HIMSS 2015 Finale: Days Three and Four
On April 16, 2015, HIMSS 2015 came to an end. Once again this year, it was an exciting and informative meeting. Over 40,000 people attended; 11% more than last year.
With each passing year, it becomes more and more apparent that healthcare has launched into a major period of transformation – probably the biggest transformation the industry has ever experienced. Data-driven healthcare, analytics, and improvement are at the core of this transformation.
Here is a summary of some key HIMSS analytics presentations on days three and four at HIMSS 2015.
Applying Analytics to Population Health Management
By Kori Krueger, MD, MBA and Kate Konitzer, MMI, Marshfield Clinic
The Marshfield Clinic is the largest private group medical practice in Wisconsin and one of the largest in the United States, with more than 700 physicians representing over 80 different medical specialties, more than 6,000 additional employees, and more than 50 locations in northern, central, and western Wisconsin.
The presenters pointed out that there are several stages to the Population Health Management (PHM) lifecycle which are illustrated in the figure below. Each of these components was discussed during the presentation using a hypertension example. At the Marshfield Clinic, applying analytics to these stages has helped to advance PHM by addressing the challenges at each stage of the life cycle. Those challenges were identified in this presentation and the use of analytics was described to address them.
Defining the population. In order to manage population health, an organization needs to be able to identify population cohorts. This is generally done using standard terminologies and coding systems. The effective identification of cohorts requires the implementation of an enterprise data warehouse (EDW) and the utilization of structured data and standard terminologies. In addition, the effective use of well maintained problem lists and care plans that are linked to the problem list are very helpful. The analytics infrastructure used at the Marshfield Clinic is illustrated below.
Attribution. The goal of attribution is to identify patient-provider (PCP, specialists) relationships. At the Marshfield Clinic, in situations when this is dependent on patient self-reported data, patients accurately know their PCP about 95% of the time. The other 5% of patients represent a challenge because they can either not identify a PCP or refuse to do so. The Clinic has implemented a process to continuously document this at points of care in order to fill this gap in the data.
Identifying Care Gaps. When managing a chronic condition like hypertension, it is important to identify gaps in care in monitoring and managing the condition. This can occur as a result of conflicting guidelines, lack of evidence, or data inaccuracies resulting from where and how data is collected (e.g., type of device, home monitoring, place of service). To fill this need, a governance system was put into place to determine best practice for monitoring and managing hypertension including consensus-driven, evidence-based guidelines.
Stratify Risks. Needless to say, it is important to identify types of risk, assess risk in each category and predict future risk. This can be challenging when multiple comorbidities are present. Good risk assessment can help to further define population cohorts. The Marshfield Clinic is constantly refining and improving its risk model.
Engaging the Patient. Engaged patients with chronic conditions tend to have better outcomes. Differing levels of patient engagement can result from disparity in or access to necessary resources, as well as under-funded or unfunded care management programs. The Marshfield Clinic has utilized its EHR and implemented various care management tools to improve patient engagement. They have also focused resources on “at risk” (high risk) patient populations. The result has been a more informed consumer of healthcare.
Managing Care. The Marshfield Clinic has developed a multi-faceted and multi-disciplinary approach to care management for chronic conditions like hypertension. Key challenges have revolved around care plan adherence, effective communication outside of patient-physician interactions, and variations in care. Evidence-based guidelines and advanced care management programs have helped address these challenges resulting in improved outcomes. Identified best practices have been integrated into the Marshfield EHR.
Feedback Loop. Accurate, consistent and timely feedback is important. This requires that good data be available, accessible and timely. The Marshfield Clinic has implemented an environment of continuous improvement (PDSA cycles) supported by actionable information in well-designed dashboards. Dashboard utilization is widespread including at departmental meetings. The plan is to integrate this at the point of care in the future. A sample dashboard visualization is shown below. Blue shading indicates the goal and green the actual performance for these metrics.
Measuring Outcomes. It is important to develop a consistent approach to measuring outcomes. Challenges that need to be addressed include addressing data quality issues and eliminating inappropriate variation in order to achieve the best possible outcomes. Success requires integrated data (ideally clinical, operational and claims data) in a single source of truth and implementing an effective quality improvement program.
Using this approach, the Marshfield Clinic has achieved some impressive results with respect to managing hypertension. They include the following:
- BP control rate has increased from 49.8% controlled to 77.3% of patients controlled
- This has resulted in an additional 15,182 patients now at goal that would not have been at goal in past
It is known that there is a need to treat 18 patients for 5 years to goal in order to prevent one heart attack or stroke. This means that the Marshfield Clinic has made considerable progress toward promoting health. Based on this data:
- An additional 674 heart attacks potentially avoided. Savings over 10 years (2010 $): $56,953,000
- 169 strokes potentially avoided Savings over 10 years (2010 $): $31,045,000
- Total Potential Savings: $87,998.000*
*Estimated using the CDC Chronic Disease Cost Calculator for State of Wisconsin including only direct medical expenses, not indirect societal costs
Data Analytics in the Military Health System Reorganization
By Colonel Albert Bonnema, MD, MPH, Chief, Information Delivery Division in the Defense Health Agency’s Health IT Directorate
The Military Health System (MHS) is the enterprise within the US Department of Defense that provides healthcare to active duty and retired US Military personnel and their dependents. Its mission is to provide health support for the full range of military operations and sustain the health of all who are entrusted to MHS care. The MHS includes 56 military hospitals and about 360 clinics.
In support of the MHS, the Department of Defense has launched a Healthcare Management Systems Modernization project designed to update all clinical systems throughout the MHS. This includes an $11 B contract to upgrade to a single EHR. In addition, the plan is to develop state-of-the-art analytics capabilities for the MHS. This presentation reviewed the challenges and plans associated with this analytics initiative. Contracts will be awarded later this year.
The health IT modernization effort is focused on five areas including upgrading the EHR, implementing an health information exchange (HIE), integrating applications, consolidating medical networks and modernizing analytics. The planned health IT infrastructure is illustrated in the graphic below.
While the MHS currently has analytics capabilities, they face many challenges. Demand for data and information to support MHS modernization exceeds analysts and data marts capacities. The analytics capabilities are also too fragmented. Colonel Bonnema described twenty years of legacy data marts, tools, and analysts as “feudal kingdoms.” An analysis of data objects of the ten largest data marts showed greater than 70% duplication. Colonel Bonnema described the work of a core group of analysts scattered among data marts as outstanding to ‘cutting edge’ health services research, but there is too little collaboration among them. With so much fragmentation, meticulous control and effective use of data is difficult.
In order to build analytic capacity, the MHS is focused on three areas.
- Creating an effective governance process to manage data quality.
- Standardizing and upgrading analytic skills and capabilities.
- Implementing an advanced analytics infrastructure.
The scope of the MHS analytics requirements is broad. The MHS is more than a care provider system. It is also involved in many other health related initiatives including an insurance plan, health promotion, research, military mission support, disability system, occupational health, humanitarian, public health, etc. MHS leaders and managers want a single “view” of their mission in order to manage and optimize clinical care and operations. This requires integrated data and a single source of truth.
The MHS seeks to create a ‘high reliability’ organization by focusing on five key goals as illustrated in the graphic below.
The MHS analytics service will be managed by the Information Delivery Division (IDS) which will be in charge of data services (including the creation of registries and dashboards), web strategies, enterprise intelligence and HIE. This division will be focused on high reliability goals 2, 3 and 4 in the above graphic – improving care, population health and lowering costs, inspiring trust in health IT and data, and empowering individuals to improve their health. The IDS will focus on data quality, data availability and data access tools including visualization tools with drill down capabilities, data discovery tools and other types of analytical tools.
The MHS has embarked on a comprehensive, complex and challenging analytics journey. It will be interesting to see how it unfolds.
Business Intelligence for Sepsis and Heart Failure Readmissions
By Christopher Kodama, MD, MBA President, MultiCare Connected Health
MultiCare is a not-for-profit health care organization with more than 10,000 employees and a comprehensive network of services in western Washington. MultiCare is made up of five hospitals, numerous outpatient specialty centers, primary and urgent care clinics, as well as a variety of other services and community outreach programs. MultiCare has a history of commitment to ongoing quality improvement and developing collaborative care partnerships in the community.
Under the leadership of Dr. Christopher Kodama, MultiCare Connected Care, an Accountable Care Organization (ACO) and Clinically Integrated Network (CIN), was developed in response to health reform and market driven opportunities. MultiCare Connected Care is an opportunity for MultiCare and independent community providers to participate in risk- and performance-based contracting with payers, and to participate with regional employer accountable care networks. The guiding principles of the MultiCare Connected Care Network are:
- Commitment to quality
- Enhancing the patient experience
- Driving waste out of the system
Physicians participating in the MultiCare Connected Care Network need to participate in a health data information exchange process that will insure that real-time information that supports care coordination across the continuum of care is available to all providers on a patient’s care team. Each practice must also agree to participate in the guiding principles of MultiCare Connected Care through the use of clinical guidelines and care pathways developed by the physician-led Clinical Collaboratives. Each group must support specific coordination of care activities to achieve improved patient outcomes, satisfaction and specified targets for cost and quality improvement. The collaboratives are characterized by being physician-led, evidence-based, multi-disciplinary and adequately resourced with support skillsets. To date, six Clinical Collaboratives have been formed including critical care, women’s care, surgery, cardiac care, medicine and pediatrics.
This presentation reviewed how MultiCare leveraged data warehousing, analytics and knowledge management to reduce mortality from septicemia, demonstrate promising results in lowering heart failure admissions and prepare for value-based care. Dr. Kodama also discussed strategies for using data and analytics to scale quality improvement and for developing an organizational culture of continuous improvement.
The Collaborative journey began in 2010 when a DRG-based analysis identified sepsis as a good candidate to serve as a “proof of concept” pilot based on identified cost and quality issues. In 2012, the sepsis initiative evolved into the Critical Care Collaborative. By 2014, this had evolved into the six collaboratives listed above, each with two work teams. This year, it is anticipated that other collaboratives will be launched including back pain/spine and glycemic control.
An initial step in the sepsis initiative was to develop a standard, evidence-based clinical pathway and order sets. The Modified Early Warning System (MEWS) was used to monitor and identify patients, and it was automated to improve detection of sepsis patients. MEWS is a tool for nurses to help monitor their patients and improve how quickly a patient experiencing a sudden decline receives clinical care. Analytics was used to develop a sepsis dashboard for the improvement team. The dashboard was carefully designed to provide easy visual identification of important trends and issues. An example of the sepsis dashboard is illustrated below.
Results of this improvement effort have been impressive. The sepsis mortality rate has decreased from 15-17% to 7-8.5% as of Dec 28, 2014. Dr. Bonnema pointed out that this improvement work is never expected to stop. All improvement efforts launched by this Collaborative and all others is expected to be focused on continuous improvement.
Similar to the sepsis effort, the heart failure readmission project began with a clear analysis of the problem and the development of a standard, evidence-based clinical pathway and order sets. The involved clinicians identified that a standard way identifying the heart ejection fraction was a critically important data point, and this was addressed early in the process. It was also recognized that the management of heart failure readmissions is more of an outpatient challenge than an inpatient challenge. The following steps were identified as very important in preventing heart failure readmissions:
- Notify Primary Care Physician of discharge
- On-time discharge summary
- Readmission Risk Assessment
- Education with “Teach Back”
- Advanced Care Planning
- Discharge instructions
- Medication Reconciliation
- Interim follow-up phone call
- Timely follow-up appointments
- Specialty Consult
Like sepsis, a dashboard was carefully designed to provide easy visual identification of important trends and issues. An example of the heart failure dashboard is illustrated below.
Dr. Kodama also reviewed the technology infrastructure, data governance structure and the Operations Council which determines what collaboratives are launched, authorizes improvement projects by each collaborative and provides the necessary support for success. He also reviewed several lessons learned including:
- The importance of patience and proof of concept before rapid deployment of successful pilots
- The importance of attention to good change management
- The critical role of education to success
- The importance of prioritization and focus (do not try to “boil the ocean”)
- The importance of thoughtful design and deployment
- Not to let “perfection” stand in the way of progress when “good enough” will get you where you need to go
A clear approach to decision-making and process approval