Darren Dworkin is the senior vice president of enterprise information services and CIO of Cedars-Sinai in Los Angeles and oversees the health system's data analytics and management efforts.
Here, he discusses the biggest challenges for understanding data analytics at his system and how Cedars-Sinai is building an internal team to support their efforts.
Question: What are the top data management challenges you hope to solve in the next 12 to 24 months?
Darren Dworkin: The way we talk about our pursuit of data is very much a journey, and I would describe the last 10 years or so as us largely being on a quest to ensure we can collect as much data as we could while also developing platforms that do more than collect data. We have been producing a ton of data, and going forward we have very specific and broad ambitions to release and make that data available to our patients, either through direct applications or connecting to external applications so patients can choose how they access their data.
Internally, the academic medical center is built on the basis of discovery and curiosity, and we want to release our data and make the information easy for investigators to use. Historically, our operational teams have hired and built teams with the hunter-gatherer mentality to the data; there are a few mature teams that were looking to normalize their data. We weren't spending much time analyzing, thinking and hypothesizing about the data.
However, we had a large initiative to help support out clinical, operational and research department to enable them to get their data more simply and make it as pre-normalized and pre-translated as possible. Then the experts that understand their area can spend their time analyzing their information.
Q: What is your organizational approach to data registries and how do you make sure you're getting the right data input?
DD: We have a team that is dedicated to and focused on data quality and governance. These people ensure the right data is getting to the right place; they make sure the data is correct and translated for the end user. One observation we made some time ago was that our CEO became frustrated if he received two different reports on the same issue, which led to two different conclusions. For example, length of stay can have multiple definitions and include both clinical and financial perspectives.
We also need to provide the data to the folks who are working with it every day. We have a centralized team that manages, builds and deploys reports for our departments so we can be effective. Then, experts within those areas create dashboards or reports using the data so that team members know the important information and oversee improvements.
Q: As healthcare organizations partner with more vendors and collect an increasing amount of patient data, how do you ensure that data is secure?
DD: To date, we have largely worked closely with a select group of companies to provide necessary enabling tools for us to perform the work. The partnerships do not involve us giving information to a third party and having them analyze it and then providing us with recommendations. Our partnerships are more focused on bringing in tools we need. For example, we partner with Tableau to use the toolsets it provides to build dashboards and present visualization of the information for both internal and external users in a way that meets their needs. We hold large internal training classes for individuals throughout the system to learn how to use Tableau and its functions. We teach them how to use the information, and how the information is categorized and defined so they can hit the ground running.
We do not outsource the data science responsibilities, so we haven't encountered challenges that some others may be facing.
Q: I'm sure you receive requests daily for partnerships or collaborations with technology companies or vendors. How can CIOs handle that situation?
DD: As a general comment, we view our need to be very cautious in terms of how we are working with patient information and ultimately who we are extending the partnership of that information to. It's very rare that we cross into a relationship where we share that information, and when we do, we have to understand the reasons for it. We consider data a trusted resource that patients share with us, and we take making this information available to patients very seriously. We try to be aggressive and creative in sharing information back with the patients. It's their information and they should have the choice of what to do with it. We have been very conservative and purposeful around sharing any information with a third party.
Q: What do you think are the most interesting opportunities for partnering with big tech companies currently or in the future?
DD: I don't think we know yet. We are cautiously exploring relationships and the benefits they would bring. There are a couple of tricky things around that, and an increasingly gray area of navigating the rights to data. There are companies that have the goal of monetizing data and we don't find that compatible with our goals. As we explore a new horizon ahead of us, we need to dig deep and understand what the motivations and goals are for our partners to make sure they are compatible.
Some of the tech companies are easier to fathom because they are more transparent in the enabling tools and technologies that they provide. Partnerships will make us better, faster and more efficient, which lead to a win-win because we are able to do more to help patients with that information and insights. There are other big tech companies whose motivations aren't clear to us, and that's at minimum a yellow flag, if not a red flag.
Cedars-Sinai has been around for a long time. There is a Silicon Valley mindset of moving fast and breaking things. We don't by any means consider ourselves a luddite, we are awfully innovative, but there is a time and place to be cautious and we have to make sure we have the right information before we enter into these partnerships because it's hard to get data back after you give it to another organization.
I would never say that we won't do a big partnership; if it were the right situation and we were able to understand all aspects of it, we'd consider a big partnership.
Q: As a large academic medical center, how do you aim to stay at the forefront of healthcare delivery and innovation without the big partnerships that several other health systems have entered into?
DD: There are organizations that are moving quickly and signing up with large companies. But I think academic, community-based organizations like Cedars have a responsibility to really understand all aspects of the partnerships. Without signing on with one of the big three or four tech companies, we still accomplish a lot and I think we are among the leaders in the industry for advanced analytics capabilities.
At the moment, we aren't finding ourselves hurting in terms of skillsets. We have operationalized over a dozen machine learning models with clinical and operational impact; we have automated machine learning for greater than 700,000 to 800,000 'what if' scenarios on inpatient days; and we are creating new inpatient dashboards at two per month. We continue to find all sorts of amazing things we can do and add them to our repository of more than 1.5 billion livestream datapoints so we can inform our machine learning models.
We have a lot to brag about, so it's not so much that if you're not in a partnership you're not leading; it's about the proof of focus, tools and teams to be able to lead forward. Cedars-Sinai certainly doesn't have as many data scientists as google, which is out of reach for most companies, but we don't need that. We need the appropriate level of data scientists to solve the real-world problems of data flow, admissions, follow-up care and gaps in care.
A lot of that low-hanging or medium hanging fruit meets our needs without going to the most advanced corners of data science.