Jonathan Hirsch was studying neuroscience at Stanford University when he wandered into two oncology classes and saw an opportunity to change the way health systems handle genomic data.
“I started getting really immersed in molecular oncology, and the challenges in implementing molecularly guided treatment started coming together with the challenges in utilizing complex data,” Hirsch said to The Cancer Letter.
Eight years later, Syapse, an informatics software program that integrates oncology data from electronic health records with genomic data, is being used in 79 hospitals and 800 clinics across 11 states.
“Our focus as a company is on care transformation, so we want to be at the point of care with the oncologists when they’re making that treatment decision for the patient,” Hirsch said. “We want to be there when they’re considering ordering a molecular test, selecting a targeted therapy, considering a clinical trial.
“We charge the health systems a software-as-service fee, specifically an annual subscription fee for our software platform and a per-user fee. We do not sell data, so our revenue comes from health systems paying for the value of our software.
“As far as our support for and work with Vice President [Joe Biden’s] Cancer Moonshot, I’m afraid you’ll have to stay tuned for June 29 to learn more.”
Hirsch spoke with Matthew Ong, a reporter with The Cancer Letter.
Matthew Ong: What is Syapse, and what unmet need or problem does it solve in oncology bioinformatics?
Jonathan Hirsch: These days, everyone is talking about precision medicine, but few have implemented it at scale due to the tremendous operational challenges in making these complex programs work. Syapse enables health systems to clinically implement oncology precision medicine through a software solution that pulls together all of the data, decision support, workflow, and quality improvement pieces that are truly needed for real-world clinical operations.
Specifically, oncology groups have a huge problem pulling together the relevant patient clinical, molecular, and treatment data to understand the full care journey of the patient. They have a problem understanding what the best practices are for treating a patient using precision medicine. They have a problem operationalizing complex precision medicine clinical workflows, such as molecular tumor boards, and finally they have an issue both measuring outcomes and rapidly learning from those outcomes at scale to improve best practices.
Each of those problems is massive, and in many ways are fundamental problems within the medical system today. Syapse is the software system that solves those challenges for those health systems, with an eye towards making precision medicine a reality for the largest number of cancer patients possible.
MO: How did Syapse come about? When did you start thinking about big data problems in health information technology and oncology?
JH: We started up in the earliest form back in 2008. At the time, I was in graduate school at Stanford studying neuroscience, and I had previously done a lot of work in molecular mechanisms of neurological diseases, specifically diagnostic and treatment paradigms based on those molecular mechanisms.
I didn’t have a big plan at the time to start a precision medicine company, but I had worked in academic medical centers, I’d worked in clinical research, I’d worked in biopharma, so I had personally seen that physicians and other members of the healthcare ecosystem lacked proper software tools to make proper use of complex data and information. Taking one look at the new EMRs being put into medical centers at the time more than confirmed this!
The issues of molecularly guided care were on my mind when I, frankly, wandered into two oncology courses focusing on molecular diagnostics and the molecular basis of oncology treatment. I started getting really immersed in molecular oncology, and the challenges in implementing molecularly guided treatment started coming together with the challenges in utilizing complex data. I thought, “How are oncologists and care teams, particularly outside of the academic centers, going to move into this era of molecularly-driven treatments if they don’t have the necessary tools, software, services, etc., to deal with this complex data, help guide treatment decisions, and learn from real-world outcomes at scale?”
During one of my Stanford classes, I said, “I think that there is a company to be built here,” and I dropped out of grad school and started Syapse. This was in the depths of the recession of 2008, so for the first three years of the company it was basically just me full time, with two co-founders working part-time, and a bunch of grad students we hired on as interns building initial prototypes and testing it out in the market.
We got going as a “real” company a few years later, around the end of 2010 or early 2011. That’s when we hired a small team, moved out of the dining room and garage, and into a real office. We had our first production software and early adopters in mid-2011 through mid-2012. The first two health systems we started working with were Stanford and Intermountain back in late-2012 and 2013.
MO: How exactly does Syapse solve these precision medicine challenges in its design?
JH: Our software focuses on four key areas: data integration, decision support, clinical workflow, and quality improvement. The first component that we have is we build out is our data integration platform that is very good at going into the various different electronic systems within the medical center, such as the electronic medical record, imaging systems, pathology systems, etc., and extracting all of the relevant cancer information about those patients, pulling it into a centralized database, and structuring and normalizing that data so you have the full patient record.
This is an extraordinarily complex process that many claim to do, but few do successfully. We are able to be successful in solving this data integration problem because of the semantic computing platform we have built, which normalizes messy data into best-practice oncology knowledge models that we have spent years building, and the packaged data integration pipes we have built to hook into the databases of EMRs and other standard clinical system.
Additionally, we bring together the clinical data with molecular (genetic, genomic, etc.) test results from in-house and send-out molecular labs. We have developed an interoperability framework for molecular testing, called the Syapse Lab Certification Program, which enforces a best practice data schema for the exchange of structured molecular test results.
The second area of focus is clinical decision support functionality. Having integrated the full, longitudinal patient data, our software can fire decision support rules at different points in the patient’s care journey, suggesting molecular tests to order, matching the patients to drugs and clinical trials, and more. Unlike others, we do not pursue a “black box” approach to our CDS functionality, meaning Syapse is not developing algorithms and new clinical protocols behind the scenes and forcing doctors to use them.
Instead, we allow each health system to fully control the clinical best practices that are embedded in our software and used for CDS. Whether those are the best practices developed by the institution themselves, standards in the public domain or developed by groups such as ASCO, created by a third party vendor, or even shared by one Syapse-partner health system with another.
The third focus area is the clinical workflow and care coordination software framework. It is critical to streamline complex processes such as molecular test ordering, molecular tumor boards, specialty drug procurement, and clinical trials eligibility assessment so that all of the key clinical stakeholders can focus on patient care rather than fill out forms and manage logistics.
An example of this workflow optimization is using the data we’ve integrated on the patient to automatically fill out a clinical trial eligibility assessment form. We focus a lot of providing workflow tools for care coordination, such as a dashboard that a nurse navigator can use to monitor the status of patients in a precision medicine clinical program, and solve workflow bottlenecks such as drug procurement. Additionally, our workflow software integrates with the EMR, so that a physician has a seamless experience moving from their current clinical software environment to ours.
The fourth area of focus is our quality improvement and learning health system framework. Our software tracks patient outcomes, both through direct documentation in our software and well as through data we integrate from imaging, drug administration, and other system. What we do that is a bit special is we can link the outcomes directly to the full care journey and the decisions made in our software, such as changing a patient from chemo to a genomically targeted agent as a result of a molecular test.
Our software enables this outcomes tracking at scale within a health system, so the end result is that each health system is building a massive real-world evidence database in our software, linking clinical history, molecular and genomic data, treatment decisions and implementation, and outcomes. Our software then contains tools to enable physicians at point of care, expert review groups such as molecular tumor boards, and health system administrators to all use this information to advance quality initiatives.
For example, an oncologist at point-of-care can say, “I’m seeing a patient with a rare signature of tumor type and molecular aberrations, what do I do for this patient sitting in front of me?” Our software can help that physician contextualize that rare patient case into a larger population and use the real-world population information to determine the appropriate treatment course. We call that feature “Similar Patients.” And, we can enable the physician and the administrators at the health system to understand the broader population dynamics and trends, and derive practices from their real world data that then get codified in our CDS functionality.
Enabling precision medicine as a real-world clinical program in a community setting is quite complex, and really does require a focus on these four pillars of data integration, decision support, clinical workflow, and quality improvement.
MO: Who is currently using Syapse and what is your business model?
JH: We work with large health systems and physician networks. Those might be large integrated delivery systems such as Intermountain Healthcare, community health systems such as Providence Health & Services, or academic medical centers such as Stanford Cancer Institute. Those are some examples of organizations we work with.
Our business model is that we charge the health systems a software-as-service fee, specifically an annual subscription fee for our software platform and a per-user fee. We do not sell data, so our revenue comes from health systems paying for the value of our software.
We tightly align ourselves with the health systems—the health systems are both our customers as well as our long-term strategic partners. We’re really trying to enable the health systems to improve the quality of care that they’re delivering, and use precision medicine as the lynchpin for a broader transformation into a value-based care framework.
MO: You’re saying that this is a solution that’s available to both academic centers as well as community health systems?
JH: Absolutely. I think the power of the solution increases when you have a broader network that you need to serve. For an academic center, for example, you may have a physician who’s an expert in non-small cell lung carcinoma, and that’s all they see and all they treat, but that academic center may have a network of community affiliates who aren’t experts.
Traditionally, it has been very difficult to coordinate the sharing of knowledge of best practices between the academic center and their community affiliates. One of the things our software does for an academic center is help them disseminate those best practices out from their experts to the community affiliates, and then receive back information about the care journey of those patients, compliance with those best practices, and outcomes.
When it’s time to have the patient maybe referred to the academic center or to have the patient matched to a clinical trial at the academic center, our software can help automate that process rather than what occurs today, which is essentially the phone calls back and forth between different organizations and emails and disorganized mess.
To give you a concrete example, an academic center may run a molecular tumor board that provides treatment guidance for patients seen by their community affiliates. Our software supports the MTB referral workflow, the data aggregation and case presentation, recording and disseminating the treatment guidance back to the community affiliate, and tracking adherence to the guidance and outcomes.
MO: How are you different from other data software? What is it that you offer that is unprecedented compared to what else is out there?
JH: There are a few things that are really unique about us. The first is the focus on clinical care transformation. There are many others playing in the precision medicine and oncology data space who are focused on the research side, which is certainly a worth place to focus.
But our focus as a company is on care transformation, so we want to be at the point of care with the oncologists when they’re making that treatment decision for the patient. We want to be there when they’re considering ordering a molecular test, selecting a targeted therapy, considering a clinical trial.
We’ve built our company and product around this mission. That means our product needs to be comprehensive, to satisfy the needs of our clinical users and health system customers. For example, providing molecular data in isolation isn’t helpful; you have to provide the molecular data in the context of the clinical information. Providing decision support in isolation isn’t helpful if you don’t connect the drug recommendation to the subsequent clinical action to help start a procurement process for that drug. And doing all that without tracking outcomes isn’t useful either, because you don’t know what’s working and what’s not working, and you can’t rapidly bake that knowledge into updated best practices.
Our focus on point of care decision-making and clinical transformation, and the comprehensiveness of product we’ve built to support that, is the primary thing that makes us unique as a company.
The second thing is the fact that we are aligned directly with the health systems and the physician groups, rather then aligned with the pharmaceutical companies, for example. We’re not selling data. We’re not monetizing data. We’re really aligned with the health systems and patients in their pursuit of better care and better treatment decisions.
The third thing is our demonstrated success in implementing precision medicine across different environments: academic medical centers and community practices; in multiple health systems across the country—West Coast, Pacific Northwest, Midwest, East Coast; and in the context of a variety of different IT ecosystems and challenges such as different EMRs, no EMRs, lots of documentation, lack of documentation, in-house versus send-out molecular labs, and more.
MO: So Syapse can be customized to meet specific needs of each health care system?
JH: We have a set of best practices—not medical best practices—but workflow best practices and data best practices. We come to a health system with that robust set of operational best practices based on all of the successful precision medicine implementations we’ve done with health systems. Each new health system benefits greatly from that shared set of operational best practices.
But at the same time we do recognize that each health system has unique aspects to them, and in particular, that all health systems desire to be in control of the clinical best practices that they and their care teams are using. We can configure our software to use the medical best practices of any individual health system that we work with, or we can help them share best practices with each other.
MO: What is the Oncology Precision Network and how many members and patients does the consortium currently have?
JH: The Oncology Precision Network (OPeN) is an effort to rapidly improve patient care by making it easy for health systems to share de-identified cancer patient data and the knowledge gained from this real-world evidence. Syapse has partnered with some of our key customers to launch this initiative: Intermountain Healthcare, Providence Health & Services, and Stanford Cancer Institute are the founding members of OPeN.
OPeN unlocks that data from the siloes of the different health systems so we can as an industry more rapidly advance our state of knowledge and understanding in cancer precision medicine and make more rapid progress in figuring out what the right treatments are for particular patients. The network gets together these health systems and helps them share this information in a regulatory-compliant fashion that respects patient privacy. Data being shared includes demographics, clinical history, tumor genomics, and treatments. OPeN members agree not to monetize the aggregated data asset, but rather to use it for clinical care purposes. In those aspects, OPeN is unique among the major cancer data sharing initiatives out there.
Amongst the three founding health systems, Intermountain, Providence, and Stanford, OPeN covers, 11 states, 79 hospitals, and when we’re at full implementation we’ll be at about 50,000 new cancer cases per year and 200,000 active patients under management. It’s a big network, but it’s just a starting point where we are at right now.
OPeN has deliberately chosen a controlled launch strategy. The three founding health systems and Syapse have been working very closely together to build and successfully launch the network. That has resulted in a rapid launch of the software framework with a large number of patients and integrated data already in the system. It is very exciting to see it live!
OPeN will start onboarding additional health systems later this year, and our goal is full geographic coverage of the U.S.
MO: Do participating health systems use a standard protocol for consenting and registering patients in a HIPAA, PHI-compliant way? What’s the process?
JH: There will not be a standard research protocol that has to be followed. The patients whose data are in OPeN represent real-world cancer patients, not a rigidly formalized research study.
There is plenty of room for different members of OPeN to have different approaches. For example, Swedish Cancer Institute at Providence has a Personalized Medicine Research Program that has a specific protocol and study calendar, but the other institutions have their own approaches and protocols. The point of the network is to capture as much real-world experience as possible and to rapidly learn from this data.
Of course, there has to be consideration for HIPAA, HITECH, and the other regulations. The three founding health systems and Syapse have come together to create a standard data use and data sharing agreement that all OPeN members will sign on to, and that agreement goes into the details of what information can and cannot be shared, the HIPAA and HITECH protections, and much more. Constructing the OPeN contract may have been as challenging as implementing the data aggregation, normalization, and sharing software!
MO: Is the consortium novel or are there efforts that are similar?
JH: The consortium is novel from the standpoint that we are capturing and tracking the real-world experience of cancer patients across many different health systems and EMRs, including not just clinical data but also genomics and treatments. OPeN is doing so without prescribing that a specific research protocol be used.
We are taking data from each of the organizations participating in OPeN and we are normalizing and mapping all of the data so that the end result is a standardized dataset but not from a standardized research protocol. The reason why that’s really important is we want to actually see the variance in workflow and the variance in treatment decisions across the different organizations—across the academic centers, the non-academic research environments, and the community environment—represented in the system and the very different geographies and patient populations in the system.
The point of the network is to learn, and to do so quickly. It’s to learn what works better than not, what situation is better than others, are there treatments that work well in academia and clinical trial settings that don’t work well in community practice for certain reasons. OPeN is really intended to, first of all, support point of care decision-making and second, to learn from the real-world treatment experience rather than run another clinical trial.
That’s what differentiates OPeN versus all the other efforts that are out there.
MO: What role do you foresee science playing in Vice President Biden’s Moonshot? We all know he has selected and endorsed NCI’s Genomic Data Commons. What does Syapse have to offer in furthering the Moonshot’s data-sharing goals?
JH: As you can tell from the description of Syapse and OPeN, Syapse and OPeN are very much spiritually aligned with what the Vice President is trying to do with the Cancer Moonshot. We are very much in favor of unlocking data from siloes, and enabling the use of real-world cancer patient experiences and data in informing clinical decision-making for today’s patients
As far as our support for and work with the Vice President’s Cancer Moonshot, I’m afraid you’ll have to stay tuned for June 29 to learn more.
MO: What’s up next for Syapse? What are your projections for Syapse in oncology 5 or 10 years from now?
JH: Certainly, one of our major goals is to democratize access to precision cancer care to as great an extent as possible. What you’ll see Syapse doing over the next 5 years is growing our relationships with large community health systems, whether those are integrated networks or large hospital and physician groups, and driving the expansion and utilization of precision cancer care throughout the community.
We’re trying to move precision cancer care up earlier and earlier in the care journey of the patient and figure out when it makes sense for an earlier stage patient to receive a precision medicine-guided approach to cancer care.
The next thing that you’ll see Syapse doing is increasingly merging the precision medicine approach with a value-based or at-risk payment paradigm. We are big believers in the fact that precision medicine is not only clinically effective, but also a cost effective mechanism for treating cancer patients when you set up the proper financial relationships and all parties involved are at risk.
You’re going to see Syapse doing more and more work to enable health systems to go at-risk for larger portions for their cancer population, and to establish relationships with payers—whether it’s an owned payer or a third-party payer—to make that happen.