HIMSS 2014 Day 1: The State of Health Care Analytics

My Folder

himss 2014 booth

Day 1 Greetings From the Land of HIMSS 2014 and Health Care Analytics

Fifty (yes 50) of us from Health Catalyst arrived, with only one flight delay casualty.  Orlando was a perfect temperature today, and the show kicked off very well.  Last year, the show seemed crowded and dense, but the halls and the floor seemed much more spacious and well-laid-out this year.  It had a more open and productive feel, whereas last year it felt frenzied and sometimes uncomfortably crowded.

Luckily, our booth was set up well, and we had non-stop traffic. We thought last year was hard to match, but we had more visitors who stayed much longer and asked great questions.  We had two meeting rooms constantly occupied with customer  meetings, interspersed with members of the media and analysts coming to visit us.   We heard over and over again that “everyone” was talking about “analytics” or “population health.”  Which was good because we had a great story to tell.

Today I attended two HIMSS education sessions that stood to me because of their challenges with data and analytics in their very unique circumstances:

Enabling Analytics for Patient Outcomes to Reflect Your Care Model

By Efrem Castillo, MD, the Regional Medical Director of WellMed. 

Wellmed is a health system treating over 130,000 patients while employing 130 physicians in 82 clinics across Texas and Florida.

What intrigued me was the common set of challenges they had with their data

  • Internal quality reporting  did not reflect the macro outcomes
  • Data was inaccurate and not timely
  • Metrics were based on data available
  • Different sources of truth
  • Net results:  Physicians did not trust the reports

I thought the following slide summarized their data issues very well.

himss 2014 data problems

Do these health care data problems look familiar?  These data  issues are such common problems we see almost every day in our discussion with health systems.  The biggest downside of these data issues is that doctors learn to not trust the data.  Without the physician support, data-driven healthcare cannot scale and thrive.  We see this problem over and over again.  A health care data warehouse is the only solution that can fundamentally and systematically solve these problems.

Data-driven:  A Pioneer ACO applies data to improve patient care

Domnique Morgan-Solomon, MPH and Vice President of Population Health of the Stewart Health Care System.

Dominque told a fascinating story of how Steward Health Care has navigated the mostly unchartered territory of being a successful pioneer ACO.  Steward Health is the second largest care network in New England, including 11 hospitals, almost 2100 beds, and 2700 integrated physicians, most of which are private practice (2000).

Dominque shared how she was literally told to get a program staffed, up and running in 90 days!  With no real blueprint.  She shared many challenges, but several were particularly interesting.

  • They quickly learned that measuring risk was fraught with risk.  The claims data showed a set of high risk patients that in many cases overlooked a different set of high risk patients that could only unearthed by clinical data.  They learned that you absolutely cannot manage risk without having all three set of data:  clinical, payer, and patient provided.
  • They learned that the high risk patients were not necessarily the patients to focus on for ACO improvement.  In many cases, the high risk patients were already getting the best care possible at their stage.  Instead, they needed to learn to identify the “rising risk” patients – the patients at risk for rising costs where they still had time to intervene and prevent.
  • And they experienced mindset challenges with physicians and patients.  It was hard to convince the physicians to focus on treatment that would only show fruit in 18 months.  And it was difficult or confusing for patients to understand why they were getting treated differently or why they were in “a program.”

Some of their key learnings included:

  • They learned that collecting and reporting health care data was meaningless unless it actionable and was targeted at a specific audience.  They discovered they needed different reports for physicians, care management teams, executives, and patients – which each being able to see a similar situation with different data needs.
  • They learned they needed an entirely different set of metrics.  In fact, a suite of balance metrics.  No longer could they laser focus on a narrow set  – they needed a balance of metrics.    I thought the following slide captured well the types and categories of metrics they learned to collect and monitor.

himss 2014 transforming data elements into performance metrics

  • They also learned that it took them a while to evolve their dashboards to the point where both the clinicians and the leadership could understand and evaluate performance.

Highlights of Their Year 1 ACO Results

  • They achieved a surplus in 2012
  • They had measurable decreases in readmission rates and SNF ALOS
  • They improved overall care retention by 5%

Overall I was particularly impressed with the resourcefulness of Steward Health in figuring out what worked and what didn’t quickly.  This really captured the agile principle, which is a core principle in Health Catalyst in both how we develop products and how we roll out implementation at our clients.