Sepsis is a word that can evoke feelings of dread for most people and yet, many fail to understand its meaning and implications fully. It is a life-threatening condition that arises when the body’s response to infection results in organ dysfunction or failure. It is one of the leading causes of deaths with 6 to 9 million deaths per year worldwide and 250,000 deaths per year in the United States. To make matters worse, sepsis is also one of the leading causes of preventable deaths worldwide. The key to prevention is early detection since every hour of delay increases the odds of death by 20%. However, sepsis can be very hard to detect especially in its early stages when it is reversible. The symptoms often point to other conditions and patients can deteriorate rapidly. Traditional risk models such as SIRS criteria to detect sepsis generate a lot of false positives leading to inefficient or ineffective care and nursing fatigue.
Leveraging Machine Learning
Due to PCCI’s close collaboration with Parkland Health & Hospital System, we recognized sepsis detection and prevention as an area that could benefit from an advanced machine learning-based algorithm. To be effective, the model needed to fulfill the following criteria:
A high enough positive predictive value (PPV) to ensure that false positives do not end up causing frustration for clinicians. It is very challenging to have a high PPV for use cases with low prevalence rates in the target population. Despite the high mortality numbers, the overall prevalence rate for sepsis in an inpatient setting (our target population) is 3% to 4%.
A prediction interval (how far ahead in the future can the model predict) that gives clinicians a sufficient heads-up to intervene.
PCCI developed a predictive model to fulfill the criteria identified as effective. This real-time model is designed to predict an individual’s risk of becoming septic within the next 12 hours.
Instead of relying on traditional ways of detecting the risk of sepsis through a handful of physiologic variables, we cast our net wide and started out with approximately 120 variables as potential predictors, such as socio-demographic, vital signs, co-morbidities, hospital utilization, medical conditions, clinical history, and lab results.
Testing the Lasso Logistic Regression Model
We trained and tested a Lasso Logistic Regression model, a decision tree and a neural net on 27 months-worth of Inpatient encounters from Parkland. Initial data sets included 54,629 encounters and 30,922 patients. We chose the Lasso Logistic Regression model as the final model because of its high performance and interpretability by clinical users. The resulting model that compared favorably to other community models resulted in:
95% Accuracy
30% PPV
55% Sensitivity
Incorporating the Model into Clinical Workflows
A model is useful only if it is actionable and available at point of care. PCCI’s Sepsis model is incorporated into clinical workflows through industry-standard APIs. The model accesses EHR data in real-time every five minutes and alerts the clinician if the risk is above a certain threshold. The risk threshold can be tailored to local populations and clinical best practices of an organization. The alerting mechanism can vary based on EHR functionality, but it can include a simple alert, queuing up of order sets based on the risk threshold, and sending a real-time alert to a secure mobile device or pager for a member of the rapid response team.
1. Purpose Driven Change-Leadership Workshop by David Shore, PhD, Harvard
Throughout the summit weekend, many workshops were presented by thought-leaders in the space. David Shore’s workshop was a veritable delight of new twists on old themes to jog the mind and start a new race for transformative change within one’s sphere of control. Some key points included:
Spending extra time shaping questions to ask increases the efficiency at arriving at solutions
Project life cycles should be front-end loaded with interrogations of reality and refuting assumptions
Conduct a sequence of smaller projects that feed into a cohesive program instead of long drawn out projects
It’s only innovation if you effectively solve meaningful problems, which you can scale and spread
Sustain with the “Science of Spread” methodology
According to research, the optimal size of a project team is seven to eight people – if it takes more than two pizzas to feed your team lunch, you have too many people!
Many interesting points of view of healthcare providers regarding SDOH
While 40-50% see their important influence on outcomes, 70-90 % don’t necessarily think it’s their job to respond to those needs.
A personal favorite: go beyond lessons learned to lessons leveraged!
2. Extensions of the Triple Aim Statement Reframing the Importance of SDOH
First, we had the Triple Aim, then quadruple and now… the quintuple aim:
Cost
Quality
Patient Experience
Provider experience
EquitySDOH
As this Triple Aim Statement continues to expand, what do you envision to be the sixth?
3. Social and Healthcare Platforms
Early stage entrants working on cloud platforms to connect care, patient created and social data are seeing encouraging early gains. Below are some notable platforms to keep an eye out for:
The Real-World Education Detection and Intervention (REDI) Platform:
Currently deploying in border towns along southwest Texas by UT Austin Lynda Chin, MD’s team in collaboration with PWC (pro bono), AWS, and Walmart
They report a 1.7% decrease in Hgb A1 c of diabetics in an integrated data sharing program with remote monitoring
An Oregon 1115 Waiver project is driving to get large numbers of community health workers across the state to document on and create closed-loop referral
They focused more on the human aspects of this and it seems that they may still be in technology development
4. Powerful Visualizations for Action
This was a “blow you away presentation” with some truly powerfully meaningful novel approaches driven by Jason Cunningham, MD, CMO of West County Health Centers. Below are suggested steps one can take to innovate the virtual world of healthcare:
Use a mix of vendors to include Tableu, Unifi + KUMO + Argis
Create visualizations for actions. For example, zip code areas affected by wildfires were targeted and cross referenced with their patient list allowing the ability pinpoint their patients for proactive outreach
Allow for early identification and replacement of lost belongings including medications, medical supplies, and strong patient experience feedback approval
5. Early findings and Interesting Metrics to Prove the Value of SDOH Intervention
While the consensus opinion and extensive research clearly indicate the magnitude and causal nature of SDOH’s influence on health outcomes, quality, and cost, most interventions depend on unique funding streams. This is because ROI hasn’t been proven to hit mainstream reimbursement. Examples include:
WellCare Insurance Plan reported a decrease of $2,400 per year per member for those who received social needs interventions versus those who did not
Sutter Health used a Health Equity Index to target risk populations affected by disparities and used the index to prove intervention effects
Special Needs Plans (SNP) used a “Loneliness scale,” which contributed to disease progression and longevity to target and monitor at-risk individuals
Return on investment is largely focused on health outcomes, but how can we measure the social outcomes of Social ROI?
6. Speeding Up Patient Transport
Getting patients where they need to go, when they need to go is a top priority that has an impact on not only outcomes but patient experience in terms of ease and convenience. Just think about your own stress when your car is in the shop, stress can agitate any clinical state. Interesting approaches to speeding up patient transport include:
Ordering patient transport through referrals in their EHR
Superimposing public transport routes onto patient location density and using the information to advocate for new routes
7. New Term – “Patient Disengagement”
Patient engagement often is a “catch-all” bundled term. But new ways of disentangling the terms unlocks possibilities, such as:
Disengagement Vulnerabilities- a method of enumerating characteristics of individuals and their circumstances that can interfere with engagement to target and develop personal connectedness
Tangible incentives are used to increase participation and encourage healthy choices
8. Payer Pressing Mobile Engagement for the Homeless Who Are “Not Ready to be Housed”
“Housing first” advocates began changing the landscape and the dialogue on the all-too-common reality of homelessness. One size doesn’t fit all in this multidimensional problem. A notable example is the homeless and housing resource team created by ANTHEM Indiana Medicaid. If patients aren’t ready to be confined by walls, the program provides a cell phone and a mobile app to engage them with online tools.
True population health management requires at least three fundamental elements to drive transformational change and meaningful results:
Aligned incentives across payers, health systems, post-acute care providers, and physicians.
Technology and framework fit for engaging an entire community that leverages resources for care coordination and addressing social, economic, and behavioral needs
Personal engagement to drive activation, behavioral modification, and create a foundation for shared decision making.
HOW ASTHMA AFFECTS POPULATION HEALTH
While this framework is required for managing all populations, it becomes critical and complex when managing chronic disease. For example, asthma is a common disease, but it’s rarely recognized as one of the most common chronic diseases for children under the age of 18 with 6.2 million affected individuals. Over 8% of children have asthma, most with symptoms occurring before five years of age. Asthma disproportionately affects low-income, minority, and inner-city populations with African-American children being at the highest risk. It is a significant driver of school absenteeism, with an estimated 12-15 million school days lost per year.
Asthma impacts both families and the healthcare system financially as well as socially. Childhood asthma is the cause of nearly five million physician visits and more than 200,000 hospitalizations each year. Medical expenses for a child with asthma are almost double than those for a child without the disease. Given these statistics, there’s a compelling need for early identification and effective intervention to control this disease.
PCCI’S POPULATION HEALTH FRAMEWORK
The Parkland Center for Clinical Innovation (PCCI) has developed sophisticated predictive models to proactively identify children at risk for asthma exacerbations and has combined this powerful engine with a comprehensive population health framework to:
Reduce asthma emergency department visits and hospitalizations
Increase patient adherence to medication and clinic visits
Increase evidence-based leading practices at the provider level
Figure 1. PCCI Pediatric Asthma Framework
PCCI’S PEDIATRIC ASTHMA MODEL AT WORK
Through tailored clinical workflows, monthly provider reports, point-of-care EHR integration, and patient-centric mobile messaging applications, the framework can engage providers, communities, patients, and their families to optimize care, drive engagement, and reduce unnecessary utilization.
Within three years, deployment of the program in the Dallas metro-area by a large health plan resulted in:
32-50% increase in the appropriate use of controller medications
31% reduction in ED visits
42% reduction in asthma-related inpatient admissions
This framework has resulted in more than a 40% drop in the cost of asthma care with the health plan saving over $18 million for both patients and healthcare providers.
ENGAGEMENT IS KEY
The key to PCCI’s pediatric asthma framework is that the clinicians, patients, and health systems are all engaged which generates value for all parties involved. With a foundation in literature-based evidence, the framework aligns with national and international guidelines. It is also both modular and patient-friendly – offering different levels of interventions based on patient risk score, needs, and available resources.
By utilizing our sophisticated predictive model to identify children at risk for asthma proactively, we are able to combine it with a comprehensive population health framework to:
Increase patient adherence to medication and clinic visits
Educate patients in care and self-management
Optimize health plan to care manager outreach and workflow
Engage physicians via direct EHR alerts
Reduce preventable asthma ED visits and hospitalizations
CURRENT DEPLOYMENT
ENGAGEMENT WITH THE POPULATION HEALTH FRAMEWORK
High and very high-risk patients can engage through succinct, precise, and educational text messages delivered by a simple effective mobile platform. Patients are surveyed about their condition, emergency inhaler usage monitored, and they are reminded of upcoming appointments and medication refills. These innovative features allow continuous symptom monitoring by the clinician to ensure continuity of care and positive outcomes. Patients have displayed satisfaction with the program, with over 70% top box satisfaction and an attrition rate of less than 15% on mobile engagement over a two-year period.
Community engagement is a critical element when trying to ensure not only coordination of care, but referrals and connections to community resources. Providers at either hospitals or the clinics, receive best practice alerts and utilize technology to identify SDOH needs. They can also refer families to community-based organizations (CBO) for assistance with critical daily needs. Currently, we’re in the process of expanding interactions and engagement with local schools, so that school nurses concurrently receive alerts on high risks children and can help coordinate care in those settings.
COMPETING POPULATION HEALTH FRAMEWORKS
While there are multiple and broad initiatives occurring in every market, results have displayed limited to incremental progress. PCCI has demonstrated that transformational change and meaningful results are achievable. Meaningful results require concurrent engagement and coordination of payers, providers, community, and patients via advanced risk-predictive stratification algorithms and deployment of information via new/updated workflows at the point of interaction. Figure 2 highlights the difference and impact our comprehensive program has across a market.
Figure 2. PCCI Pediatric Asthma controlled analysis. Comparison 1: DFWHC Medicaid <18 yo: 5% drop in asthma ED visits. Comparison 2: All Health Plan Members <18 yo: 10% drop in asthma ED visits. PCCI Asthma Program: 31% drop in asthma ED visits
Additional Innovations
Despite tremendous success, opportunities still exist to improve results and continue to further engage providers and patients. We are designing and rolling out two additional innovation pilots:
Testing the effectiveness of disease-specific in-home personal assistance devices (Amazon Echo) to engage groups of homogeneous, high-risk, pediatric asthma children in a gamified home environment.
Integrating within a home or community to allow remote monitoring of asthma medication adherence and in-home air quality by using “Internet-of-Things” integration.
The framework is designed with adaptability in mind and is ideal for environments where providers hold risk-based contracts. Future applications will include other patient populations and health conditions. PCCI’s Pediatric Asthma Population Health Framework has not only reduced unnecessary utilization and costs, but it has also improved the healthcare experience for hundreds of pediatric patients and their parents.
Central to the work of PCCI is its strength in predictive analytics/modeling and building connected communities using intelligent, integrated, electronic information exchange platforms. Our state-of-the-art programs and advisory services deliver exceptional value to Parkland Health & Hospital System, the local community, and the broader healthcare market.
Recently, PCCI had the opportunity to welcome students from two Dallas ISD schools– Townview Science and Engineering Magnet and Emmett J. Conrad High School. This group of students are finalists in the Lenovo MIT App Developer Competition and have been invited to Washington D.C where they will be recognized in an awards ceremony for their innovative app aimed at addressing social and health needs in their communities.
A Different Type of App
Nine Dallas ISD students were given the opportunity to design an app for a competition. While most kids would opt for creating a gaming app or the next social media hit, these students took a different approach. Two-thirds of their classmates come from economically disadvantaged families, and many of their friends, neighbors, and family members have little access to food or healthcare resources. Deeply rooted in their passion for helping their community, they decided to develop “Klinik” an app that could enable them to address these immediate needs.
What is Klinik?
“Klinik” is a mobile app that provides users with crucial information on how to get access to basic resources such as food, healthcare, and shelter. The team of prepared and articulate students shared their project with PCCI innovators, spurring conversation around various aspects of app development. By brainstorming effectively, canvassing their communities to create a database of resources, and having potential users provide feedback on the design during development, these students were able to create an app with a sophisticated design and scope mirroring methods seen in well-established companies.
Learning how to compile accurate community resource inventories that enable those who need them according to their preferences is an active area of research in the field and a strong focus for PCCI. The students having developed the basic app are assessing how it can be scaled for broader use across geographies.
PCCI Strategizes with DISD Students
PCCI will strategize with DISD Academy programs and NAF Future Read Lab to create opportunities for collaboration and learning during innovation cycles and to extend the applicability of initiatives. We think that these students and their peers have the potential to work closely with PCCI through worksite tours, internships and a variety of other NAF sponsored activities.
Their work was sponsored by the Lenovo App Inventor program. Lenovo Scholars provides STEM education through NAF academies in the U.S. in partnership with the Massachusetts Institute of Technology (MIT) to enable the next generation of app developers.
Over the last few years, it has been very clear from research that Social Determinants of Health (SDOH) variables have a major impact on health outcomes. It is estimated that close to 80% of health outcomes are impacted by SDOH. With the rise of population-based risk contracts in both the commercial and government sector, it is essential for both providers and payers to collaborate in the identification of best practices to address these SDOH variables. This is especially relevant as providers such as hospitals assume greater risk in arrangements with plans throughout the country such as Accountable Care Organizations (ACO) and bundled payments.
NATIONAL INTEREST AND PROGRESS
Many national associations such as the American Hospital Association (AHA) and America’s Essential Hospitals have developed resources and launched learning collaboratives for hospitals and health systems to address these variables such as food insecurity, housing, and transportation. Health system innovation and care-redesign models driven by organizations such as Healthbox and AVIA have launched collaboratives and forums to educate and address SDOH initiatives. The May 3, 2018, Healthbox forum discussion on “Challenging the Status Quo of Social Determinants” visually captured the opportunities and challenges ahead into one image (Figure 1):
Figure 1: Image captured during Healthbox Executive Panel Discussion, May 3, 2018. Chicago, IL
These variables have always been a focus of many health systems in terms of articulating their benefit to the community, but now they have particular importance given the rise of more population risk contracts.
Several major barriers have impeded the industry’s progress in addressing SDOH variables: funding and regulations. Fortunately, we have begun to see opportunities in both areas emerge in 2018!
MEDICARE UPDATES AND THE BENEFITS OF SOCIAL DETERMINANTS OF HEALTH DATA
Medicare Advantage (MA) has a regulation titled “Uniformity Standard” that requires all of the plan’s benefits, including cost-sharing, be the same for all plan enrollees. On April 2, 2018, the Centers for Medicare & Medicaid Services (CMS) outlined several widespread changes in this regulation that both providers and plans have advocated for over the last several years in their 2019 Medicare Advantage Call Letter. CMS expanded the flexibility of lifting the uniformity of supplemental benefit to allow different segments of an MA plan to offer specific benefits to a targeted population like diabetics. This can begin in CY 2019 (January 1, 2019) after the plan designs are approved by CMS. An example could be reduced cost sharing for foot or eye exams. In their official bids that were submitted by the June 4, 2018 deadline, the MA plans can include any of these supplemental benefit elements. Hopefully, providers will see many of the plans deciding to include these additional benefits in their MA bids to address the SDOH variables.
Additionally, in the Bipartisan Budget Act (BBA) that was passed in early 2018, Congress has taken it further by extending the lifting of the uniformity of the supplemental benefits to all chronically ill members of the MA plans effective January 1, 2020. This reinforces the need for us to gain valuable lessons during 2019 in order to determine what works and what doesn’t before it is transitioned to a broader population.
The Chronic Care Act of 2018 extended the Center for Medicare & Medicaid Innovation’s (CMMI) Valued-Based Insurance Design Model to all 50 states in 2020. This model was launched in 2017 to allow Medicare Advantage plans to offer supplemental benefits and reduced cost-sharing to seven conditions including Coronary Artery Disease or Congestive Heart Failure. The model focuses on four approaches:
Reduced Cost Sharing for High-Value Services
Reduced Cost Sharing for High-Value Providers
Reduced Cost Sharing for enrollees participating in disease management
Coverage of additional supplemental benefits such as transport or meal delivery
The creation of more supplemental benefits will enhance the quality of services we provide for our patients especially in terms of addressing the SDOH. Encouraging the inclusion of these targeted supplemental benefits will allow us to partner with payers to improve the health of the country in a more innovative way.
ADDRESSING SDOH WITH HEALTHCARE PROVIDERS AND COMMUNITY RESOURCES
At PCCI, we have been directly involved in national and state-driven education forums, presentations, and roundtables directed to design and deploy local models for the Connected Communities of Care program (previously known as the Information Exchange Portal) that bring together providers, payers, philanthropic organizations, community-based organizations (CBO), and local/state government entities. While most markets continue to be in a learning mode, significant and tangible activities are being initiated in a number of municipalities, including Dallas, Raleigh-Durham, Louisville, Detroit, Chicago, Phoenix, Salt Lake City, as well as across whole regions. For example, North Carolina recently requested proposals for the development of a North Carolina Resource Platform via the Foundation for Health Leadership & Innovation. The goal of this multi-year program is to connect over 3,000 statewide community-based organizations via technology, and facilitate SDOH. This will be completed through a programmatic coordination of referrals between healthcare providers and community resources to comprehensively identify and address the needs of individuals across the state. On a broader level, the Accountable Health Communities Model deployed in 2017 is engaging 31 organizations across the country to address a critical gap between clinical care and community services in the current healthcare delivery system. This is being done by testing the process of systematically identifying and addressing the health-related social needs of Medicare and Medicaid beneficiaries through screening, referral, and community navigation services to see if it will impact healthcare costs and reduce healthcare utilization.
SUCCESS IN SIX TRACKS
Our experience over the last five years across Dallas tells us that models will need to address six tracks to be successful: Governance, Legal, Technology, Clinical Workflows, CBO Workflows, and Sustainability (Figure 2). The maturity and evolution of the models need to develop and be staged within a multi-year deployment framework (concentric circles in Figure 2 represent the progression and evolution of the model with outer circles representing mature and more sophisticated models).
Figure 2: Connected Communities of Care program multi-year deployment framework
There is also a critical upfront readiness and deployment/implementation assessment that is important in order to stage the deployment of a Connected Community of Care program. This broad representation of the community’s fabric is critical to ensure that:
A community is ready to undertake the operational and financial requirements associated with deploying a Connected Communities of Care program
The healthcare and social needs of the community are at the forefront of the customized design of the platform (something most for-profit technology vendors offering an out-of-the-box solution either cannot do or fail to do properly)
The design is sufficiently flexible to adjust as the healthcare or social needs of the community change
Addressing SDOH is finally moving from a “buzz” word to implementation pilots. While we talked a lot about population health over the last 10 years, doing population health without a truly engaged and “Connected Community of Care” is like focusing on rescuing people from drowning in a river vs. building a bridge so they can cross it safely. As we continue this journey, let us make sure we build a bridge that adapts to the needs of each community and has emerging local and national models of care to ensure sustainability. We don’t want to end up with a bridge like the Choluteca Bridge in Honduras, connecting nothing to nowhere.
Acknowledgments: Valinda Rutledge, PCCI Executive Advisor and Lindsey Nace, PCCI Marketing and Communications have contributed to this article.
How often did you get the question (or have asked it yourself) during an interview: “Why are you interested in our organization?” Simple, standard, mundane – on the surface, some might even call the question un-inspiring. I disagree. It tells me right off the bat how much effort the candidate put into learning about the Parkland Center for Clinical Innovation, our organization, our work, our team and how they synthesized and interpreted the information. I can tell within the first two minutes how interested I will be for the next 28. Regardless of the level of experience, I’m way too often disappointed by the response.
Using Python to Create a Sentiment Analysis
Recently, I was blown away! I asked the same “boring” question to a candidate interviewing for an entry data science position at PCCI. As soon as I finished the question, his eyes lit up and he quickly pulled out a document from his bag. With great enthusiasm, he replied:
“In addition to my own research, I wanted to know what others are saying and feeling about PCCI. So, I wrote code in Python to create a Twitter sentiment analysis. I used it to create a word cloud and analyzed it to see if the keywords match my passion and interpretation of my own research. These four words really resonated with me because … I also wanted to understand PCCI’s reach and brand recognition, so I analyzed the top 10 famous people and companies talking about PCCI. I was impressed to see @HarvardBiz, @washingtonpost, @NIH, @HHSGov, etc, but most importantly to see @KirkDBorne. He’s so influential. Finally, the outputs of the Sentiment Count Plot Analysis and the Sentiment Subjectivity Distribution reconfirmed that this is a great place and the place I want to be.”
A Match Made in Data Science
I know I’m a geek at heart and this answer resonated with me more than it would with most (did I mention earlier how important it is to know your audience and their interests when answering a question?), but regardless of what approach you take, this is how you do it!
