons. It's that delta and that change that a radiologist, a breast physician will look for to say, "Wait a minute here. Something has happened. I now see a growth that wasn't here before that might be abnormal. Let's look at it, etc."
Having those prior exams is an absolute requirement here. The problem, of course, is getting that information from where it might be held or born, in a timely way at the right quality, into the hands of physician when you come in and order as a woman to get a subsequent exam and just determine, do I have a problem or not? Do I ultimately need to be treated?
Life Image helps do that, by making sure that data, that mammogram and maybe five years of prior mammograms is available at the time I do my next screening. It's the difference between, in many cases, whether you have a good outcome or not, whether you have to have additional testing that stretches you out in your life and your experience.
Think about, and this happens commonly, of a woman that will go in for a mammogram and not know that those priors should have been available. They're not available, and because they're not she gets an inconclusive result that maybe there's a problem. We're not sure. You need to come back for further testing.
That's a terrible human stressful experience of, "Oh my God. My physician can't tell me that I'm clean and I'm clear. Wants me to come back for testing so we can sure."
That might be another MRI. It might be a CT exam. It might be an MRI that's really, really expensive. It might even be surgical biopsy for women who have dense breast tissue or need a lumpectomy.
All of that is an incredibly stressful journey for a patient that may take weeks or a month or six weeks leading up to surgery, that we know, from lots of clinical research and work that we do here, that there's a high degree of occurrence where all that additional testing past that first mammogram was actually a false positive and unnecessary.
Everyday in America, I'll focus just on the United States, there are women that because they didn't have the prior data, the right data available at the time that's needed for diagnosis, that they undergo unnecessary treatments here, which is costly and stressful to the patient and terrible experience.
Closing that gap is ultimately really, really important here for patients.
Now, back to that earlier comment, which is, why is medical information so hard here? Why does the healthcare industry continue really to struggle making data like mammograms here interoperable and available to patients all the time?
I mentioned one of the challenges was the science itself. We're changing the science constantly to adapt to new technologies. That affects the ability to standardized.
In mammography itself, several years ago, we created the technical ability to use tomography, which is a very highly, experts will tell you, improved way of looking at breast tissue.
The focus was, that kind of imaging is incredibly more dense, incredibly more complicated to move and to access and to move around. Frankly, lots of healthcare systems couldn't accept that. They couldn't bring it. They couldn't upload it. They couldn't access it in their database.
It's taken several years to modify that. But the science of creating it was frankly, in some regards, years ahead of the ability to take that type of imaging and move it through.
A second kind of cause here is really around just behavior. Of people who participate in healthcare system, it's really, really slow to adapt. It's very hard to change practice patterns and approaches and physician workflow.
Think of it this way. A physician's time is incredibly valuable. They can't see enough patients in a day. We create, in our healthcare system, teams that wrap around physicians.
Whether it's a physician office or whether it's an hospital or whether it's just a surgeon, there's teams of support people that are there to prepare the patient, get the paperwork, do the consent, get the test data, get the equipment ready, get the surgical suite set up so that the physician's time can ultimately maximize here.
When you create that kind of workflow that affects three people, 10 people, it takes a lot of energy to create it.
All the sudden now, you create a new technology, a new data type, and say, well, I can't use this computer system anymore. I want to use this one. Or we're swapping out this UI that you experience on your Android phone with this one on your iPhone.
That is time that then has to be adapted for in the workflow. People have to be retrained. All of that has to happen. For physicians and healthcare providers who's daily life is to treat as many patients as possible, putting all that burden for change occurs, but it's costly and it's slow.
Lots of physicians and their support teams just don't adapt that quickly within healthcare. Again, they're orientated to deliver as much care as possible, not necessarily orientated to, hey, I want to adopt new technology, as the best thing we need to do.
The technology change itself is also a major factor here, because that's really slow and it's really costly. As I had described a little bit earlier, we got a large number of vendors inside the healthcare domain.
Whether you're large, large companies like General Electric or your smaller companies that are more nimble in a technology point of view, any given hospital, if we talk about a hospital, may have hundreds of technology providers.
It is not uncommon when I say hundreds...I'm not even talking 100, I'm talking hundreds of technology providers.
To connect them all and integrate them all and move data between them and try to create a seamlessness is just this massive technology challenge and barrier here, again, to interoperability.
But I will tell you, the inverse of that is also potentially problematic. Because when you come in and you say, "We're going to wipe out all of these 100s, and we're going to replace them in many regards with one system...We're going to create one EHR, and that system is going to layer across the entire hospital and be the one way we do it."
Then, in many regards, what we found and find in the industry has been that then one big system becomes really, really slow and burdensome to change. Because it says you have to do it in one single way, then new science can't adapted as early enough, and changes can't adapted as early enough.
It's this really interesting balance in healthcare around this barrier around healthcare data, and how do you solve that to get into the hands of patients and providers when they need it in the quality they need at that time that they need it.
You want lots of small companies that innovate to move fast. If too many, it's a problem. If they're too disparate it's a problem.
On the other hand, if you have one big institution or two or three big vendors, then it's really bureaucratic, and it's a whole different process. So then, it's slower to change in a way that can adapt. That's what we deal with in healthcare here across the board.
Then we have really the last thing of why are we struggling getting this done as an industry? It's just the practice approach to medicine itself.
People talk about medicine being a science, and it's an art here too. What we have in the healthcare system, if I focus just on physicians, as one consistency, they have a lot of varied experience. They have a lot of varied skills. They have a lot of varied capabilities.
There are some physicians that are really fantastic and great and cutting edge and really high performing specialties that have insight that no one else on the planet have.
Then there are physicians who aren't. I don't make a value judgment. There's a wide population of people that have different skills and experiences and capabilities, based on their exposure to the treatments they're talking about, on their training, just on their individual philosophies and approaches.
What that ultimately means, since our system is set up that allows an individual physician and an individual patient to ultimately make a decision of treatment, a diagnosis and treatment.
Through that discovery process, we have across our healthcare domain many, many, many pardon my terminology here, but endpoints of individual physicians with varying practice patterns and experience, with patients with various disease conditions and circumstances, and societal issues making what that very best decision is for that physician and that patient.
With hundreds of thousands and millions of those occurring everyday, it's very hard to say, how do I standardized healthcare delivery? Even if the science is built, even if the FDA has approved large clinical trials...
And everyday we're learning new things about what today we thought was good and tomorrow isn't so good to do anymore. Better research shows it's good. Imagine that in an environment. We have millions and millions and millions of individual decisions in order to make that happen.
With that, I just ultimately would want summarize here is that big picture here among this complexity and the one thing that would ultimately drive it, it's having data in the hands of patients and the patients in the center of the healthcare system.
It's this increased drive of consumerism, which requires consumers to be knowledgeable, that requires the government to continue to influence providers to cooperate and change their practice patterns to deliver to patients.
That is the major factor that solves all these problems overall in the long run, is patient demand in healthcare when they have the data to evaluate the service they received, the cost they received, and the ability to own their own data so it's transportable. That's what's going to create the greatest change in the system over time.
Thank you for your time and appreciate the opportunity to share with you.