Predictive Analytics: A Prescription for a Better Patient Experience

Healthcare consumers are increasingly aware that they have options. And they want stakeholders—whether health plans or providers—to help them make the best decisions about their care through personalized communication and education.

As consumer expectations continue to raise the bar on the patient experience, healthcare organizations must shift their focus upstream in care episodes to help patients make the best decisions. Advanced predictive-analytics techniques and artificial intelligence infrastructures are helping industry stakeholders identify the specific needs of patients in the populations they serve. Armed with deep analysis of consumer healthcare purchases and behavior, care teams can target engagement and educate patients about the financial implications of their options.

The Rise of Predictive Analytics

Analytics strategies aimed at predicting consumer behavior are common in industries such as retail, which focuses heavily on understanding the buying preferences of potential customers. Businesses then can interact with consumers on a more individual level, targeting them with smarter product offers and building stronger relationships.

In health care, patients considering their options increasingly are seeking personalized decision support that considers specific nuances of their care. In contrast, broad communication that is not relevant to individual needs (such as sending a reminder to schedule a screening to a patient who already has done so) is likely to annoy a patient and erode trust rather than improve the relationship.

ealthcare stakeholders should consider the track record of consumer-driven trends in other industries and incorporate market-focused, consumer-oriented strategies that reflect precise communication with patients. By leveraging claims data enriched with clinical data and past-usage patterns, organizations can better identify future health indicators and prepare patients for the healthcare decisions that lie ahead.

For example, claims information might reveal that a patient is receiving physical therapy or recently had a cortisone shot. When combined with consumer insights, predictive-analytics platforms might point to a future need for a CT scan. Care teams can respond to these signals by educating patients about their options, taking great care to not alarm patients, but rather nurture decision support after having earned trust through individual relevance. The earlier healthcare organizations begin messaging and communicating about pending needs and options, the more empowered patients feel in understanding the implications of their decisions.

A More Precise Strategy

Most healthcare organizations understand the power of analytics strategies in data-driven care. Through analytics, organizations have the opportunity to steer appropriate interventions to the patients who need them, improving both the overall experience and clinical outcomes.

The first step is understanding the difference between descriptive and predictive analytics. When health plans and providers target consumers based on healthcare utilization and demographic information such as gender, age, education, and employment, they engage in descriptive analytics. For example, a strategy may entail sending reminders to all women ages 45 and older to get a mammogram if they have not scheduled one in the last 12 months.

In contrast, predictive analytics targets consumers based on healthcare utilization and other behavioral signals to uncover a pattern and predict future outcomes. Although predictive analytics strategies have the potential to notably elevate the patient experience, health plans and providers must consider that a delicate balance exists when communicating with patients about future healthcare needs. Simply put, health care and finances are highly emotional subjects for most people. In formulating strategies, healthcare organizations need to carefully nurture relationships while also considering the impact of sensitive information.

A health plan may deliver tips to a patient regarding the best orthopedic surgeons in the area or alternative treatment paths, for example, when claims activity indicates a future need for knee surgery. However, model imperfection exists (generating false positive signals), so great care must be taken in the sequence and tone of a trust-based nurturing campaign.

Ultimately, this data-enhanced model helps care teams prioritize communication based on risk, reaching those patients needing the most help first. Taking it a step further, financial counseling strategies can be deployed to help patients set up financing in advance of future care encounters..

Healthcare stakeholders must first understand consumer needs and then build solutions around those needs, as opposed to building the solution first. This is accomplished most efficiently not by asking consumers what they want, but by observing and then iterating prototypes with real users.

Applying Predictive Analytics

The opportunities presented by advanced machine-learning techniques to understand and predict consumer health behavior are notable. Artificial intelligence frameworks that exist today aggregate key claims and clinical data by leveraging machine learning and predictive modeling to enhance patient relevance by key indicators, and these platforms continue to learn from themselves and improve. Real-time monitoring of pre-identified procedure bundles can produce signals to predict timing of coming medical needs and offer care teams the timely decision support needed to educate patients about their choices. This approach, paired with clinical training and evidence-based guidelines, could offer a more proactive way to identify key areas for intervention.

Providers and health plans should begin with high-risk patient populations to effectively reach those people who need the most help with decision making. Conditions and procedures that are characterized by wide variability in cost and outcomes are good starting points. Once these areas are managed successfully, organizations can target lower-frequency treatments and eventually focus on prevention.

Positive patient experiences increasingly hinge on partnerships that empower patients with support for making the best choices. Today’s healthcare consumers want to align with “brands” and clinical relationships that they feel are proactively invested in their health for the long term. Healthcare organizations that leverage advanced machine-learning techniques stand to successfully improve clinical outcomes and patient experience while building and retaining market share across a complex and competitive industry.

Keith Roberts is vice president of engagement, Change Healthcare Engagement Solutions.

Lucas Lukasiak is manager of analytics at Change Healthcare Engagement Solutions.