HIMSS 2014 Day 2: The State of Health Care Analytics
Day 2 Greetings from the Land of HIMSS 2014 and Health Care Analytics
The second day of HIMSS 2014 was jam packed with many activities and more information related to healthcare analytics.
Congratulations to North Memorial Hospital
First, we were pleased to see North Memorial Hospital, one of our first Health Catalyst clients, win a Healthcare Informatics Innovator Award on Monday evening. We are so grateful to work with both Dr. Nielson and Dr. Vespa in many projects.
Tom Burton speaks at a completely sold out breakfast meeting
Next, our co-founder Tom Burton spoke about healthcare analytics in a breakfast meeting. We had more than 100 people sign up with a room capacity of only 80! Tom was particularly excited because he and his team have developed some learning materials to help better explain the advantages of a Late-Binding™ data warehouse versus traditional data warehouse technologies. These learning materials included a new “shopping video” demonstration as well as some interactive card learning exercises that Tom distributed to the entire audience. Tom was particularly happy when one of the industry analysts said afterwards, “This was the best educational session I have ever attended.” We will post that Late-Binding ™ education video on our web site and blog next week.
Chilmark Analyst Overview
Chilmark, which has produced an extremely thorough and in-depth, independent review of the clinical analytics market to date, also gave a breakfast briefing with a few key industry insights:
- Chilmark included 60 analytics vendors in their 2013 review, with the assumption that they may have missed another 20. It is definitely a confusing time right now in the analytics market.
- Larger IDNs, academic medical centers and ambulatory ACOs are leading the way with analytics adoption.
- Barriers to analytics adoption include preoccupation with HIPPA and ICD-10 compliance, and the difficulty of data integration and unstructured data.
- Chilmark is already seeing some evidence of an analytics market shake up with some analytics vendors proving very effective while others are going out of business.
Data Governance Strategy
Ohio Health shared three key learnings about how they established a successful data governance strategy.
Learning #1: Building the case for change on data governance
- The following twos slide illustrate well the overwhelming task that many data analysts face when putting together a quarterly score card in many organizations. Without a centralized data warehouse, this extremely manual process often takes hundreds of hours just to produce key reports What happens when they are asked to produce these reports monthly or weekly?
Learning #2: Establishing a data governance committee
- Ohio Health was able to design and implement an impressive data governance structure with support all the way at the executive level. They established four types of bodies.
- An executive committee – chaired by the CIO and CMO
- A data governance committee
- A prioritization subcommittee
- Specific work teams chartered to help with the prioritized tasks
- The role of the executive committee included
- Provide strategic direction and a shared vision for data and analytics
- Champion and communicate EDW initiatives with respective constituents
- Review staffing models to support the program
- Assess business benefits of new project ideas and requests
- Advocate adoption, integration and sharing of data in the EDW
- Emphasize “data” as an asset, “information” as evolutionary
- Establish/oversee the EDW Data Governance committee to ensure:
- Data quality standards
- Effective and efficient use of data
- Appropriate data access policies and security
- The role of the data governance committee included
- Facilitating decisions on the accountability of data policies and processes
- Establish and document business rules
- Validate data, create policies on data access and security
- Identify data stewards and SMEs for the data
- Establish data quality expectations and strategies
- Identify KPI owners and define business rules for those KPIs
- Assist in the formation of a data dictionary
- Provide a centralized vehicle for effective communication of data-related initiatives
Learning #3 Analyze which data definitions to focus on in the initial stages of data governance processes
- Starting by trying to define the most important metrics fostered key important strategic discussions and helped facilitate the understanding of the need for data governance. For example, Ohio Health found that they had 34 length-of-stay dashboards within their system
- The three most important metrics and challenges they focused on first included
- Readmissions (within the same hospital or health system)?
- Length of stay (the finance definition or the quality definition)
- Mortality (the finance definition of the quality definition)
Key Challenges in Data Governance
- The time commitment in the beginning is long
- Decisions made in the committees have to be proactively communicated down the chain
- Different groups have a natural reluctance to lose control (“This is a great project but don’t touch my data”)
- Fear of losing jobs – a tangible anxiety that automating processes would reduce the analyst needs instead of freeing up the analyst’s time so he or she could analyze
Multiple Sessions on ACO, Population Health and Shared Risk Lessons Learned
There were multiple sessions from ACO members who shared their learnings about how important data and a foundational data platform were for ACO success. Some of the key themes from these systems were as follows.
- Catholic Health Initiatives stated that in shared risk arrangements, data is the new currency. To manage the health of populations, you have to harness the power of data to be
- Mercy Health Select cited Atrius Health CIO Dan Moriarty as saying, “To be an ACO, you need three things … if you don’t have those things, you are flying blind.”
- A common EHR
- A robust data warehouse
- A care coordination platform
- The Advisory Board was quoted as saying “51% of ACOs believed the biggest problem in their first year was related to data or IT operations.”
- UPMC shared that healthcare “data liquidity” was the foundation for accountable care
- They shared a great study, sponsored by a joint MIT Sloan Management Review and IBM, which showed the primary obstacles to widespread analytics adoption. The number one obstacle is data access – which is what an EDW provides.
- The next five reasons were organizational issues. This highlights the importance of data governance and executive sponsorship in order to begin the systematic path to data driven clinical decision making.
Michael Sills, MD (CMIO/Vice President Informatics Technology, Baylor Quality Alliance) shared key learnings about predictive analytics.
- Predictive analytics will help drive the industry towards “fee-for-value” healthcare.
- For the best chance of success, predictive applications should focus on clinical areas that are:
- Patients with data dissonance and discrepancy can often be a surrogate for high utilization of care.
- Longitudinal and ongoing cohorts are best.
- Effective change will require “time-of-care” analytics.
- We all should be thinking in terms of care “transitions” not hospital “discharges.”
Finally, Day 2 had some very fun moments including:
Health Catalyst’s Chris Keller found Cinderella waiting for him….
And of course, what would a HIMSS conference be without a walking carrot top, apple, banana and Microsoft sock men?
Oh yes – and Health Catalyst was named as one of the top influencers in social media for Tuesday.
Which was really saying something once we learned the daily volume of tweets at HIMSS!