Healthcare Analytics Summit 2017 Final Day: Thursday Recap
Paul Horstmeier opened day two of the 2017 Healthcare Analytics Summit™ (HAS) conference with a recap of the day one, mostly notably the “Dan Burton ping pong challenge” and photos from Wednesday’s superhero night, which showed a lot of fun, costumes, and stiff ping pong competition. He shared some day one analytics, including that the conference answered more than 15,000 poll questions the previous day. Also, the HAS™ 17 app data showed a lot of connections made over the past few days.
Horstmeier shared a pre-summit survey finding that 44 percent of attendees believe that population health is moderately successful. When asked which new important data sets are still missing, attendees overwhelming chose “social determinants of health.” The number one impediment to using machine learning and associated risk scores is lack of data.
And, for those who were wondering, the first day’s sessions ended at 5:03pm.
Keynote: The Patient’s Power in Improving Health and Care
Maureen Bisognano, President Emerita and Senior Fellow, IHI
When Thursday morning keynote speaker, Maureen Bisognano, President Emerita and Senior Fellow, Institute for Healthcare Improvement (IHI), opened her presentation, she said she felt like she was “in the future” at HAS 17. After Bisognano gave a nod to the technological innovations discussed this week, she shared her vision for the future of healthcare, one where the industry achieves the IHI Triple Aim.
Maureen conveyed an example of how schools can be “flipped,” where students watch teacher lectures at home, and then do homework in the classroom. Students use smartphones, laptops, and other mobile devices to complete their lessons. The results have been remarkable, with graduation rates increasing to 90 percent, and college attendance rates rising 17 points to 80 percent. It helped teachers identify the “silent failers” (the students most in need) and then work with them in the classroom. Teachers went from being “the sage on the stage” to being “the guide on the side.”
Bisognano discussed a strategy for improving the patient experience and health of populations while lowering cost, which she called “flipping healthcare.” In a flipped healthcare system, the patient—and what matters most to them—drives care, versus the traditional, paternalistic approach in which physicians and organizations make decisions. It’s a movement from “what’s the matter?” to “what matters to you?”
Patient engagement, Bisognano said, may be the next “blockbuster drug” in healthcare. In her experience, Bisognano explained that patients aren’t looking for more office visits, but instead more community around health concerns, easy access to information (i.e., the ability to email or text questions to their care team)—in short more human connection without having to go to the office.
Bisognano presented examples of programs to increase patient engagement in their own care. In one based in Sweden, patients performed their own dialysis in a hospital wing dedicated to their care. This saved the hospital significant money and allowed opportunities to patients to improve quality of life in addition to healthcare (i.e., employment counseling for those out of work). Bisognano also discussed a program called “centering pregnancy,” where expecting mothers received care as part of a group, versus individual sessions. This allowed for social support from others in the group and resulted in reduced risk of preterm birth, reduced racial disparities, and increased breast feeding.
Keynote: The Cost of Healthcare-A Revisionist History
Robert DeMichiei, Executive VP and Chief Financial Officer, University of Pittsburgh Medical Center
Robert DeMichiei says that healthcare has a lack of focus on costs. In fact, we’ve become bystanders to costs, where we see it and comment on it, but do nothing about it. We think it’s too hard, that it would cost too much to implement activity-based costing (ABC). There are plenty of excuses.
On the other hand, we are obsessed with revenue, driven by volume, reliant on commercial payers, and addicted to growth. We push more and more costs to consumers, and state Medicaid programs are nearing bankruptcy.
During his presentation, he questioned whether the policies and practices this obsession has inspired are sustainable in today’s landscape of financially overburdened patients and overstretched Medicare coffers. Whereas as other industries have shifted to more sustainable models, healthcare has yet to do so. Accurately measuring cost, he said, is critical to understanding areas for improvement.
“We have a cost problem!” DeMichiei said, and added that healthcare sustains the problem by ignoring that costs are too high.
DeMichiei explained that healthcare’s traditional financial dependence on volume, revenue, and commercial payers doesn’t position the industry to succeed in the transforming landscape of healthcare, where health systems must reduce cost and deliver value to patients.
As a cost management solution, DeMichiei proposed activity-based costing (ABC), an approach where cost is assigned based on patient utilization versus time spent receiving care. He has applied ABC at UPMC to drive cost productivity, identify clinical variation, measure service line performance, and link cost and quality. UPMC has also partnered with Health Catalyst to develop the CORUS™ suite, an ABC platform to manage the true cost of care.
University of Pittsburgh Medical Center saw $42 million in cost reduction opportunities for service lines; shaved $5 million from supply costs; created transparency in identifying practice variation; and improved processing time by 97 percent.
ABC, DeMichiei explained, provides the needed scientific approach to reduce cost and improve quality in healthcare, particularly as healthcare continues the transition from fee-for-service to value-based reimbursement. It generates critical, accurate data to drive actionable, defensible change.
Keynote: The Population Health Engine
David Nash, MD, Founding Dean of the Jefferson College of Population Health
David Nash, MD, founding dean of the Jefferson College of Population Health, divided his presentation into four parts: 1) how healthcare got into its current predicament, 2) defining population health, 3) closing the gap between population health and the current state of health, and 4) what it all means moving forward.
How Healthcare Got into Its Current Predicament
Dr. Nash referenced a lesson from Robert DeMichiei’s earlier presentation—that healthcare can’t fix cost by focusing on cost. Fixing costing is instead about examining variation and physician practices (versus continuing to follow a revenue-driven model).
Dr. Nash referenced a National Academy of Medicine (NAM) study that ranked the U.S. number 17 in the world in patient quality of life and value in spending and predicted shortening lifespans for millennials compared with their parents’ generation. Other, higher-ranking countries in the NAM study spent more healthcare money of social aspects of care, while the U.S. spent more on expensive test.
The U.S. suffers from variation and overutilization, said Dr. Nash. Healthcare won’t transform until it addresses this, as social determinants of health (i.e., environment and credit score), predict more for lifespan that healthcare utilization.
Defining Population Health
Dr. Nash shared a framework for population health. The structure linked health outcomes with morbidity, mortality, and quality of life; health determinants with medical care, socioeconomic status, and genetics; and policies and interventions with social, environmental, and individual factors. He referenced the NAM report “Vital Directions for Health and Healthcare,” which designated the following priorities for population health: pay for value, empower people, active communities, and connect care.
Closing the Gap Between Population Health and the Current State of Health
In the quest for population health, Dr. Nash said the industry needs to shrink the gaps in social determinants of health. Part of this is adopting population health intelligence, where physicians self-evaluate how well they’re caring for patients and where they can improve. Next steps will include leveraging artificial intelligence to determine patients at risk as well as early identification of adverse events, such as sepsis and chemotherapy failure.
What It All Means Moving Forward
Key themes moving forward in the quest for population health, according to Dr. Nash, are transparency and accountability. As he said, no outcome will mean no income. Dr. Nash explained that transformation will require a change in culture that includes practice based on evidence, reduction of unexplained clinical variation, less adherence to professional autonomy, continuous measurement, and engagement with patients across the continuum of care.
Breakout Sessions: Wave 3
Session 22 – The Data Operating System: What It Really Means and Why You Will Need It (Technical Session)
Imran Qureshi, Chief Software Development Officer, Health Catalyst
By 2025, it is projected that 41 percent of hospitals will be operating with negative margins. Risk management and population health need far more data than we currently have. Our current single monolithic app model isn’t working. What we need is to evolve from our data warehouses to a data operating system (DOS). Imran Qureshi walked participants through what a DOS™ is and why it’s crucial for a healthcare system’s survival.
A DOS contains aggregated data from multiple sources, is designed for data synthesis, works with all EHRs in your health system, supports web and mobile app creation, enables SQL and machine learning queries, updates data in real-time, and provides centralized security at the app and data levels. DOS provides clinicians with all the data they need in one place as well as a place to see synthesized data that can inform the decisions they need to make with their patients. Hospital IT find that a DOS enables them to scale existing data warehouses and enable a great deal of reports and apps without the need to hire additional people. Calculating risk and trending it over time is what a DOS can do for hospital leaders. A DOS even helps keep data secure while still providing access to the data, which data administrators love.
Session 23: Yours, Mine, Ours – Provider-Payer Convergence and the Future of Data Analytics (Edu Session)
John Moore, Founder and Managing Partner, Chilmark Research
The move to value-based care is accelerating, but providers remain ill-prepared to shift to a quality-driven business model. Early experiments with at-risk arrangements are inconclusive, while ACOs struggle to invest in requisite organizational, process and technology strategies. Provider/payer relationships are increasing, but there are no clear winners when it comes to financial and clinical models.
The emerging provider/payer market requires new strategies that bring these two groups together to provide a health service that is responsive to consumer needs, offers differentiation in the market and delivers high-value. Unfortunately, this is easier said than done. The silos that currently exist between payers and providers defeat the move to a consumer centric model and the current infrastructure doesn’t support convergence. What is needed is a common data platform that provides a single source of truth to facilitate better care and reduce costs.
Convergence is not for everyone and best practices don’t exist. If you decide to move forward, start by getting your internal house in order. Look at what the early innovators have done and make sure you have executive leadership that is in it for the long haul. Whether or not you choose a convergence model, migration to VBC is inevitable. And, as Moore said, “you need to get on board, or get out of the business.”
Session 24: A Bad Moon on the Rise? How We Raised the Sense of Urgency and Built a Strategy Shift with our Board of Directors (Case Study)
Duncan Gallagher, President, Donegal Advisory Services, Former Executive Vice President, CFO, CAO at Allina Health
Duncan Gallagher, President of Donegal Advisory Services, opened with a slide that said: “133 million Americans are enrolled in government healthcare programs—today.” This was to emphasize the fact that if anyone’s worried about a government-driven healthcare system coming to America—well, it’s already here. It’s now up to private systems to adapt to the value-based payment system.
For the rest of the session, Gallagher explained Allina Health’s shift in strategy to pursue greater risk sharing and embrace partnerships that would enable Allina to be more efficient. This was accomplished through many discussions with Allina’s management team and board about topics like the health system’s economic outlook, the insurance market, and the critical factors in transitioning to population health management.
It wasn’t easy, though. They had to work through uncomfortable questions. For example, they needed to decide how to balance the tension between near-term operating margin goals and accepting more outcomes-based risk to accelerate their capabilities in the transition to outcomes-based payments. They also acknowledged they may not be able to answer everyone’s questions now—but they expected the answers would be clearer later on.
Gallagher’s parting words on how to move forward in an evolving, uncertain market:
- Prioritize your organizational migration forward.
- Make difficult choices.
- Move in many increments, but move with haste.
- Partner to access capabilities that you do not possess.
- Be prepared to adjust your initiatives.
- Set measurable goals and hold leaders accountable.
- Embrace transformation and disruption from external sources.
- Engage your board.
Session 25: Achieving the Promise of PROMs (Case Study)
Neil W. Wagle, MD, MBA, Associate Chief Quality Officer, Partners HealthCare
Neil W. Wagle, MD, MBA, Associate Chief Quality Officer, Partners HealthCare, opened his HAS 17 breakout with a personal story on selecting an orthopedic surgeon for his mother. As many people do, he asked around for advice. He turned to his respected mentor, who recommended a surgeon. Dr. Wagle and his mother, however, were disappointed in an abrupt and unsatisfying visit with the recommended doctor. When Dr. Wagle considered how his mentor made the choice, he concluded that his mentor based his recommendation solely on surgical outcome metrics for the surgeon, where patient-reported outcomes measures (PROMs) would have provided a more accurate picture of what he and his mother could expect.
In many similar medical decision-making processes, PROMs are the outcome metric of relevance, Dr. Wagle explained. For example, if a patient considering radical prostatectomy only looks at five-year mortality rate for the procedure, he’ll learn little about what to expect (i.e., rates of incontinence and sexual dysfunction) and won’t be able to make an informed decision. PROMs provide the deeper dive into the full outcome story (i.e., how long incontinence or sexual dysfunction might last and the frequency of comorbidities such as depression).
Benefits of PROMS, Dr. Wagle explained impact both patients and health systems. In addition to better care for patient, PROMs help physicians save time and enable more personalized care. They highlight clinically meaningful change by telling the whole story quickly, including history, improvement, and setbacks.
Dr. Wagle explained that while implementation of PROMs is challenging, it’s feasible with the right approach. He shared seven steps to implement PROMs:
- Get leadership buy-in. (Tip: use examples from other clinics that have improved outcomes with PROMs.)
- Form three sub teams: implementation, technology, and data and analytics.
- Choose a technology platform (i.e., EHR native, integrated side-car, or third party).
- Choose a collection modality (i.e., at home, clinic, or multimodal) and get the associated IT working well.
- Choose an initial target concern. (Tip: choose an area where PROMs is the outcome of relevance.)
- Choose a priority: implementation versus data and analysis.
- Agree on which PROM to pursue. Choose one with a short list of questions, that’s relevant to clinical care, validated, an industry standard, and covered by a PROMIS domain (Patient-Reported Outcomes Measurement Information System).
Session 26: One ACO/Integrated Delivery System’s Governance Journey (Case Study)
Christopher Kodama, MD, President, MultiCare Connected Care, and Florence Change, Executive Vice President, Chief Operating Officer, MultiCare Health System
Representing MultiCare Heath System, Dr. Kodama and Florence Chang shared two case studies, one operational the other clinical, to demonstrate the importance of analytics and governance in improvement work. MultiCare uses a shared governance approach that enables them to align their improvement efforts with organizational priorities. It integrated financial data into outcomes improvement initiatives to support decision-making and prioritization, and discovered that change management is critical and so is focus (don’t boil the ocean).
Dr. Kodama introduced the VUCA principles, traditionally used in the military to train leaders in highly volatile situations. He then connected the VUCA principles to what is needed to succeed in this ever-shifting, volatile healthcare environment:
- Volatility yields to vision and vision requires governance.
- Uncertainty yields to understanding and understanding requires analytics.
- Complexity yields to clarity and clarity requires analytics.
- Ambiguity yields to agility and agility requires governance.
Vision and understanding include:
- Prioritizing opportunities – focus on the critical few.
- Involving stakeholders in setting the vision.
- Explaining the why.
- Setting clear expectations for all leaders.
Clarity and agility include:
- Making informed decisions.
- Diffusing transparent performance data.
- Rapidly checking and adjusting.
The bottom line is that availability of information, data, and action plans will drive quality. When we make data available in a relevant way, others will access it and their desire to become a high-performer will increase.
Session 27: A Population Health Management Diabetes Case Study
Rona Y. Sonabend, MD, Medical Director, Clinical Systems Integration Process Improvement, Texas Children’s Hospital
Diabetes Mellitus (DM) affects 208,000 children in the U.S. and this number is increasing. The national cost for all diabetes care was $245 billion in 2012. So, Texas Children’s had a straightforward mission: Deliver the highest quality of care to patients with diabetes mellitus at Texas Children’s.
Texas Children’s needed to reduce variation and fragmentation in the care it provided. But the turning point for them was realizing that incremental change would not be enough to transform diabetes care. It needed a comprehensive, cross continuum care improvement program for diabetes patients.
It began by developing mission statements for the five diabetes care process teams in the hospital: ambulatory, inpatient, high risk, community, and education. The teams improved preventive care and assigned every patient a risk score to establish risk-based interventions. This didn’t previously exist for pediatric care and it decreased preventable readmissions by 30 percent.
From a population health management standpoint, the program significantly increased the percentages of influenza vaccine rates and patients with individualized school packets. It improved pediatric provider knowledge. The program deployed continuing education for physicians, nurses, staff, and patients, and built a standardized diabetes education model. It standardized and improved inpatient care; delivered care earlier, decreased LOS, and decreased the time to get insulin on board.
This was all made possible by empowering nurses to make medical decisions and be part of evidence-based guidelines. In fact, Texas Children’s created a model for cultural change that had vision, leadership, support and infrastructure, dissemination of information, and was multidisciplinary and inclusive.
The improvement programs progressed through four informatics maturity stages:
- data reporting – defining how many patients.
- data analytics – defining who the patients are and where they are.
- predictive analytics – defining who’s at risk.
- prescriptive analytics – defining how to prevent it.
Ultimately, Texas Children’s accomplished its objectives through preparation readiness and engagement, developing infrastructure with technological support, removing barriers to implementation, and delivering care across the continuum to create success.
Breakout Session: Wave 4
Session 28: Unleashing Data: The Key to Driving Massive Improvements (Two-hour Deep Dive, Education Session)
Tom D. Burton, Co-founder and Chief Improvement Officer, Health Catalyst
Tom Burton provided another exciting game experience designed to teach about the importance of unleashing data and focusing efforts across the improvement spectrum. He has created a game called Spectrum, which is based on the popular strategy game, 7 Wonders. Spectrum teaches about the necessity of investing in analytics training and infrastructure and recognizing how to sustain and spread improvement.
Burton explained that across the spectrum of value and effort, organizations need to focus on light touch and organic improvements, fast track improvements and the high effort and comprehensive improvements.
The game Spectrum teaches that in order to do improvement work, you must invest in data infrastructure and people before you can progress on your journey toward a data-driven culture. You must also perform ongoing opportunity analysis to determine what your next best move is for your organization. Balancing the light effort and deep continuous improvements win you points in Spectrum; but in real life, this balance of improvement work earns you meaningful outcomes at your organization. Don’t forget the need for interdepartmental collaboration, system-wide adoption, and reduction of conflict. Spectrum is a fun and engaging way to learn strategies that can be implemented at your organization to unleash the data across the spectrum of improvement.
Session 29: Supercharge Your Improvement Efforts with Predictive Analytics (Technical Session)
Chris DeRienzo, MD, MPP, FAAP, Chief Quality Officer, Mission Health; and Andrew O. Johnson, PhD, Manager, Data Science, Clinical & Business Analytics, Mission Health
Chris DeRienzo, MD, and Chief Quality Officer at Mission Health started the session by reinforcing the importance of using analytics for improvement efforts by saying: “We are dead in the water without a culture of continuous improvement, grounded in analytics, that permeates everything we do, and all that we are.” He also stated that “to survive, we have to evolve.” One way to accomplish this is by building a robust internal data science team that’s integrated into a continuous improvement analytics strategy. Here’s are the steps Mission Health took to do just that:
- They began with the BIG(GER) aim by asking the question: “What do we need to best deliver on this promise in a population health world?”
- They created a vision and support for data-driven continuous improvement grounded in analytics across their organization by coaching clinical leaders on the power of data science to improve everything they do.
- They built their analytics team with in-house data science expertise.
- They cultivated a strong organizational motivation, capability, and process to move to predictive analytics operationally and clinically (initially focused on readmissions).
Session 30: Using Predictive Analytics and Machine Learning to Lower Systemwide Readmissions (Case Study)
David M. Wild, MD, Vice President, Lean Promotion, The University of Kansas Health System; and Chris Harper, MBAi, MPM, Director Business Architecture and Analytics, The University of Kansas Health System
Many systems struggle to reduce all-cause readmissions, as national statistics show. 20 percent of Medicare patients are readmitted within 30 days. The total cost of these readmissions is $26 billion, of which $17 billion is preventable. University of Kansas Health System believes it must provide patients efficient, value-added, effective, and patient-centered care during and after discharge from the hospital—which in turn, will reduce hospital readmissions and lower costs.
To begin, the University of Kansas Health System defined the readmissions problem in general terms:
- Hospital readmissions may result from actions taken or omitted during the initial hospital stay.
- Readmissions are expensive, and consume a disproportionate share of expenditures for inpatient hospital care.
- Readmission rates have been established as a basis for comparing hospital performance.
Then they defined their particular readmissions problem:
- No clear process to prevent readmissions and identify high-risk or readmitted patients to guide clinical practice.
- High variability across units and service providers.
- No specific process for readmitted patients.
- Lack of documentation around education to the patient.
- Documentation system and access does not support transparent
University of Kansas Health System developed a Lean approach and supported organizational design that resulted in a Continuum of Care Advisory Team (CAT), which drove the overall program. It created education and engagement in five key areas. Then it deployed machine learning and predictive analytics, using an EDW with data from multiple sources. The objective was to understand the risk factors driving readmissions.
Early results were promising. From the fall of 2015 to the fall of 2016, all-cause heart failure rates dropped from an average of 29 percent to an average of 10 percent. Heart failure readmission rates dropped from 12 percent to 3.8 percent, and non-heart failure readmission dropped from 15 percent to less than 8 percent.
Machine learning, predictive analytics, and Lean were used to identify patients at high risk of readmission, guide clinical interventions, and redesign care. Key to this process was having dedicated and focused business and clinical project SMEs engaged to drive the analytical work.
Session 31: Have Data, Need Analysts: Lessons Learned from the Woodworking Industry
John Wadsworth, Senior Vice President, Client Engagement, Health Catalyst
John opened his presentation with a story about love in the 7th grade. It’s when he met his wife (!), but what he really meant was that this is when he fell in love with woodworking. And as only a man with a passion for woodworking can, John showed the audience how it can be a metaphor for healthcare data analysis. He described that both required special tools with skilled operators and both require efficient layout and flow. Attendees were challenged to build balsa-wood airplanes to demonstrate what happens when you get the right tools, people, and skills involved—and what happens when you don’t. The key takeaways from this interaction session:
- Understand the analytic total cost of ownership
- (TCO) = tool + operator + skill.
- EDWs and EHRs are not competitors; they’re complementary tools in your analytic ecosystem.
- Don’t force your best analysts into management.
- Data analysis is more valuable than data query, data movement, data modeling, and data visualization.
- Pay your best analysts like you pay your physicians.
Session 32: Chronic Disease Management Reduces Readmissions (Case Study)
Amber Theel, RN, BSN, MBA, CPHQ, CPHRM, Director, Quality Outcomes and Metrics, MultiCare Health System
Rooted in a long-standing commitment to quality of care, MultiCare Health System’s focus is on doing the best thing for patients the first time, every time. So, when the system found that costly COPD 30-day readmissions were higher than expected, and that many patients being readmitted for heart failure also had COPD, they set out to develop an organizational strategy that would improve patient care processes and reduce readmission rates.
To launch this effort, MultiCare’s existing Medicine Collaborative formed a workgroup that was tasked with identifying opportunities for improvement. Supported by strong leadership, the group included a broad base of engaged participation, with representation from many care locations and disciplines. The end product was a chronic disease management program that resulted in:
- 13 percent reduction in readmission rate.
- 82 percent of patients’ PCPs were notified of patient discharge.
- 95 percent of patients with COPD were assessed for readmission risk.
- 89 percent increase in COPD order set utilization.
MultiCare credits the success to leveraging the insights of an interdisciplinary team, using evidence-based care bundles at discharge (to standardize care and improve outcomes), and leveraging analytics to promote adoption.
If you are interested in initiating a similar program, Theel says go with what you already know works, and apply lessons learned from other improvement efforts. Also, recognize that thoughtful, data-driven planning up front saves time in the long run and ultimately leads to better decisions that result in better care.
Session 33: Using Predictive Analytics and Real Time Decision Support to Reduce Harm (Educational Session)
Stan Pestotnik, MS, RPh – Vice President, Patient Safety Products, Health Catalyst
Stan Pestotnik, MS, RPh, Vice President, Patient Safety Products, Health Catalyst, presented on the essential role data and analytics must play in reducing the rate of patient harm in U.S. health systems. Given declining rates of patient safety (a tenfold rise in patient harm since 1999 and 400,000 lives lost per year) and adverse financial implications with the transition to value-based care, organizations must look at patient safety.
Health systems, explained Pestotnik, face consistent roadblocks in improving patient safety. It’s often displaced by other organizational concerns, struggle with metrics glut and system literacy, and organizations tend to think they’re complying and that incentives will improve quality. The solution, he explained, is a systems approach to safety that addresses culture, as culture will always lead technology.
Once an organization builds patient safety culture, it can leverage data and analytics to identify and profile risks for harm. Organizations can use analytics to first build reactive capabilities, then proactive capabilities, and, ultimately full integration of capabilities where safety tools integrate across workflows across the health system. The aim of integrated capabilities is scalable, sustained safety outcomes improvement. Applying machine learning to triggers will further strengthen their positive predictive value.
Pestotnik offered several use cases to show how data and analytics can improve safety. Among them were a program targeting heparin use, which resulted in 7 percent improvement in percentage of patients therapeutic in 24 hours and zero incidence of major bleeds and an opioid example, in which an organization used analytics to identify opioid use and reduce prescription volume by almost 1 million pills.
In conclusion, Pestotnik said that making safety a cultural priority (and backing it up with analytics and technology) benefits everyone—from patients and their loved ones to frontline clinicians and health systems.
Breakout Session: Wave 5
Session 34: How to Advance Beyond “Regular Data” with Text Analytics
Michael Dow, Director, Product Development, Health Catalyst, and Carolyn Simpkins, MD, PhD, Chief Medical Informatics Officer, Health Catalyst
Mike and Carolyn started their presentation with a clear definition of Text Analytics: “The process of deriving high-quality information from text by applying NLP to transform text into data for analysis.” Then, they shared why it’s so important: 80% of clinical data is locked away in unstructured notes. Walking through an example of how to find “LVEF” in a patient’s records to identify possible proactive interventions, they showed the precision and accuracy involved. They also shared the five components needed to use NLP with clinical text: Validation, Extraction, Context, Discovery & Synonyms, and Search Engine.
Session 35: The Population Health Template: A Roadmap to Drive Successful Health Improvement Initiatives (Educational Session)
Michael Kobernick, MD, MS, MS-PopH, CPE, Chief Medical Officer, SmartHealth, Ascension Health
During this hands-on session, participants learned about the importance of using a population health template to guide them in their improvement work. Sections in the template include (1) defining the health concern, (2) defining the current state, (3) determining the desired future state, (4) identifying the project plan and how to execute it, and (5) evaluating the program.
When organizations don’t use a tool like this, they run the risk of gaps and failures. For example, when Michael Kobernick, MD, and Chief Medical Officer of Ascension Health’s SmartHealth, was working with a wellness vendor, he asked them to provide specific examples of how they would demonstrate value from receiving health coaching. This was a several million dollar project, but the vendor was not able to identify the problem they were trying to solve, nor were they able to provide metrics that would show the project had any value.
After describing what not to do, Kobernick asked attendees to fill out the templates at their tables and give report outs on their findings. Then Kobernick provided the group with the following key takeaways:
- Be sure there’s a clear articulation of the health issue being addressed—this is missed in most
- It’s important to do a thorough assessment of the subpopulations in question by paying attention to any social determinants of health, disparities, and behavioral issues.
- Use pre- and post-measures of value that are defined in terms of the quintuple aim and are clearly related to the health issue in question.
Session 36: Perspectives from a CEO and Physician Executives: How Effective Governance Can Drive Sustained Improvements (Educational Session)
Cate Ranheim, MD – System Medical Director, Hospital Programs, UnityPoint Health System; and Tim Hobbs, MD, MBA, EVP – Chief Physician Officer, Community Health Network; and David Grauer, MHSA, MBA – Former CEO/Administrator, Intermountain Medical Center
David Grauer, MHSA, MBA, Senior Vice President, Professional Services, Health Catalyst and former CEO/Administrator, Intermountain Medical Center, presented on outcomes improvement government. Cate Ranheim, MD, System Medical Director, Hospital Programs, UnityPoint Health System (UPH), and Tim Hobbs, MD, MBA, EVP, Chief Physician Officer, Community Health Network (CHN) joined him to share best practices and experiences in outcomes improvement governance at their respective organizations.
The aim of outcomes improvement governance is leadership that benefits patients while also meeting organization goals in today’s value-based care environment. Grauer opened the session with the three-systems governance framework, which, done well, accelerates outcomes improvement. These three systems are best practice, analytics, adoption, and when fully leveraged, can help achieve outcomes improvement.
Grauer defined outcomes improvement governance as “the leadership, structure, process, and organizational culture needed to support sustained, system-wide outcomes improvement.” He reminded the audience not to confuse this with data governance—the details of managing data.
The outcomes improvement team, Grauer explained, consists of a leadership team, which oversees data governance committee, content and analytics team, and domain guidance team. They follow these must-have outcomes improvement governance principles:
- Stakeholder engagement: Starting at the top, engage all stakeholders around a common vision.
- Shared understanding: Have a common understanding of organizational need, capabilities, and readiness.
- Alignment: Adopt a consistent improvement methodology, align incentives, and balance polarities.
- Focus: Practice disciplined decision making to prioritize, fund, organize, and sustain initiatives.
Dr. Ranheim discussed her experience at UPH—specifically in the best practices of stakeholder engagement (leadership structure and committee composition and physician engagement) and focus (opportunity analysis and prioritization of projects). She also shared the outcomes improvement structure at UPH, which started with the clinical leadership group, then the Health Catalyst team, then the clinical service group, then work teams, and finally, regional implementation teams.
Dr. Hobbs spoke about his experience in best practices with shared understanding and alignment. For shared understanding, he underscored the importance of engagement with board members and developing a common definition and vision for quality. On alignment, he stressed conducting an organizational readiness assessment and ensuring aligned incentives, including a long-term incentive plan to create clear expectation of executive team.
Session 37: Getting Operational Leaders on Board to Deliver the Triple Aim (Case Study)
Lauren Anthony, MD, System Medical Director, Allina Health Clinical Laboratories
Allina Health believes that patients deserve to receive the optimum level of care through the use of a systemwide comprehensive blood conservation program. In 2011, facing a transfusion rate that was 25 to 40 percent above national benchmarks, the system launched a blood conservation program to reverse the trend. The initiative proved to be successful, resulting in $1M+ annual savings.
Key strategies that contributed to this success included:
- Initial benchmarking of blood product utilization by an outside consultant
- An educational splash with multiple sessions by national experts
- Formation of a system transfusion council to govern the work
- Revision of transfusion guidelines along with educational videos to support the new guidelines
- Implementation of a new order set with analytics decision support at order entry
If you are considering starting a similar program, Anthony says securing support from senior leadership is essential. Then, partner with physicians and nurses who already have an interest in reducing unnecessary transfusions. Having an engaged team championing the initiative is critical to success.
Session 38: Clinical and Financial Partnership Reduces Denials and Write-offs (Case Study)
David Wild, MD, Vice President, Lean Promotion, The University of Kansas Health System; and Colette Lasack, MBA, Vice President, Revenue Cycle, The University of Kansas Health System
No matter how awesome the clinical care in your hospital, it can all be undone if the patient doesn’t have a good experience with the revenue cycle process. If that’s bad, the patient will never recommend your hospital. In support of the patients it serves, University of Kansas Health System strives for the highest quality and shortest lead time in the revenue cycle process by eliminating all possible waste. This equals what the system defines as great patient financial care.
The patient is at center of everything University of Kansas Health System does. It says value is in the eyes of the patient. Its guiding formula is world-class patient outcomes + world-class patient experience delivered by competent committed and engaged staff = strong sustainable growth and financial performance.
In general, CMS denies 26 percent of all claims. Up to 40 percent of those claims are never resubmitted. Managed badly, this can cost a hospital tens of millions of dollars. This is why University of Kansas Health System strives to get it right the first time, a concept it refers to as “chasing zero.”
The need for revenue cycle improvements was evident in above average denials, challenges in effectively communicating with stakeholders, variation in processes, not leveraging available technology, and non-specific reporting. The challenge was getting to a 5 percent (of gross revenue) denial rate.
University of Kansas Health System recognized that, in order to effectively reduce revenue cycle and implement effective change, it needed to get upstream in the revenue cycle process. The system needed a change in organizational commitment. It focused on a Lean culture of continuous improvement rather than one-time tasks. It developed a framework that a team must reflect a long-term fixture rather than a short-term fix. It created more visible executive support. The organization made key investments in technology, analytics, reporting, people, and capacity.
In the end, University of Kansas Health System came away with a renewed commitment to continuous quality improvement that has enabled it to envision a new way to think about denials.
Keynote: “A Coalition of the Willing: Data-Driven Population Health and Complex Care Innovation in Low-Income Communities”
- Carolyn Simpkins, MD, PhD, Chief Medical Informatics Officer, Health Catalyst
- Kelly Craig: Chief Strategy & Information Officer, Camden Coalition of Healthcare Providers
- Maryann Vienneau: Program Director, iCMP and Palliative Care, Partners HealthCare
The HAS 17 documentary, “A Coalition of the Willing,” opened with Don Berwick (senior fellow at the Institute for Healthcare Improvement) explaining that the bulk of healthcare in the U.S. is consumed by small a percentage. Five percent of patients in the U.S. account for approximately 50 percent of the total cost of care. The U.S. spends so much on healthcare, but still isn’t getting the best quality. As a result, healthcare struggle with fragmentation, siloing of services, lack of services, and uncertainty.
The documentary looked at several care teams, all working in their communities to end the cost curve to make healthcare more humane. Each is leveraging data to identify patients who need help the most and ways to intervene. They employ an approach called “hot spotting” in which they use data to identify patients who need the most support and work to devote resources where they’re most needed.
In Camden, New Jersey, where a significant portion of the population is eligible for Medicaid and has complex healthcare needs, the Camden Coalition of Healthcare Providers is working with residents to understand how to help. The coalition’s goal is the make people feel part of the healthcare system, not victims of it. It aims to empower patient and care team to work together and coordinate better care. With hot spotting, the coalition targets people who need assistance the most—whether a result of poverty, drug addiction, or any situation the interferes with effective healthcare.
Health Quality Partners in Pennsylvania is working to improve care for the elderly with whole-person, whole-system approach. Nurses, data analyst, and outreach specialists use data-driven hot spotting to engage individuals in need and identify other issues that contribute to their healthcare challenges.
Larger systems are also discovering how to scale care management and maintain the real-time decisions and positive outcomes the Camden Coalition and Health Quality Partners have achieved. Common priorities are root causes, social determinants of health, and determining what people need versus what they’re actually getting. It’s consistent with the population health goal to keep beds empty and devote resources to keep people healthier outside the care setting.
These innovative care management programs define success per individual—a shift from previously looking to what healthcare leaders and organizations consider success.
Dan Burton began his closing remarks by reflecting on some of the many highlights of HAS 17. He announced attendee selection for favorite keynotes and breakout session. He also reflected on the fun (runs, walks, and the super hero party), which were also a meaningful part of the summit, as they allowed participants to connect and celebrate. He closed by asking those who sometimes feel overwhelmed to remember the work of healthcare outcomes improvement matters. With compassion, drive, and passion, healthcare transformation will succeed and be impactful.
Would you like to learn more about this topic? Here are some articles we suggest:
- 2017 Healthcare Analytics Summit: Wednesday Recap
- 2016 Healthcare Analytics Summit: Wednesday Recap
- 2016 Healthcare Analytics Summit: Thursday Recap
- 2015 Healthcare Analytics Summit: Wednesday Recap
- 2015 Healthcare Analytics Summit: Thursday Recap
- 2014 Healthcare Analytics Summit: Wednesday Recap
- 2014 Healthcare Analytics Summit: Thursday Recap