Customer Journey Analytics: Cracking the Patient Engagement Challenge for Payers

This report is based on a 2018 Healthcare Analytics Summit presentation given by Christer A. Johnson, Principal, EY Analytics, and Alexander (Alex) Marano, Customer Analytics Lead, Cigna Information Management and Analytics, entitled, “Using Machine Learning and Big Data to Drive Patient Engagement and Better Health Outcomes.”

There are moments in every patient journey in which a well-informed, well-timed intervention (e.g., phone call, personal visit, etc.) can effectively engage patients and positively influence their health-related behavior. Random engagement, however, is far less effective than analytics-driven strategic engagement. To know when to reach out to and which patients to focus on, payers need an approach that leverages machine learning and big data.

Other industries have used analytics-driven engagement strategies, or customer journey analytics, to analyze real-time data on customer interactions, demographics, and lifestyle events and use these insights to influence customer behavior (e.g., buying behaviors in the retail sector). The healthcare industry is now working to leverage a similar approach to engage effectively with healthcare consumers to help them avoid risky health behaviors in favor of behaviors that improve their health and lower healthcare costs. This report describes a collaboration between a global analytics services provider and a large payer organization to leverage customer journey analytics among healthcare consumers.

Patient Engagement Solutions Target Avoidable Risky Behaviors

Patient engagement solutions support outcomes improvement and cost reduction by targeting behaviors that have negative health consequences but are largely avoidable. These behaviors have a significant impact on patients with chronic and multiple health conditions, who are among the costliest populations.

Considering the following examples of unacceptable high rates of avoidable risky behaviors, improving patient engagement must become an industry priority:

  • Twenty-four percent of adults with diabetes don’t know they have it.
  • About 13 million adults with high blood pressure don’t know they have it.
  • One in five deaths is caused by cigarette smoking.

Addressing the above avoidable health concerns above and their like presents a sizeable healthcare improvement opportunity. But to impact these statistics and the patients who comprise them, health services providers must understand behavioral patterns and factors most linked to engagement.

Patient Journey Analytics for Strategic Engagement

Customer journey analytics follows patient behavior to identify patterns most likely associated with engagement. Analysis may show behaviors and characteristics linked with likelihood to engage and key times to reach out to those patients. For example, the period following a specialist visit can be an engagement opportunity as patients tend to seek information (e.g., internet searches) at this time. If a case manager reaches out to the patient after the specialist appointment, she may have the answers the patient is looking for.

To leverage journey analytics to improve patient engagement, payers create the journey data, follow a framework for and execute the four phases of journey analytics, and operationalize the analytics.

Creating Patient Journey Data

Patient journey data creation (Figure 1) starts with a reference library that holds events, population criteria, outcomes, steps, attributes, and respective SQL logic. From that library come the patient attributes, time-sequenced journey, and journey step attributes. These datasets allow for ongoing and future applications:

  • Ongoing: opportunity sizing, an automated insights engine, individual path analysis, and a rules engine for sensing/responding/anticipating/engaging.
  • Future: simulation analysis to model the likely impacts of health services interventions and automated data mining for new opportunity identification.

Figure 1: Patient journey data creation

The Patient Journey Analysis Framework

The framework for patient journey analysis (Figure 2) asks questions about the patient journey in three key categories:

  1. Understanding value:
    • What is the total value of the opportunity to move customers from undesired to desired outcomes?
  2. Understanding negative outcomes:
    • What kind of customers are having negative outcomes?
    • What kind of providers may be influencing negative outcomes?
  3. Understanding the relationship between journeys and positive outcomes:
    • Which are the most frequent journeys?
    • Which journeys lead to the most positive outcomes?

Figure 2: The framework for patient journey analysis

The framework also considers patient attributes (e.g., demographics, conditions, height, weight, and risk score) and provider attributes (e.g., network status, treatment rules, operating hours, and percentage correct treatment) and patterns (“steps”) most associated with engagement or lack of engagement (“event”). Steps may include internet searches, in-network primary care physician visits, in-network specialist visits, care gaps, ED visits, and more. These insights can help payers improve the timing, channel, and content they use to engage members with chronic and complex conditions in coaching that lowers medical costs and improves healthcare outcomes.

Journey analytics also helps payers identify the most impactable candidates for engagement using likelihood-to-engage score leverages. When the payer mentioned in this report compared using the likelihood-to-engage score with a traditional approach to outreach, predicted overall engagement increased from 18 percent to 31 percent.

Payers have previously based engagement strategies on claims data. As a layered approach to data, a time-sequenced customer journey data lake (Figure 3), combines claims and EMR data, negative events, physical and verbal signals, and digital signals. Payers can apply data to the journey to find moments when interventions will increase the probability that a patient will choose a better path (e.g., keeping follow-up appointments or taking medications as prescribed). These insights identify paths that help create alerts for good times to take action to improve outcomes and reduce cost. Actions include the following:

  • Predict future events and behaviors.
  • Identify signals suggesting changes to expected behaviors.
  • Prescribe interventions to shape the right outcomes.

Figure 3: A layered approach to data 

The Four Phases of Patient Journey Analysis

Patient journey analysis occurs in four phases. Each phase yields detailed journey data for patients identified for case management, a list of the most important journey steps by case management category, and potential engagement and engagement lift by case management category (e.g., oncology or OBGYN):

  1. Data collection and journey data creation
    • Identify selected patients for case management.
    • Query multiple sources to design a time-sequenced journey dataset.
  2. Journey steps importance test
    • Calculate importance of and rank each journey step.
    • Use above results to select the most important journey steps.
  3. Rules selection
    • Modify journeys to include only the most important journey steps.
    • Identify the most prevalent rules (combinations of journey steps).
    • Calculate score for each rule as total engaged/total outreached.
    • Create rules to prioritize certain rulesets over others.
  4. Results
    • Use the selected rules to score patients by case management category.
    • Combine scores at case management level to determine overall engagement rate. (The payer who implemented journey analytics identified the opportunity for a 30 percent increase in engagement [Figure 4]).

Figure 4: Example of increasing overall engagement rate

Operationalizing Patient Journey Analytics

To be effective, payers must operationalize customer journey analytics. The key differentiators between basic analytics and operationalized analytics is that basic analytics only analyzes and reports (informs), whereas operationalized analytics, anticipates, engages, senses, and responds (the middle orange box in Figure 5). With the added capabilities of operationalized analytics, payers can better sense opportunity, engage in real time, and personalize these interventions over time.

Figure 5: Operationalizing customer journey analytics

Operationalized journey analytics have significant benefits over basic analytics:

  • Leverages all big data available (versus focusing only on probable data).
  • Identifies drivers of outcomes (versus tests hypotheses).
  • Run hundreds of tests simultaneously (versus limited manual evaluation).
  • Learns and modifies actions continuously (versus retrospective annual turning).

Anticipate, Sense, and Respond

Healthcare organizations can’t help patients if they can’t engage them effectively, and in an increasingly complex health landscape, better patient outreach is an imperative. Engagement improves significantly with an analytics-driven strategy that identifies whom to engage, when, and how. With patient journey analytics, payers and other healthcare organizations can influence better health behaviors to improve outcomes, customer satisfaction, and lower healthcare costs.

Additional Reading

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

  1. Six Challenges to Becoming a Data-Driven Payer Organization
  2. A Guide to Governing Healthcare Claims Data Successfully: Lessons from OSF HealthCare
  3. The Secret to Patient Compliance: An Application of The Four Tendencies Framework
  4. Healthcare Payers and Providers: The Best System for Process Improvement
  5. Improve Patient Engagement with Five Public Health-Inspired Principles
Loading next article...