The 2018 Healthcare Analytics Summit: Thursday Recap
Paul Horstmeier opened day two of the 2018 Healthcare Analytics Summit™ (HAS) conference with a recap of day one, including some great photos from Wednesday’s beach party. There some great Hawaiian shirts, hula dancing, a limbo contest, and the return of “Super Dan.” Mr. Horstmeier shared some analytics from day one. Fittingly for this year’s theme, 46% of attendees shared their ideal vacation is to the beach while 2% prefer to stay home. He also shared numbers showing the great networking happening this year. Attendees went on 145 Braindates, 47 group Braindates, and posted 219 topics.
He also shared the breakdown of attendees, with the largest group (53%) identifying as IT and analytics employees. Thirty percent of attendees reported being a senior level in their organization, with 21% front line staff, 18% mid-level, and 14% executives.
Some more fun analytics from the recap:
- Attendees booked 145 individual Braindates.
- There were another 47 group Braindates.
- Braindates covered 219 topics.
- 21 percent of Braindate participants were frontline clinicians.
- 83 percent of HAS attendees think value-based reimbursement will improve care (a more optimistic outlook than 2017).
- Overall, HAS attendees prefer a beach vacation to visiting big cities or historical sites or staying home—in keeping with the HAS 18 beach party theme.
And, for those keeping track, the first day’s sessions ended at 5:29:47 pm (13 seconds to spare)
Penny Wheeler, MD
President and CEO, Allina Health
Penny Wheeler, MD, CEO of Allina Health, refers to herself as an accidental CEO. She started her career as an OBGYN, which she saw as a specialty that combines science and human understanding. She’s managed to bring that empathy into her subsequent roles as a Board member, Chief Clinical Officer, and finally CEO.
Dr. Wheeler presented on Allina’s commitment to whole person care—caring for each person according to their mind, body, spirit, and community. The whole person approach, she explained, improves care by addressing what matters to the patient. Whole patient insight enables the health system to provide the care the patient needs and minimize care that doesn’t align with patient goals to reduce waste. In this way, Dr. Wheeler said, better care is more affordable (less wasteful) care. Her health system has leveraged analytics to better understand the whole patient.
Addressing social determinants of health is paramount to the whole patient approach, and Allina has focused on five key points:
- Does the patient feel safe at home?
- Does the patient have enough food?
- Does the patient have transportation (including to medical appointments)?
- Does the patient have a home?
- Can the patient pay utilities?
Dr. Wheeler shared a video of an Allina customer story to show the impact of whole patient care and leveraging social determinants. A young mother of two responded that she didn’t meet the five points above. These insights allowed Allina to connect her with community resources to help with food and repair her washer and dryer—allowing her to save money on laundry and free up more time for her sons’ medical care.
Healthcare is moving fast, said Dr. Wheeler—a point she underscored with a slide that showed immense healthcare activity in just a week in December in 2017. This amount of activity, she explained, demonstrates that the public and industry knows healthcare can, and must, get better. In particular, Dr. Wheeler said, healthcare care can take the lead from other consumer industries by understanding what the patient really wants. “Never forget who and what you’re serving,” she said.
Dr. Wheeler described at a macro level what she sees as the main problem in healthcare right now: the tension between economic leadership and compassionate coverage for all. She sees quality improvement as the driver of cost containment. This is the bridge between economic pressure and compassionate care. She also sees powerful analytics as the key to these outcomes and provided some powerful examples from Allina Health.
- Almost 2 million fewer pills prescribed – an 8% reduction this year, 20% reduction over two years
- 6,000 patients fewer patients receiving opioid prescriptions for more than 20 pills – a 12% reduction
- 885 fewer patients receiving eight more opioid prescriptions over 12 months – an 11 % reduction
- 78% relative reduction in elective colorectal systemwide SSI
- 19% relative reduction systemwide length of stay, saving $90,000 in six months
- 4% compliance with preoperative and postoperative order set use
She also discussed a Whole Care pilot program which looks at five social determinants of health: housing instability, cost of utilities, food insecurity, interpersonal violence, and transportation issues. The margin for some people is so thin that if their utility bills are too high, they can’t pay for a preventive health appointment that might end up saving them hundreds or thousands of dollars down the road. This idea, that making care better and more impactful while eliminating waste, is the solution to the tension between economic pressure and compassionate care.
Keynote Session Digital Innovators
Scott Shreeve, MD
CEO, Crossover Health
Day Two keynote sessions featured a panel of Digital Innovators. The first panelist, Dr. Scott Shreeve, CEO of Crossover Health, dove right into the beach theme promising to teach attendees how to surf–to paddle out with some of the big tech companies entering healthcare and catch the wave of innovation.
Healthcare inflation is driving innovation in the employer space, said Dr. Shreeve. As employers pay more for healthcare than the raw materials of their trade and are still not getting sufficient care (access and quality are not good enough), they’re motivated to innovate ways to manage employee health and control cost.
Just as major companies from non-healthcare industries (e.g., Amazon, Google etc.) are disrupting healthcare with their technology and processes, employers are applying other business experience to healthcare. They’re becoming healthcare activists by working toward health plans with more value for the employer and the employee. Dr. Schreeve said that employers are finding complex practices historically organized by payers, with no focus on the patient. By engineering a better employee health plans centered on patients, Crossover Health has seen patient satisfaction rise to 96.3 percent, wait times decrease to less than five minutes for 89.1 percent of patients, and a reduction in total cost.
So why do big tech companies like Apple, Google, and Amazon want in on healthcare? They are employers with intimate knowledge of insuring employees, they’re used to disrupting industries, and the cost of care is too high. These companies also already know how to use supply chains and provide great customer experiences. This provides them with a roadmap for how to move into the healthcare space. Dr. Shreeve refers to these companies as Health Activist Employers.
Crossover Health partners with these Health Activist Employers to create a healthcare model with patients as the center. The idea is to surround patients with a team to take care of their needs, guide them through the process, and provide technologies and services that changes the way they can access their care. With a heavy emphasis on patient experience, Dr. Shreeve sees this as an entry point to quality of care. The model invites them to engage in their care in order to move from a transactional situation to a relationship. He sees a future of care delivery that starts with a platform of care and ties geographies together to create a care network.
Executive Director, Cleveland Clinic co-branded with Oscar Healt
Kevin Sears, Executive Director, Market and Network Services at the Cleveland Clinic, said that in 2015 the Cleveland Clinic was losing inpatient market share and had a compound annual growth rate -8.2 percent. The clinic found that narrow payer networks had come the Cleveland, as broad networks were diminishing across the country. To keep up, the Cleveland Clinic needed to work with these smaller networks. Leveraging experience managing its own employee health program, the Cleveland Clinic was able to create a community care program that bent the cost curve and saved $254 million.
When Mr. Sears and team redesigned the Cleveland Clinic community care program 2017, they knew cost containment and reduction weren’t sufficient; they had to redesign the care model to excel at population health management and standardize care. They engaged with a smaller insurance company (Oscar Health) to drive engagement with Oscar’s mobile engagement tools. With the program, the Cleveland Clinic could engage patients from day one and direct care to reduce costs.
In 2015, Cleveland Clinic was losing inpatient market share at a compound annual growth rate of -8.2%. They discovered they were underrepresented in the narrow networks that were popping up. Cleveland Clinic didn’t have network offerings, but they did have experience managing their own employee health plan. In 2010, they completely restructured their employee health plan and were able to “bend the curve” in order to save $254 from 2010-2017.
Next, they went to work on redesigning a Community Care Program. This team approach intended to manage the business within a different framework with different measures of success. They would be accountable for per member per month utilization rates rather than RVUs and other traditional measures.
They developed three strategic partnerships with three different insurance companies in different spaces. In the individual space, they partnered with Oscar Health. Kevin Sears said he is frequently asked why they made that decision. Oscar Health was losing money. What they saw in the company was a fully-integrated business-to-consumer technology platform. This created greater transparency of data, digital engagement for members, and drove net promoter scores. They had the ability to engage patients from day one through mobile interfaces and a dedicated concierge team.
This created an ability to direct care more effectively resulting in notable reductions in the cost of care. Today, six months into relationship, they’ve enrolled 10,000 members (three times what they had projected). Mr. Sears said these results are possible because of the incredible level of integration between the digital platform, Cleveland Clinic, their EMR, and other data platforms.
John Rogers, PhD
Professor, Biomedical Engineering and Medicine, Northwestern
Wearable biosensor technology as it exists today (e.g., FitBits) will soon be outdated, according to John Rogers, PhD, Director of the Northwestern University Center for Bio-Integrated Electronics. From an innovative clinical perspective, current sleek and modern watches, heartrate monitors, and the like have limitations in the future of biosensor technology. The next generation of wearables, he said, would be electronics that look like and interface with human tissue to better integrate electronics with body. Electronics for the human body are allowing a move away from bulky block electronics to something that looks more like biological tissue. This requires a reformulation of electronics to be compatible with the human body. Dr. Rogers explained how this concept of applying electronic membranes to conform to tissue is being applied to the brain, heart, or skin to measure various biological responses. Continued success in this realm will result in an incredible amount of data.
These ultrathin, skin-like electronic membranes interface with the brain, heart, skin, and more to produce clinical-grade measurement of physiological health (e.g., blood pressure and EKG), continuously and wirelessly transmitting that data to the patient’s healthcare team. Existing wearables are silicon wafers, which do not move with body tissue. They can also require strong adhesives and wiring, affecting comfort and ability for continuous wear and data transmission. Patients can also continue to easily wear the monitors after they leave the care setting, allowing for better care through ongoing monitoring.
Dr. Rogers showed two comparison photos of infants in the NICU—one heavily taped and wired, the other wearing only a few small, skin-toned patches. The latter baby was much less impacted by the skin-like sensors, able to move naturally and interact with her mother.
Inspiration for epidermal electronics came from an unlikely place – temporary tattoos. They’re thin, soft, and comfortable. The idea was to form skin-like electronics with those properties only with biosensor capabilities. These sensors are designed to laminate onto the skin’s surface in order to take clinical grade measurements of physiological health.
He shared early successes in this realm, specifically with neonatal intensive care patients and pediatric intensive care patients. EKG monitoring of NICU and PICU patients requires the use of heavy, bulky wires on the most medically fragile patients. This prohibits movement, interferes with patient’s ability to interact with the baby, and often leads to lifetime scarring. A study of 60 babies allowed researchers to monitor vital signs successfully while nurses monitored a graphical user interface. There was no indication the babies were even aware of the sensors.
Dr. Rogers discussed other ways biosensors are being used including monitoring tremors in Parkinson’s patients, cardiac stress monitoring in expecting patients, and monitoring stroke patients.
Breakout Sessions: Wave 3
Session 24 – How Automating and Virtualizing the Hell Out of Healthcare Is the Only Way to Save It (Strategy, Innovation)
Lyle Berkowitz, MD, FACP, FHIMSS
Chief Medical Officer and Executive Vice President, Product, MDLIVE
President, MDLIVE Medical Group
Healthcare will not continue to work in its current form, said Lyle Berkowitz, MD, FACP, FHIMSS, Chief Medical Officer and Executive Vice President, Product, MDLIVE, and President of MDLIVE Medical Group. In this failing healthcare system, 40 percent of Americans do not have a primary care physician (PCP), and 50 percent don’t get even the basic care they need due to access issues.
Saving the primary care system requires automation and virtualization, said Dr. Berkowitz. While the public tends to blame insufficient care on PCP shortages, he explained that there isn’t a PCP shortage; there’s a time shortage for PCPs (they don’t have enough time to do their jobs). Healthcare must rethink primary care, including rethinking the upside-down pyramid where care focuses on the top complex 5 percent of patients. Instead, healthcare needs to look at bottom majority (the less complex patients) and consider ways to automate their care. Dr. Berkowitz likened the transition to automated virtual care for less complex patients to how people use travel agents: they tend to do simple travel tasks themselves online (self service) and use an agent only for more complex travel planning.
Virtualization, said Dr. Berkowitz, is also critical to providing care to patients who can’t easily get into a PCP office due to distance, time, or cost. Automation makes virtualization scalable by allowing a doctor to take care of more patients in a significantly higher-quality manner.
Dr. Berkowitz projects that 80 percent of care can and should be done online, so that the 20 percent (complex patients) who really need an office visit have that access.
Health systems must do three analytics reports to automate virtualize and automate:
- The Burden Report
- The Convenience Versus Loyalty Report
- The Time Versus Complexity Report
Session 25: Toward Proactive, Predictive, and Personalized Care: How Startups Are Using Data Science to Build a Better Future for Healthcare
Chief Executive Officer, MATTER
While disruption has been a consistent theme of HAS 18, Steven Collens, CEO of MATTER, said he thinks the disruption in healthcare will happen in concert with big institutions. Large institutions have the wisdom of experience in the healthcare space, but usually lack the ability to innovate or pivot quickly. Entrepreneurs, however, are comfortable with risk and are able to iterate rapidly in concert with customers.
Between 2010 and 2017, the rate of acceleration for digital health financing was four times higher than venture capital overall. The healthcare system is ripe for innovation because of the explosion of data and technology. Mr. Collens discussed a number of exciting startups in the healthcare space in concert with two major healthcare trends: empowering people and equipping clinicians.
Startups that Help Empowering People
- Telesofia – enables patient education and engagement with personalized patient videos, accessible on any device.
- Tapcloud – customizes care plans for any condition and simplifies complex care into daily steps.
- Aiva – the world’s first voice-powered care assistant – hands-free communication for happier patients.
- HelloJoy – discovers patterns in symptoms and risk factors for depression using the sensors in a smartphone and AI
Startups that Help Equip Clinicians
- Quant HC – detects a patient’s clinical deterioration hours or even days in advance of current methods
- Ariel Precision Medicine – Integrates a patient’s symptoms and genetics with complex medical information. Intended to enable screening, diagnosis and monitoring of chronic disease progression.
- PhysIQ – The only FDA-cleared personalized analytics platform that learns a patient’s baseline and detects subtle changes to provide unprecedented insight.
- SonarMD – Delivers care management algorithms to help providers and patients dramatically improve the management of chronic diseases.
He also shared 5 ingredients needed for a culture of innovation:
- Imprimatur of the CEO
- Empowered innovation leaders
- Dedicated budget
- IT support
- Process for vetting solutions
Session 26: Increasing Capacity Without Construction: A Collaboration of Analytics and Frontline Operations
Chief Analytics Officer, Stanford Health Care
Rudy Arthofer, RN, BSN, MHA
Administrative Director, Hospital Operations Center, Stanford Health Care
Stanford Health Care is the #1 trauma center in the San Francisco and San Jose areas, with over 75,000 ED visits and 25,000 inpatient visits each year. In 2017, Rudy Arthofer, RN, BSN, MHA realized that the health system had a big problem: roughly 50% of the time there was a greater demand for beds than the hospital had available. To alleviate the issue, Stanford canceled surgeries, denied transfer requests and opened a mobile surgery unit, but despite these efforts, ED visits continued to increase. To tackle this issue, Arthofer decided to take action. The first step was to create a list of organizational mandates:
- No additional bricks and mortar
- No OR cancellations due to bed capacity
- Decrease ED boarding hours
- Maintain patient and staff satisfaction
- Increase transfer center acceptance
To achieve these mandates, Arthofer next created an interdepartmental leadership team that met (and still meets) daily to assess issues and propose solutions. In order to take things to the next level, Arthofer engaged Yohan Vetteh and his team to build a dashboard that allowed for the statistical analysis of historical capacity data. That dashboard now includes a complex forecasting model that helps Stanford’s teams to make capacity predictions across the next 12 months and plan for seasonal swings.
As a result of these efforts, Stanford Health Care was able to meet the organizational mandates including decreasing ED boarding hours by 24% and eliminating the need for the mobile surgery unit despite ED visits, OR volume and inpatient days increasing.
Session 27: Turning Data Analysts into Data Scientists
Chief Software Development Officer, Health Catalyst
Data Scientist, Health Catalyst
Evolved Artificial Intelligence (AI) resources now allow data analysts to become citizen data scientists, using methods formerly reserved for data scientists. Imran Qureshi and Taylor Larsen of Health Catalyst discussed use cases this approach and when to consult a data scientist for help.
The current top-down method of AI means hiring data scientists and training them to learn your organization, problems, and data. While this strategy works, it can be expensive and slow to produce. A majority (~80%) of most data scientists’ time is spent on data wrangling and cleaning, tasks that data analysts already perform. A complementary approach is the bottom-up method of AI, which leverages data analysts who already understand your organization, problems, and data, to solve problems using AI with data scientist consultation/supervision.
To determine whether an AI problem can be solved by a data analyst, look at whether the problem is simple and fits a known pattern. Complex problems are best reserved for dedicated data scientists; however, if the problem fits a bottom-up approach, the first step is to understand some common data science algorithms such as regression, classification, and clustering.
Qureshi walked through the creation of an AI spam filter for Hotmail based on rules to establish probability and then adjusting the threshold for either high sensitivity or high specificity depending on the goal of the organization.
When data analysts use data science methods, it’s important to look for indicators of problems with an AI model. The area-under-curve (AUC) value is one such indicator that shows you need help from a data scientist.
Session 28: Population Health Innovations Deliver Significant Cost Savings and Improved Health Outcomes
Executive Director, Crowfoot Village Family Practice
Dave Jackson, MBA
Chief Technology Officer, Airdrie & Area Health Co-op
This session showed how Alberta’s Crowfoot Village Family Practice (CVFP) provides world-class care through innovation—inspiring the nearby town of Airdrie to set a goal of becoming Canada’s healthiest community.
As one of Canada’s first medical homes, CVFP commits to giving patients same-day access, a consistent provider, and efficient multidisciplinary care. Shauna Thome explained that culture is a priority—physicians “share the halo” as part of the care team and time is protected for team collaboration. The clinic prioritizes quality improvement, communicating measures on a central scoreboard. Based on improving measures such as health screenings and diabetes interventions, CVFP patients have lower ED use and reduced lengths of stay compared to similar patients.
Dave Jackson explained that Airdrie saw the clinic’s example and asked, how do we expand this for the community? Based on two years of public engagement around the community’s strengths, needs, and health priorities, their plans include:
- Creating a Health Park—a hub to address siloed services and provide healthy senior housing
- Forming networks to guide health improvement efforts for people with similar needs
- Using technology to share data, provide answers, and create applications
While the project is in its early stages, health is now a regular discussion among stakeholders and in the community. In fact, Airdrie’s Mayor is now counting his steps—and wants to share his data.
Improvement depends on the building blocks—data itself doesn’t help unless you have the right conditions to act on it. It’s not just about the data; it’s about the people.
Session 29 – Integrating Clinical Improvement and Activity-Based Costing Identifies Pathway to Healthier Moms and Babies
Director, Women’s Health Service Line Finance Lead, UPMC
Beth Quinn, MSN, RNC-MNN
Program Director, Women’s Health Services, UPMC
Hyagriv Simhan, MD, MS,
Executive Vice Chair, Obstetrical Services, UPMC
Early detection and prevention of gestational diabetes mellitus (GDM)—hyperglycemia first noted during pregnancy—not only improves outcomes for mothers and babies during pregnancy and delivery, but also in the future. Fifty percent of women develop with GDM develop type II diabetes in the 10 years following pregnancy. Other consequences of GDM include high birth weight, hypoglycemia, jaundice, increased risk of C-section, and more.
Through the GDM initiative, a multidisciplinary group at UPMC developed a project plan and framework as a model for other projects and initiatives. The GDM initiative involved:
Mapping the patient pathway—the team developed a data logic map that pulled in finance, IT, and clinical data to “get everyone to speak the same language” and a process map that demonstrated intersections between process, milestones, and ideal state:
- Creating a stakeholder map—mapped by power/influence and interest in change/impact.
- Integrating clinical, financial, quality, and IT team members—an activity-based costing system enables UPMC to present true cost of care data to the team (e.g., they showed that GDM deliveries were $4,000 more per episode than non-GDM deliveries; as a result, they set a goal to reduce C-sections from 43% to 37% of total births).
- Setting process metrics—identified an opportunity to increase screening and diagnosis.
- Setting outcome metrics—neonatal hypoglycemia, macrosomia, neonatal jaundice, babies with 2+ conditions.
The early results of this initiative have been impressive—they identified best practices for lab ordering, identified areas of opportunity to improve patient engagement, developed best practice alerts, and more.
Breakout Sessions: Wave 4
Session: 30 – Machine Learning Marketplace (new to HAS 18) (AI, innovation)
In the Machine Learning Marketplace session, Paul Horstmeier, Chief Operating Officer for Health Catalyst, hosted a showcase of 12 innovative use cases for machine learning in healthcare. In the first hour, presenters gave five-minute overviews of their 12 stations and explained how they are using machine learning to address particular healthcare challenges.
During the session’s second hour segment, audience members walked around and visited all station and presenters. The presenters explained their machine learning projects more deeply, including results and key lessons learned, and answered attendee questions.
The 12 Machine Learning Marketplace sessions cover topics including using machine learning to assist in real-time inpatient care, detecting errors in medical data (comorbidities), optimizing no-show rates, determining risk modeling in falls, maintaining machine learning models after launching, predicting no-show patients, determining opioid risk, assessing readmission risks, and analyzing unstructured data, VOC data, and social media data to create actionable insights:
- Operationalizing Predictive Analytics Within Critical Care Environments (Baylor College of Medicine)
- A Natural Language Processing Toolkit for Healthcare (Sutter Health)
- Machine Learning in the Real World: Improving Accuracy of Readmission Risk Reporting Enabling Service Line Reporting (Pulse Heart Institute, Multicare)
- No-Show Forecasting (Dartmouth-Hitchcock Medical Center)
- Congratulations, It’s a Model! Now What? (Mission Health)
- Claims-Based Machine Learning Applied to Opioid Use Disorder in Oregon (Health Share of Oregon)
- Forecasting Inpatient Census for Operational Efficiency and Smarter Resource Allocation (Westchester Medical Center)
- Clinical No-Show (Sanford Health)
- Using Machine Learning to Detect Errors in Medical Data (Fresenius Medical Care North America)
- Risk Modeling for Falls in Value-Based Healthcare Using NLP and Other Advanced Methods (Acuitas Health)
- Use the Healthcare Data Operating System (DOS™) to Turn Data Analysts into Data Scientists (Health Catalyst)
- Shaping the Future of Healthcare with Machine Learning (Sanford Health)
Session 31 – Analysts Surf the Tsunami of Healthcare Data
Senior Vice President, Client Engagement, Health Catalyst
On a trip to Oahu in December 2017, John Wadsworth watched the best surfers in the world surf the North Shore for hours. As he watched them analyze the waves, he gleaned three surfing principles, which he also applied to analysts:
- Shifting environment. As a storm swell approached, John noticed the waves getting bigger; the surfers changed their approach. Analysts need to understand external factors that affect health systems (at-risk contracting, reimbursement models, evidence-based medicine)—and that understanding should inform their strategy.
- No, no, no, YES! A good surfer is patient. John thought, “this is a lot of waste” because the surfers committed to only the best waves. Analysts need to select only the best opportunities by shifting from reactive analytics to prescriptive analytics—root-cause analytics that explain the “why.”
- Get into position. As surfing great Gerry Lopez told John, “Surfing is 90% paddling. It’s gut-busting work.” To capitalize on the right opportunities, analysts need to identify the fundamental problems and pressures on the system—and find the opportunity that matters to the healthcare system. It may not be the one right in front of you.
- Top analysts pair these principles with critical skills—domain areas (healthcare operations, healthcare data), technical skills (data query, movement, modeling, analysis, visualization), and technology.
Organizations don’t hire analysts to run reports or manage tools. They hire analysts to solve problems. When analysts do that well, they touch the lives of tens of thousands of patients because they help clinicians deliver better care.
Session 32: The Data Maze Game: Navigating the Complexities of Data Governance
Thomas D. Burton, MBA
Co-founder and President, Professional Services, Health Catalyst
Senior Vice President, Professional Services, Health Catalyst
Do you see data as an expense—or as your organization’s most important strategic asset? Your answer may depend on the effectiveness of your data governance.
In this lively interactive session, Tom Burton and Mike Noke led participants through a collaborative game designed to illustrate how data governance can maximize the value of data, helping organizations surface and quantify opportunities and transform insight into better decisions—and massive improvement in health, cost, and experience outcomes.
In three game rounds, players moved tokens across their gameboards, racing to add “data sources,” access “self-service dashboards”, and achieve “improvements”—all under time pressure and without speaking to their teammates! In between rounds, Mr. Burton explained the recommendations that underlie the maze-game analogy:
- Combine qualitative (pain point) and data-driven analyses to identify opportunities for data governance.
- Charter a data governance committee with top-level participation and support; align with organizational and improvement leadership and strategies.
- Organize data governance subteams around your key clinical and business work processes (not the technology).
- Use the concept of the data life cycle—the flow of data from its capture, to integration, to access, insight, and action—to identify data issues across three dimensions:
- Quality: is data timely, accurate, and complete for purpose?
- Utilization: is data “right”? Are we delivering the right information to the right audience, at the right granularity, at the right time, with the right visualization/modality?
- Literacy: do people have the knowledge, skills, and inclination to use data for better decisions?
- Define a portfolio of data quality, utilization, and literacy projects representing a mix of effort/value and aiming at a range of clinical, financial, and operational improvements.
At the conclusion of the session, game-winning tables received prizes—and all participants received a copy of the Data Maze game and Health Catalyst’s Healthcare Data Governance: Improving Decisions from the Bedside to the Boardroom handbook.
Session 33: Addressing the Super-Utilizer Patient Challenge
Chief Strategy Officer, Community Care & Systems Chair, Super-Utilizer Care
The top 5 percent of patients, sometimes referred to as “super-utilizers,” account for more than 50 percent of total healthcare costs in the US. Scott Pingree, Chief Quality Officer at Intermountain Healthcare, discussed their journey to understand and engage this patient population in order to provide better care at lower costs.
In 2011, Intermountain began a quantitative assessment to find out who these super-utilizers were, what resources they were utilizing, and why the costs were so high. In a study of the top 1% of high-cost patients, they found that the median number of unique attending physicians was 12 per patient. This creates a huge challenge in care coordination that today’s healthcare model is not well-equipped for.
Today, Intermountain is working on a pilot project that extends care management and the medical home in order to provide whole person care and improve coordination of care for these patients. This Integrated Community Care model includes a super-utilizer care team made up of a medical director, nurse care manager, social work care manager, transitionist, and pharmacist. While the costs to provide this care model are more expensive in the short-term, they save money down the road (starting at about 24 months). The next phase of their journey will be partnering upstream to focus on prevention and adopting best care practices.
Key takeaways include:
- Understand complex patients from their perspective.
- Improved and more efficient care.
- Identify specific areas of opportunity.
- introduce pilots and interventions.
- Measure progress.
- Analytical process is a marathon, not a sprint.
Session 34: Transforming Emergency Care with Analytics and Technologies
Linda Hummel, MS, BSN, RN
Vice President, Quality and Patient Safety, Mission Health System
Rick Lee, MSN, RN, CEN, NE-BC
Executive Director, Emergency Services, Mission Health System
90% of hospitals report that they regularly operate at or above capacity, and Mission Hospital is no exception. The hospital was seeing 3.5% of patients leaving the ED without being seen by a provider, and another 3.5% leaving before completing treatment — leading to decreased patient satisfaction and thousands of dollars in unrealized revenue — when they decided that something needed to change.
To begin tackling this situation, Rick Lee, MSN, RN, CEN, NE-BC, initiated daily huddles with the ED practice teams at Mission Hospital in which they assessed and implemented several projects. Unfortunately, implementing all of the projects at the same time led to confusion about what exactly was driving results. Linda Hummel, MS, BSN, RN, jumped at the opportunity to dive in and help the team focus their efforts and develop a means for measuring outcomes.
The result was a retrospective dashboard that reported on the ED results of the past 24 hours, including previously unknown data points such as what areas of the hospital had the longest wait times. This information was then put directly into the hands of hospital leaders so that they were aware of and held accountable to the data and its impact on patient throughput.
In closing, Lee recounted some of the results that were realized:
- Threefold improvement in patient ranking for overall quality of care and provider communication
- 70% relative reduction in time to complete registration
- 89% relative reduction in the rate of patients who left without being seen
Session 35: Innovative Analytics: Using Analytics to Evaluate Emerging Technologies
Manager, CV Clinical Programs/Services, Allina Health
Senior Analytics Engineer, Health Catalyst at Allina Health
Clinical Program Director, Cardiovascular Clinical Service Line, Allina Health
Steven M. Bradley, MD, MPH
General Cardiologist, Minneapolis Hearth Institute
Healthcare tech is exploding—around 1600 cardiovascular devices gained pre-market approval in 2017. Some devices will be cost-effective and help patients, while others will increase cost without improving outcomes. In this session, an Allina Health team described their success in answering the question: which technologies will make a difference?
The key is partnering with providers, supply chain, and analysts to use data to identify and address unnecessary variation—keeping in mind the difference between efficacy (a device works in ideal conditions) and effectiveness (it works in the real world). Steven Bradley explained that even though cardiologists usually “love shiny new toys,” if the data shows a device doesn’t improve outcomes, providers will change their practice.
For example, Dr. Bradley described how Allina evaluated the use of higher-cost sternal plates after surgery rather than standard wire closures. Use of this technology was increasing, with variation between surgeons. While the FDA said sternal plates might bring benefit based on bone scans, the data showed no clinical benefit at Allina. After the team discussed the data with clinicians, sternal plate use decreased to 2%.
Using analytics to evaluate clinical outcomes for new technology—and transparent discussion with providers—has helped Allina ensure technical innovations are used responsibly.
- With new technology, identifying the data can be challenging—expect some detective work.
- Be transparent about why you’re evaluating new technology—bring providers into the discussion.
- Look to experts in your system to help design projects, generate ideas, and analyze variation.
Breakout Sessions: Wave 5
Session 36: Data, Insights, Action! Little-Known Principles and Skills for Making Analytics Actionable
Senior Vice President, Professional Services, Health Catalyst
Russ Staheli sees the value of data to healthcare much like the value of water to the farm he worked as a child. Used correctly, it’s a critical life-sustaining factor; used incorrectly, it’s useless or even damaging.
The flood of data in healthcare has little value without insights and action. Insights from raw data generate excitement on the part of clinicians and leaders to make change happen. And acting on the insights leads to actual improvement value.
In Staheli’s role leading around 130 analytics professionals to catalyze improvement at healthcare systems across the nation, he has found three principles to be key in making analytics actionable. First, analytics professional who have “generalist” “mile-wide-and-foot-deep” skill sets are more likely to create insights and action fast. Second, staying highly aligned with (strong principles) and loosely coupled (with minimal bureaucracy), keeps teams generating fast value with the big picture in mind. Third, a balanced investment approach with equal emphasis on both insights and analytics ensures that professionals stay passionate while creating improvement and ROI value.
Further, analytics professionals who generate the most actionable insights, tend to share seven essential skills. These include the ability to:
- Present and interact dynamically
- Create insights
- Present insights in a compelling way
- Understand healthcare data sets and sources
- Read and write SQL
- Model data
- Leverage visualization tools
Without these core skills, more advanced, future skills such as AI and machine learning are meaningless.
Session 37: Reducing Unwarranted Clinical Variation Saves Tens of Millions of Dollars
Finance Manager, Allina Health
Director, Analytics, Health Catalyst at Allina Health
Unwarranted variation in clinical care is costly, representing as much as $30 million of actionable savings opportunity per $1 billion in revenue for a typical organization. Allina Health has made reducing this variation a focus in order to improve financial margins. Matthew Brown and Sarah Jenson presented the step-by-step approach their financial and clinical analyst teams used to both reduce variation and save money. Their partnership and standardized approach helped them identify and prioritize clinical improvement initiatives with the highest projected ROI.
Before undertaking this type of project, Matthew Brown provided the following tips to help prepare for success:
- Improving quality requires executive and board leadership.
- Standardizing work in a large organization with limited resources is challenging.
- Improvement work must be prioritized across the system.
An audience poll revealed that 44% of respondents said their organization was just starting the process to identify clinical variation. Luckily, Sarah Jenson was there to provide her step-by-step approach that included specifics of their quarterly opportunity analysis, analysis approach, and graphical interpretations used to present their data.
Once the analysts present their report, the finance team asks a number of questions to predict the impact to the bottom line:
- How is the care model changing? What steps are we taking to drive results? How can we track those steps?
- What changes in care will affect out financials?
- How much of this population is in our ACO/Shared Savings Population?
- Could this project be eligible for a pay for performance (P4P) program?
Both Matthew and Sarah provided attendees with specific instructions from the beginning to end of their process and walked them through a detailed example of their spine surgical care model and the PPCs impact on Pay-for-Performance incentives.
Session 38: Proactive Patient and Leadership Engagement Delivers an Improved Care Experience
Director of Patient Satisfaction, Wolff Center at UPMC
MPA, Senior Product Manager, UPMC Enterprises
In 2017, University of Pittsburgh Medical Center (UPMC) leadership gathered to determine the organization’s priorities and landed on four key pillars:
It’s no mistake that service and people top that list, said Amy Triola. UPMC believed strongly that happy employees = happy patients, and vice versa, but unfortunately, the hospital had been ranked in the 28th percentile in the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) for the preceding year. This indicated to hospital leadership that something needed to change.
UPMC created an ambitious cascading goal system (where goals start with executive leadership and trickle down through each department to frontline employees) that aligned the priorities of those working in all levels of the organization. But what are goals without a means to measure them? Jody Madala, MPA worked with Health Catalyst to develop dashboards that reported on UPMC metrics and helped to support accountability.
To sustain positive progress over time, UPMC created accountability teams that met regularly with leadership and individuals across all facets of the organization to help keep them on track. The mantra became: stay focused on patients and employees and the rest will follow, and this has proved true. As a result of UPMC’s efforts, they were able to raise their HCAHPS percentile by 13 points within 6 months, while increasing employee engagement among hospital workers by 33% in the same period.
Through a series of evocative, funny, and heart-wrenching clips from both Hollywood and her real-life adventures, Kim Goodsell took the audience on an afternoon journey about the “patient of the future.”
Author of Gaming Aging in the ‘Omic’ Era, Kim Goodsell has been lauded for self-diagnosing a previously unknown link between two rare genetic disorders. She’s been credited by Dr. Eric Topol, one of the world’s foremost biomedical scientists, as “marking an important watershed moment in medicine made possible by the proliferation of digital technology and the online democratization of data.”
The birth of the “citizen scientist” is the empowered patient the medical community has been waiting for. However, clinicians may not be ready for Mrs. Goodsell’s specific brand of tenacity. In fact, when she asked that her geneticist sequence her LMNA gene to see if it was the culprit she suspected, her request was not well received.
Backed by video clips including the likes of “Alison in Wonderland” and “Monty Python and the Holy Grail”, rare disease patient Kim Goodsell invited the audience to “remember the future” healthcare and consider the “great inversion” (power of democratization of healthcare information) in medicine today. As a digital patient pioneer (not an MD or PhD), Ms. Goodsell inverted her own healthcare experience when she made own rare disease diagnoses based on personal internet research. She petitioned her care team at the Mayo clinic to sequence her LMNA gene, which she suspected was involved in her rare heart condition diagnosis. Her clinicians said it was not likely that the LMNA gene played a role in both disorders. Later, they called Ms. Goodsell back to tell her she had correctly linked her rare disease to the LMNA gene.
After 30-some years of an adventure lifestyle in which Ms. Goodsell didn’t depend on or prioritize digital technology (often without cell phone or internet access), she was now leveraging technology to understand and manager her health. She is also living with an implantable cardioverter-defibrillator (ICD), a high-technology device to manage her heart rate.
Years after her heart condition diagnosis, Ms. Goodsell again delved into her own research when she received a second rare disease diagnosis—a degenerative musculoskeletal condition. She was determined to find a link between her condition, which she did with her research.
Widely considered an anomaly for her persistence in research, Ms. Goodsell argues that she is in fact an average patient, or at least the average patient of future. She insists that all patients can, and should, access the data they need to understand their health and advocate for themselves. Ms. Goodsell now advocates for greater patient access to their real time data, such as her ICD diagnostics, as this insight can help them avoid risk.
Despite everything, she told the audience she has the deepest gratitude for the doctors, data, and the devices that have kept her alive.
CEO, Health Catalyst
Dan Burton began his closing remarks by reflecting on some of the many highlights of HAS 18, from new additions such as Braindate and the Digital Innovation Showcase, to returning favorites like the Analytics Walkabout. In the fifth year of this conference, both keynote speeches and breakout sessions received the highest scores ever. Top recognition went to Dr. Robert Wachter, Marc Randloph, and Kim Goodsell, among others. Health Catalyst’s John Wadsworth repeated his 2017 honor of most popular breakout session, while many others received top marks.
He also reflected on all the fun attendees had, connecting on runs, walks, and at the beach party. He closed by reminding everyone that transformation is possible when you appeal to people’s better selves and thanking attendees for being a catalyst for change.
Mr. Burton closed with a clip from the film “Moana” to offer a final learning on leadership and the choice between maintaining the status quo or searching for and trying new solutions to challenges. The movie, he said, showed the power of believing in a higher ideal and assuming positive intent.
Would you like to learn more about this topic? Here are some articles we suggest:
- 2018: The 2018 Healthcare Analytics Summit: Wednesday Recap
- 2017: Healthcare Analytics Summit 2017 Kicks Off: Wednesday Recap and Thursday Recap: 2016 Healthcare Analytics Summit™ Finale
- 2016: Wednesday Recap: 2016 Healthcare Analytics Summit™ Kicks Off and Thursday Recap: 2016 Healthcare Analytics Summit™ Finale
- 2015: 2015 Healthcare Analytics Summit™: Wednesday Recap and 2015 Healthcare Analytics Summit™: Thursday Recap
- 2014: Healthcare Analytics Summit™ Day One Recap and Healthcare Analytics Summit™ Day Two Recap