Healthcare Analytics Summit 2020: Day Three Recap

HAS Day 3 RecapPaul Horstmeier, chief operating officer for Health Catalyst, welcomed attendees to the third and final day of the Healthcare Analytics Summit 20 Virtual (HAS 20 Virtual) with a quick preview of the day’s agenda, including two keynote presentations and two waves of breakout sessions. He then turned it over the HAS 20 Virtual data analysts, who shared engagement data from the previous day’s events—the Analytics Walkabout, Machine Learning Marketplace, and Digital Innovation Showcase.

Next up, Health Catalyst chief data scientist Jason Jones gave a brief insight into segmentation in data analytics, using HAS 20 Virtual survey data as an example. Jones demonstrated that while he could parse conference attendees by age, profession, personal interests, and more, those segments still didn’t create the full picture of each attendee. His message was that data can aid healthcare decision making, but data users still need to include the immeasurable individual human factor when they use analytics.

Before Horstmeier introduced the day’s first keynote speaker (Vice Admiral Raquel Bono, MD), he revealed the finalists for Socks of HAS and invited participants to vote for their favorite. He also announced that HAS Game leaders were in a tight race, and he’d announce the winner during the day’s closing session.

Keynote Session #20

Vice Admiral Raquel Bono, MD – Former CEO and Director of the U.S. Defense Health Agency, Head of Washington State’s COVID-19 Healthcare Response Team

In her keynote presentation, Vice Admiral Raquel Bono, MD, shared the roots of her interest in medicine. During late nights waiting for her physician father to return home, she determined that she’d one day work in the hospital so she could see more of him. Though the future Dr. Bono didn’t initially think girls could become doctors (she assumed she’d aim for another hospital role), once her father told her she could, she began her path to becoming a surgeon. In addition to medicine, Dr. Bono followed her family legacy to the Navy, as both her father and grandfather had served in the military, establishing a career path dedicated to service to her country through healthcare.

In the Navy, Dr. Bono was part of a small medical core that worked to keep soldiers mentally and physically prepared to defend the United States. That value—helping people act in support of the country—resonated with her and has guided her career, including earning a three-star rank. She took lessons from her experiences in deployment (such as the first Gulf War) and interactions with the healthcare system as a surgeon and patient to understand how the U.S. health system could be better, namely by adopting a more patient-centered focus.

Dr. Bono retired from military life in 2019 and subsequently led Washington state’s COVID-19 healthcare response team. She brought her military experience to Washington, particularly insight from pandemic scenario wargames, using this understanding to influence data-informed decisions, such as immediately suspending elective care throughout the state. She knew the importance of viewing the emergency through a collective lens of all state healthcare resources to accommodate patient surges, allocate resources, and develop a statewide transportation plan to move patients to available care.

Keynote Session #21

Michael Dowling – President and Chief Executive Officer, Northwell Health
Brent James, MD, MStat

Clinical Professor at the Clinical Excellence Research Center (CERC), Department of Medicine, Stanford University School of Medicine

In this special fireside chat, Brent James, MD, sat down (virtually) with Michael Dowling, President and CEO of Northwell Health. Northwell is the largest hospital system in New York state and one of the largest health systems in the country but started as a single hospital–Northshore University Hospital on Long Island. After 21 mergers, Northwell Health now includes 23 hospitals, 800 ambulatory care facilities, and more than 70,000 employees.

Michael Dowling similarly started from humble beginnings. Growing up in Northern Ireland, he left home at the age of 16 and worked manual labor jobs to put himself through school and eventually college. He moved to the U.S. and worked on the docks on the west side of Manhattan while attending Fordham University. He went on to teach at Fordham, served as the director of Health, Education, and Human Services for New York State, and served as senior vice president at Empire Blue Cross/Blue Shield before joining Northwell Health in 2002.

In this fireside chat, Dr. Brent James asked Mr. Dowling about his journey and the competitive spirit that has defined his life and career. A lesser-known fact about his past is that he was a champion athlete in Ireland’s national sport: hurling. Picture competitive hockey played on a field without any helmets and pads, and that will provide some idea about Ireland’s favorite, and very physical, national sport.

Dr. James also asked Mr. Dowling to share insights from leading the charge against COVID-19 from the epicenter of the outbreak. The health system saw 76,000 COVID-19 patients and had 17,000 inpatients. He described the past several months as a humbling experience that’s changed his perspective forever. What impressed him the most was the dedication, courage, and compassion of the staff. Lastly, Mr. Dowling discussed the importance of innovation, the responsibility to dismantle inequities to provide access to care, and the necessity of creating a culture of continuous learning to redefine health, healthcare, and quality.

Breakout Sessions – Wave 3

Session 22 – Capacity Planning and Leadership Response: A COVID-19 Silver Lining

John Hansmann – SVP of Strategic Consulting Operations, Health Catalyst
Jason Jones – Chief Data Scientist, Health Catalyst

This session gave a behind-the-scenes view of how an analytics organization moved to address a current pressing need for its partner systems: the use of data and analytics to forecast COVID-19-driven capacity and resource issues and inform leadership response.

Hansmann and Jones described how they first identified a likely sequence of short- and longer-term needs (e.g., adjust to meet overwhelming demand, find post-acute and at-home options, prepare for additional waves of infection) and reviewed the data and tools needed to address them. Early in the pandemic, this work led to the development of a Capacity Planning Tool (March 2020) to estimate needs and buy leaders time to create plans to address them. As they continue to collect data and observe regional variations, the team deploys different statistical methodologies to support interpretation—to visualize different projected scenarios and to “separate signal from noise.”

The thinking that drove the development of the Capacity Planning Tool is instructive for organizations using data to guide contingency plans. Hansmann and Jones emphasized that leaders must do the following:

  • Determine the decision window—how far in the future you want to forecast, for what length of time.
  • Implement systems to watch your back.
  • Choose tools to fit your purpose.

Jones also noted that the challenge of COVID-19 has had a silver lining for data scientists, as they now have increased ability to deliver critical insights, common understandings, and decision support for leadership in times of uncertainty.

Session 23 – Navigating the Post-COVID World Through Data Science

Imran Qureshi – Chief Data Science Officer, Clarify Health

COVID-19 has increased care disruptions for patients—they are seeing their doctors less frequently, avoiding hospitals, and losing insurance. The session started with a question that Qureshi would answer throughout the presentation: “How do we identify the patients who will be most affected by these care disruptions and choose the best intervention for each patient?”

Throughout the presentation, Qureshi illustrated how to determine methods to accomplish the following:

  • Identify patients at the highest risk for long-term adverse outcomes caused by care disruptions.
  • Evaluate physicians in 2020.
  • Predict risk and utilization for the near future.

Although the future is unpredictable, Qureshi left the audience with valuable ways to help predict and plan:

  • Care disruptions have happened in the past—not because of COVID-19 or affecting this many patients—and we can learn from the past.
  • Care disruptions due to COVID-19 will have an impact on the outcomes for patients.
  • Data scientists can train models on these care disruptions and predict for COVID-19.
  • Data scientists can use What-If Modeling to provide visibility into the future.

Session 24 – “Bellagio Buffet” Analytics: Necessary for COVID Recovery

John Wadsworth, MS – Senior Vice President, Client Engagement, Health Catalyst

John Wadsworth opened by sharing a handful of observations he has witnessed during the COVID-19 pandemic. Amidst all the furloughs and job cuts, he has seen that analysts with best with job security have two qualities: 1) a deep understanding of the business of care delivery and 2) technical savviness.

Throughout the session, Wadsworth drew parallels between healthcare analytics and the Bellagio Buffet to highlight three essential principles:

  • The business of the buffet and healthcare analytics are very similar—they must understand the varied needs of clientele, appeal to the masses coming through the casino/healthcare system and generate revenue.
  • The buffet and healthcare analytics both have certain assumptions built into the model. For a buffet, hold a plate with food as they wander, use utensils to feed themselves, etc. For healthcare analytics, competently interpret the information provided, take appropriate action based on analytics, etc.
  • To overcome common challenges, the buffet and healthcare analytics need to improve the menu, change the diet, and teach people to eat…better.

Wadsworth ended by suggesting a few simple analytic learnings:

  • Spend less than 2 percent of the analytic budget on benchmarking.
  • Outsource regulatory analytics.
  • Invest in analytic talent capable of producing risk-based analytics (technical and care delivery business savvy).
  • Hire a partner for data literacy for your organization—this is not an IT/informatics competency.
  • Get a partner for data-driven strategic consulting on population health, CIN, ACO, and new revenue streams.

Session 25 – Hypertensive to Healthy: Keeping Patients Safe with Remote Monitoring

Hyagriv Simhan, MD, MS – Director of Clinical Innovation for the Women’s Health Service Line, Executive Vice Chair for Obstetrical Services at Magee-Women’s Hospital, UPMC
Beth Quinn, MSN, RNC-MNN – Director, Women’s Health Operations, UPMC

Hypertension is the leading cause of postpartum hospital readmissions and among the biggest contributors to maternal morbidity and mortality. The Women’s Health service line at the University of Pittsburgh Medical Center (UPMC) used its service-line structure and interdisciplinary care team approach to identify and manage postpartum hypertension and improve maternal health outcomes.

UPMC initiated a program called Hatch—an interdisciplinary group of expert providers and experienced problem solvers—to champion clinical innovation in technology, operations, and payment strategies. These efforts, which the COVID-19 pandemic accelerated—focused on three key areas:

  • Video visits: In the first month alone, over 200 providers delivered more than 1,800 video visits to more than 1,600 unique pregnant patients.
  • Standardized patient-centered care pathways: Standard pathways use key milestones to drive length-of-stay timeline and safe plans of care during follow-up.
  • Remote monitoring: Eligible patients receive texts daily for 14 days, and call center nurses monitor values and follow an algorithm to guide further patient outreach and medical management. To overcome barriers of insurance coverage of blood pressure cuffs, UPMC was able to distribute over 6,000 blood pressure cuffs to participants.

Key to success was a clinical analytics dashboard that allowed UPMC to monitor patient characteristics and results and adapt processes based on data. Results so far are promising: length of stay, readmission rates, and ED visits are lower; 94 to 96 percent of enrolled patients are highly satisfied with their care; and postpartum check-ups are 15 to 20 percent above the national average. This model of care will become the new normal beyond the COVID-19 pandemic.

Session 26 – Advanced Analytics for Medical Practices: Value-Based Care in the New Normal

Keegan Bailey, MS – Senior Leader, Acuitas Health
Andy Choens, MS – Leader and Architect, Data Science and Engineering, Acuitas Health

Many primary care providers saw their work change drastically in March 2020 when communities closed to minimize the spread of COVID-19. In a matter of days, providers shifted from delivering care in person to delivering care virtually via telemedicine. Arrived appointments (an essential metric for independent physicians) fell along with revenue. Providers had to adapt quickly to survive, and they needed new analytics tools to do it. Enter Keegan Bailey and Andy Choens from Acuitas Health.

To fill analytics gaps for primary care providers, Bailey and Choens drew on expertise across multiple healthcare domains: clinical, engineering, statistical, and operational.

“It’s not enough to have the traditional healthcare organization silos working on these things,” Bailey said, “You really need some cross-pollination between teams.”

Some of the tools they developed include the following:

  • A readmission risk tool that uses machine learning to predict patients with a low and high risk of readmission.
  • A trend tool that tracks virtual and in-person appointment trends (clearly showing the sudden drop in arrived visits (In New York, primary care providers lost more than 5,000 appointments from one week to the next).
  • A pre-visit planning tool to help providers cope with the burdens of documenting, coding, and quality measures reporting that can take up to 90 percent of their time.

While some of these tools may sound basic to large organizations with in-house analytics teams, they are critical for independent providers who don’t have dedicated analysts, Bailey and Choens said.

Breakout Sessions – Wave 4

Session 27 – Introducing the Data Science Adoption Model™: Realizing the Value of Your Investment

Jason Jones – Chief Data Scientist, Health Catalyst

What framework can guide organizations as they strive to realize the full potential and value of analytics to guide decisions? In this session, Jones reviewed several well-known models (e.g., Gartner’s ascendency model, the Davenport, and HIMSS models) that outline the phases leading from descriptive to prescriptive analytics and personalized medicine. Jones addressed the strengths of these models but also pointed out gaps. One principle “miss” is the lack of practical guidance for leveraging data science—and it’s this gap that the new Data Science Adoption Model seeks to address.

The Data Science Adoption Model outlines five non-sequential levels that range from guided analytics that self-serve analytics can usually deliver, to more complex capabilities that typically require significant expertise. Jones uses real-life example questions to explain the data science work that takes place at each level. How can we demonstrate the potential to improve heart failure readmission rate? What should we do to reduce inpatient admissions? Do higher-volume providers deliver better outcomes? The model accounts for the complexity and challenge of correctly using data to generate insights for improvement but also lights the way forward, proving that sophisticated analytics are achievable, broadly applicable, and extremely valuable for healthcare.

Session 28 – A National Analytics Strategy and Use Case: Closing the Gaps in Data to Address a Pandemic

Gabriel Brat, MD, MPH – Assistant Professor of Surgery, Beth Israel Deaconess Medical Center, and Instructor in Biomedical Informatics, Harvard Medical School
Sadiqa Mahmood, DDS, MPH – General Manager & Senior Vice President, Life Sciences Business, Health Catalyst

The United States must have a high-value national COVID-19 data asset to guide policy, population health programs, vaccine development, and outbreak management. With that imperative in mind, Sadiqa Mahmood, DDS, MPH, presented on the fundamentals of a national COVID-19 registry, potential use cases for that data set, and field observations during the pandemic about data performance and needs.

Dr. Mahmood explained that COVID-19 data has been siloed in sources across the United States, preventing population-level understanding. Furthermore, varying local standards and a lack of a common COVID-19 definition has generated data quality issues, leaving analysts with database clutter (the likes of, “patient left for Florida”). Dr. Mahmood explained that use cases for national-level, high-value COVID-19 data spanned disease surveillance, the natural history of the disease, diagnostic testing, clinical trials, drug efficacy and safety, impact on other diseases that will persist beyond pandemic, and more. A national registry must meet these use case needs with scalable, reusable common data models, an ecosystem of stakeholders, information exchange, a national patient identifier, and common nomenclature.

Gabriel Brat, MD, MPH, joined Dr. Mahmood to present a use case for the national COVID-19 registry. He and colleagues at Beth Israel Deaconess Medical Center leveraged registry data to evaluate the effects of anticoagulation on COVID-19. Autopsies, Dr. Brat explained, have shown blood clots instead of severe lung inflammation in deceased patients with COVID-19, prompting the question, “If individuals were dying of blood clots, could anticoagulants have a therapeutic effect on COVID-19?” Using machine learning models to identify confounders, Dr. Brat and team evaluated both anticoagulant and antiplatelet use among the registry data. They found that those on antiplatelets had a lower risk of developing COVID-19. Though Dr. Brat admitted this particular finding has major caveats, he hopes to reproduce the model for using registry data across the care continuum.

Session 29 – Virtual Care in the COVID-19 Era: Enabled with Enterprisewide Analytics
Patrick McGill, MD – Executive Vice President, Chief Analytics Officer, Community Health Network

COVID-19 initially drove a decline in outpatient visits of 60 percent, due primarily to a CMS call to cancel nonurgent visits, as well as provider and public uncertainty and fear of the virus. Though many healthcare organizations rapidly transitioned ambulatory visits to virtual care (e.g., telehealth), few systems were prepared for the sudden change, as outdated rules and regulations had historically made telehealth difficult.

Like its counterparts across the country, Community Health Network (CHNw) made the sudden pivot to virtual care. And like other health systems, CHNw didn’t have the technology or policies in place to execute a pandemic-ready virtual care strategy. To meet the urgent demand for virtual care and drive practice management, the Indiana health system leveraged enterprisewide analytics to measure the outcomes, patient experience, and value of virtual care and determine how to make this new model successful. Understanding that virtual care would redefine care standards, pathways, and competition, CNHw developed scenario analysis tools to project revenue, reimbursement, staffing, and workflows, as well as the impact on different populations.

With this enterprisewide-analytics-driven approach, CHNw deployed COVID-19 virtual visit analytics in just one week. Virtual visits now comprise more than 80 percent of its visits, and it’s maintained 75 percent of ambulatory outpatient revenue. The CHNw experience shows, that as telehealth remains an essential care pathway in a pandemic-ready future, enterprisewide analytics will be critical in understanding how virtual care fits into organizational strategy (e.g., operations and staffing) and where to invest in related technology.

Session 30 – Epicenter of the Pandemic: Driving Transformation at Northwell

Chris Hutchins – Vice President, Chief Data and Analytics Officer, Northwell Health

In response to the public health crisis, the use of data to support and enable preparation and response activities was critical in supporting response teams and efforts. Chris Hutchins shares the story of Northwell Health—New York’s largest healthcare provider and the hardest hit healthcare system in the United States. Northwell Health has treated 17,000 patients and counting (the most in the United States) and added 200 beds daily at the peak of the pandemic.

To drive transformation and navigate through the pandemic, Hutchins and the healthcare system did the following:

  • Organized clinical and technical staff to address clinical reporting needs.
  • Established a regular meeting cadence.
  • Leveraged organizational structure and technology to facilitate open collaboration.
  • Identified cohort parameters (always iterate and communicate).
  • Worked with executive leaders to identify organizational priorities to inform the development of an enterprise roadmap for data integration.
  • Capitalized on the enterprise data warehouse and an enterprise platform to support self-service analytics—enabling verticals to accelerate learning and deepen capabilities.

Session 31 – The Analytics Emergency: Rapid Deployment of a Real-Time, Analytics-Enabled Incident Command

Michael Houck, MS – Healthcare Data Scientist, Analytics, Albany Medical Center

Michael Houck shared the story of Albany Medical Center during an event it didn’t predict—an analytics emergency. The team quickly established an incident command center at the onset of the pandemic with representatives from the major hospital areas. At the end of the first phase, the team had a live view into all tracked patients, an epidemiologist consistently logging data into one source, a tent callback application, and dashboards containing all required data. The team needed all the data available at a moment’s notice because they never knew when the next state or government reporting requirement would come and couldn’t lose time chasing paper files.

Houck shared some recommendations that proved invaluable to the team and the health system:

  • Real-time communication is critical.
  • Automate connections when possible, audit everything manually.
  • Analytics are crucial—from the beginning and at every step of the process.
  • Nothing is perfect (especially done overnight)—there’s always an opportunity to improve.
  • Collect data whenever possible.

Houck ended with a quote from Rahm Emanuel, the former mayor of Chicago: “You never want a serious crisis to go to waste. And what I mean by that is an opportunity to do things that you think you could not do before.” Albany Medical Center faced a huge crisis and did not let it go to waste; they were able to advance significant parts of their analytics and overall hospital capability in just a few months.

Session 32 – Prioritizing the Patient in Value-Based Care: A Data-Informed Approach

Tyler Gauthier, MHA, CPHQ, CSM – Director of Value-Based Care, OneCare Vermont
Katelyn Muir, CPQH, CSM, MPH – Supervisor of Population Health Analytics, OneCare Vermont

As the ACO in the State of Vermont, OneCare had a vision for statewide care coordination strengthened in partnership with community care stakeholders. Their population health goals went beyond improving outcomes and decreasing costs to focusing future programs on prevention and wellness.

Tyler Gauthier and Katie Muir shared how the OneCare Vermont healthcare system achieved the following:

  • A 33 percent relative reduction in emergency department utilization per 1,000 members per year among care managed Medicare patients.
  • A ninefold increase in patients meeting the care managed technical definition from 2018 to 2019.
  • A 96 percent rate of care managed Medicare patients engaged with primary care.
  • A statewide system for combining and mining claims, clinical, demographic, social determinants of health, and care team and process metrics data.

One of the distinct features of OneCare’s approach to value-based care is focusing on working with community partners to design tools and processes for optimal data gathering and analysis—they prioritize partner engagement so they can better prioritize the patient.

OneCare also promotes health data literacy across Vermont by providing in-person and virtual training for data analysts and champions. This experience and network helped OneCare quickly roll out a self-service COVID-19 tool, empowering providers and care coordinators to use their data skills to identify and prioritize high-risk patients early in the crisis.

Closing Remarks

Dan Burton, Chief Executive Officer, Health Catalyst

 After three days full of speakers, showcases, and friendly competition at HAS 20 Virtual, Health Catalyst CEO Dan Burton thanked the thousands of attendees from over 20 countries—a reach made possible by this year’s virtual platform—for participating. Burton also thanked HAS 20 Virtual presenters, participants, and content contributors, and then shared the highest-rated keynote presenters:

  • Michael Dowling
  • Eric Topol, MD
  • Vice Admiral Raquel Bono, MD
  • Ari Robicsek, MD
  • Anita Pramoda

The highest-rated breakout presenters including the following:

  • John Wadsworth, MS
  • Chris Hutchins
  • Patrick McGill, MD
  • Vivian Anugwom, MS, CHES
  • Amy Flaster, MD, MBA

With a year unlike any other due to the pandemic, Burton reminded the group of the industry’s shared commitment to improvement—even in times of crisis—and his hope that attendees can apply insights from HAS 20 Virtual at their health systems to drive meaningly change. Burton concluded the summit by thanking participants for their tireless work in the ongoing journey to improve healthcare and expressed his excitement to see everyone at the next Healthcare Analytics Summit, September 14-16, 2021.

Additional Reading

Would you like to learn more about this topic? Here are some articles we suggest:

Loading next article...