Digital Patient Engagement
How an omni-channel approach to patient engagement is transforming care management
In their best-selling book Nudge: Improving Decisions about Health, Wealth and Happiness, the University of Chicago behavioral economist Richard Thaler and Harvard Law School professor Cass Sunstein make the case that human behavior is often heavily influenced by small design changes in the way a choice is presented. Whether it’s the not-so-subtle placement of irresistible desserts in eye-level display cases or less overt prompts, like encouraging people to save more for retirement by auto-enrolling them in employer 401k programs, these so-called nudges have been proven to have a major impact on human decision-making.
The theory should resonate with self-insured employers and health plans looking to improve population health through more effective care management strategies. In theory, it should be possible to design a care management program that incentivizes members toward better behaviors and discourages those associated with worse outcomes. But theory and reality do not always align. In the real world of siloed information, incomplete views of patient experience and lagging indicators, simply identifying at-risk patients and engaging with them effectively – let alone designing a strategy to change their behaviors – has historically been a challenge.
Today, however, advances in real-world patient data analytics, conversational AI and digital engagement strategies have made it possible to execute highly personalized care management strategies that intervene earlier, engage deeper and nudge patients to the best possible outcomes. In this paper, we will outline a four-step process to advanced, omni-channel patient engagement that marries cutting edge data and analytics with human ingenuity to optimize the care management process.
Schedule a meeting with our experts to begin the patient engagement journey.
Advanced analytics to identify at-risk populations
Before any patient engagement strategy can hope to be effective, it needs to be able to accurately identify and stratify the target population based on risk. Establishing a population health risk stratification framework will set the stage for all subsequent communications and provide a vital baseline for benchmarking performance of the program over time, but getting that stratification right requires more than just basic risk measures, like history of hypertension or elevated A1C levels. With advanced analytics, it is now possible to stratify health risk across a wide range of detailed attributes, including previous clinical experience, costs, social determinants and several other factors.
In one recent example from a patient engagement program we designed with a large self-insured employer, we designed a strategy to improve cancer care. With a goal of improving outcomes and experience for covered employees and their dependents by optimizing oncology care, eliminating waste and moving to high-value services, we set out to a program focused on five cancer types: breast, prostate, cervical, lung and colon.
That meant digging into EHR and clinical data, lab results, socio-demographic data, medical claims and several other sources to get a 360-view of the complete healthcare experience of the employee population. This process allowed us to not only identify the cohort of patients most at-risk of a cancer diagnosis, but also track the patterns of clinical engagement and healthcare utilization they’ve experienced over time. Through this process (see diagram below), we were able to prioritize those employees who were most at-risk, receiving suboptimal care or otherwise experiencing barriers to care.
AI-assisted patient engagement
Once employees were risk-scored, they could then be flagged for further intervention based on their specific needs. But that meant more than just putting them on a list for a generic e-mail or scheduling them for a call with nurse care manager. With a global employee population of nearly 100,000, one-size fits-all solutions would not be tailored enough and in-person outreach would be untenable. The Goldilocks solution came in a series of outreach escalations that started with conversational AI and touchless outbound calling and eventually built toward nurse care manager dialogue.
This conversational AI technology makes it possible to build a highly choreographed, patient-specific set of behavioral nudges into the care management process without over-reliance on care managers. But it also ensures that the human intervention element is optimized for more complex cases, as illustrated in the diagram below:
Measuring process and outcomes
There’s an adage that you can’t manage what you can’t measure. In the world of value-based care, which is designed to incentivize the use of the most effective interventions and therapies, accurate benchmarking and progress measurement is central to managing an effective strategy. That means being able to track each point of engagement throughout the continuum of care, from risk scoring and chatbot engagement to provider efficiency and quality scores to ultimate health outcomes and cost trends (see graphic below):
To do this effectively and efficiently all data must be stored in a secure, cloud-based solution that makes it possible to track plan milestones and achievements, identify areas for improvement and continually track outcomes.
Driving fact-based provider engagement
The final piece of the equation is a fact based feedback loop. Armed with hard data on cost and quality of care, employers and health plans deploying an omni-channel approach to care management are well-equipped to manage provider relationships. This includes everything from guiding initial benefit design to prioritize top-performing providers to negotiating direct contracting relationships with specific providers based on their individual performance data.
Tying it all together with behavioral science
Much like the “nudges” behavioral economists describe to subtly guide humans to make choices consistent with the best possible outcomes, the entire patient engagement continuum needs to be designed with a behavioral science approach. That means tapping into deep troves of consumer data on everything from social determinants to buying patterns in order to identify the most persuasive paths to a target outcome. The system must then be designed to leverage that data to drive overall care management positioning, guide the use of triggers that will be used to illicit a response and to personalize outreach strategies based on real-world patient insights. For example, a cohort of males aged 50+ with a history of smoking and a cohort of females age 21-40 who go to the gym every day will not respond the same way to a generic message. Ultimately, with powerful analytics, that message needs to be refined and hyper-personalized down to an “N of one” to connect with the right person at the right time via the right channel.
Armed with this level of patient-level detail, enough analytical power to parse it accurately, the tools to tailor custom outreach and the ability to measure results, it is possible to build a highly targeted care management solution combines the best aspects of automation and human intervention to drive improved outcomes.
The EXL advantage
At EXL Health, we apply human ingenuity and deep industry expertise with data, analytics and AI to delivering solutions that help healthcare payers develop more efficient, hyper-personalized solutions and improve the cost and quality of care. By putting comprehensive data, smart analytics and practical insights together in an enterprise solution set, we make it possible to transform the status quo, inspire new breakthroughs and build scalable business strategies.
Find out more about our omni-channel patient engagement solutions here.
Written by:
Prasad Dindigal
VP, Healthcare & Life Sciences, EXL Health