Treating healthcare data as an asset
On day two of HIMSS14, I had the pleasure of listening to leaders from UPMC, MD Anderson and Amedisys share their stories about how their organizations are turning information into an asset. At the third annual IBM Big Data & Analytics luncheon, attendees from across the healthcare ecosystem learned about what it takes to get started on their journey, how to anticipate and address potential road blocks and how to garner support across constituencies. I am very privileged to host this special luncheon every year where leaders like those listed below take the time to share their experiences with us, and inspire us to move big data mountains.
Sri Srinivasan, worldwide big data and analytics industry leader for healthcare, started the conversation by polling the audience. First he asked how many organizations treat data as an asset, in answer to which there was only a slight raise of hands. Next he asked how many people have big data initiatives on the roadmap—more hands went up immediately. Finally, he asked how many are having difficulty getting executive sponsorship and support needed internally for such initiatives: many more hands went up to create a sea of frustrated palms.
This is not a surprise. One of the most challenging things any chief information officer (CIO) or chief medical information officer (CMIO) faces is getting support from across constituencies to transform the culture in IT and lines of business, not only to be more data-driven, but to recognize the value of combining data across silos to get greater insights. So, what is the catalyst to help get things started?
What I learned from the speakers is that when the executive leadership in an organization isn’t getting the reports and insights they need to make strategic decisions that make or break the business, or when the CMO or CMIO doesn’t exactly know what is contributing to hospital readmissions or lack of adherence to core measures, the issue seems to bubble up to the top of the list!
For example, Lisa Khorey, vice president of Enterprise Systems and Data Management at UPMC, gave one example of how her organization realized they needed to make a change. She talked about how management wanted a simple comparison of how UPMC was performing compared to other healthcare organizations. Turns out, the requirements were too broad; they couldn’t bring the data together to figure it out. So they realigned the requirements to compare performance of one hospital with another hospital. It took 30 people and multiple months to figure it out. By the time they produced the report, it was out of date.
Dr. Frenzel, the CMIO from University of Texas, MD Anderson Cancer Center, spoke about the need to enable researchers to be more agile and fluid when pursuing a deep understanding of a patient’s molecular biology. Their current situation of manually pulling data together would in no way scale to support their mission of making cancer history.
Jeremy Taylor, CIO from Amedisys had a bit of a different story. Amedisys is one of the largest home health providers. The leadership team recognized the need to transform their approach to care in order to have a sustainable business and identify new areas for care delivery, and developed a five-year roadmap for enterprise analytics.
Each of these three scenarios faces significant challenges, but all emanate from the need to systematically gather data across service lines to gain an enterprise view of patient population, cost of care, treatments, outcomes, molecular biology and more.
These organizations, like all organizations, are being driven by a business need and, also like other organizations, they have big challenges with data: bringing data together, protecting privacy, ensuring data quality and supplying reports and insights to multiple lines of business. It is to their benefit that they listen and learn from those who are on the same path, as it benefits all innovative organizations who have set out to make real changes.
What follows is a partial transcript of the IBM Big Data & Analytics luncheon:
Big data and analytics offers significant upside and promise to transform healthcare, but the proper foundation and groundwork must be laid if we’re to take advantage of the richness that innovative capabilities like text analytics, natural language processing and cognitive computing can offer in the future.
Lisa: Big question. Two major challenges—first: you can check all the boxes of everything you want to buy, but the actual knowledge of how you need to do it, is critical. Teach IT organizations. We must get over ourselves and let someone else teach us. You might start it as a journey, but I never see the end. Second: there was thinking that “we already do that” for population health and more. Well, there were in pockets, but they weren’t brought together. The approach was to be additive and not disrupt what they were doing, but get some quick wins.
Frenzel: Federated institutional reporting environment (FIRE). Bring data together with context. Create a federated model for data governance. Data owners were brought in and we created a platform to migrate the data to, and then gave the data governance power. Now they have people partnering with them to give them data. Getting buy in from folks and showing value is of utmost importance.
Jeremy: Getting data into the warehouse wasn’t a problem—it will be getting the people used to understanding the richness of the data that will be the challenge. Resources are difficult to find and it takes a unique type of person to interpret the data. Not having research background in your staff is tough when it comes to locating this type of person. We're talking internally on how to expand in that area and train people on how to understand the data, use it and spread the knowledge.
How has what you’re doing impacted the healthcare consumer?
Lisa: It is important to show value sooner rather than later, and, whatever we do, it has to be repeatable across service lines. The value is in looking across the continuum of care to understand the cost of care. This perspective changes decisions we make.
Frenzel: We got access to other technology like NLP, so now that clinicians can see what they can do with structured data, they’re looking at what they can get from unstructured data. Structured dictation opened people’s eyes on what cognitive computing can do.
Jeremy: Predict the right number of visits and treatments to optimize the outcomes.
What final words of wisdom would you give to the luncheon attendees today to help them be the successful champion in their organization, driving analytics implementation steadily forward?
Lisa: Go home and make an application inventory tomorrow
Frenzel: You have to own your analytics and warehousing, wouldn't recommend hiring it out
Jeremy: Understand your data and begin with end in mind
Read more about HIMSS on the Hub and add your thoughts on healthcare data in the comments below.