Skip to content

Corporate Wellbeing Health Plans Health System Growth Technology

How consumer data (not more clinical data) will fix healthcare

Most people spend far less than 1 percent of their lives in a clinical setting, but the relative use of data and information technology in healthcare has failed to address this conundrum.

More digital data is flowing within the healthcare industry than ever before, which is commonly and erroneously perceived as transformative in and of itself.  While the healthcare industry becomes increasingly adept at applying clinical, claims and “omic” data to predict illness and improve the coordination and quality of care for sickness, it has largely ignored other data sources that provide the greatest opportunity to positively impact health status, outcomes and cost at scale. How can this be?

There is strong agreement among healthcare experts that medical care and genetics drive only about one-third of health status. So why is the healthcare system treating us as patients all the time? Because the healthcare system is a sick care system and the data model that feeds it is broken. The way to radically, but respectfully, fix the healthcare system is to apply consumer data to help all types of population health managers partner with people, not just as patients, but as consumers.

Patients and consumers

Imagine a simple device where healthcare data insights resulting from all this digital big data get dropped into an opening that branches into two tubes. Down one tube, the device translates the insight into a list-of-things that medical professionals should do to a person to fix what ails them. Down the other tube, the device translates the insights into actions that the person can take to achieve their highest health status in the context of their daily lives and in a way they can understand. We’ve spent decades working to perfect the first tube, but what about the second? Patients are not the same as consumers. Patients receive care. Consumers make choices. In reality, most people are consumers most of the time and some people are patients some of the time. In fact, most people spend far less than 1 percent of their lives in a clinical setting, but the relative use of data and information technology in healthcare has failed to address this conundrum.

Both government and private sectors have dedicated significant capital and energy to building electronic health record and claims systems to automate and record sick care transactions. This increased digitization supports consistent quality of care and payment accuracy, but this data is primarily retrospective – a story of what has been. How do we get to the story of what can or will be? This industry needs to rapidly build the pipes and the aptitude for amalgamating consumer data, which is a much stronger predictor of how a person will or won’t interact with available health resources, with medical and claims data. This creates immediately actionable information that population health managers, including health plans, large employers, health systems and even retail pharmacies, can put into play to change a health outcome.

To have any reasonable chance of changing the trajectory and value of the $3.1 trillion ballooning healthcare costs in the U.S., it is imperative for consumers to become more active participants in their health. Every day we make decisions that impact our health, from food consumption to physical activity to management of known health conditions. Keep in mind that individuals who experience daily lifestyle and environmental factors are not closed-loop experiments. Adherence and compliance is the vocabulary of doctors to patients, not consumers. We all have variations in our relationships, finances, stress, locations, and emotions. Almost every healthcare clinical and administrative professional knows this intrinsically, and yet the healthcare system continues to operate primarily in a sick care vacuum.

If we go back to the simple, hypothetical device described above, a few actions by each of the many consumers emanating from the second tube are certainly more powerful than many actions being required of relatively scarce and overburdened medical professionals. And it is no stretch to consider that consumer incentives tied to consumer actions are a powerful ally to provider incentives that support value-based care delivery.

Consumer data is imperative and full of opportunity

Surprisingly little work has been done to date to understand the relationship between a consumer’s daily living with the impact on their health and how they interact with the healthcare system. Over the past five years, we have worked to change that. To build predictive models and make these data correlations, we created a proprietary database comprised of multiple private and public sources that represents 275 million consumers across America. Machine learning and cognitive computing also allow us to analyze and apply the massive amounts of unstructured data today’s consumers are generating – as well as add in data from physical activity tracking devices and social networking activity. We have identified approximately 800 variables that are indicators of consumer behavior and intent that supplement existing clinical and claims data.

Our data scientists have assembled a machine-learning model to assess consumer impactability and receptivity to various types of health-related activities. A single predictive model instance, such as who would be most receptive to a care management program, requires at least 20 million algorithms. Machine-learning builds, tests and learns from every single combination of data points such as a diagnosis, plus income level, plus commuting pattern. Every 10 seconds, our machine-learning model generates a new set of predictive data insights – also known as intelligence – based on the available data variables, before moving on to the next set of combinations and permutations. These predictive models reveal the behavioral side of healthcare and identify interventions, programs, and messages to drive the best value and highest measurable impact. Most importantly, the findings are understandable by ordinary people.

Perhaps not surprisingly, non-healthcare data such as home ownership, transportation choices, and buying habits are much more telling about what a person can or may do related to their health. These data points are leveraged in nearly all other industries to target prospective buyers or influence consumer behaviors, so why not in healthcare?

For example, household composition and voting history can be leading indicators for emergency room usage. Households with children are more likely to use the ER inappropriately, which exposes an educational and incentive opportunity for alternative, lower-cost care options. Taking it one step further, data can also reveal the preferred and most effective engagement channels – text, automated voice, email, mail, phone, coaching or a combination approach. Just for fun and inspired by the Super Bowl, we identified that Philadelphia Eagles fans are more likely to join condition management programs if contacted by text than by email.

The proof of value is in the data

By combining key consumer attributes including socio-economic status, voting history, education level, medication adherence rates and demographics with client data, we generated a 16 percent increase in medical cost savings and 10 percent improvement in program engagement for a national health plan. This was achieved by identifying individuals at risk of 30-day readmission within six months as well as those who would likely be impacted by a care management program and receptive to outreach.

In another instance, the identification of key predictors, including previous year results, education level, health status, home value and age, determined the likelihood of an individual not recertifying, but still eligible for a Medicaid program. The predictive models successfully identified the top 25 percent who were 1.8 times more likely to not recertify and the optimal outreach channels (automated voice and live agent calls). This targeted and coordinated outreach resulted in a 39 percent decrease in recertification failure rate compared to the control group.

Another use case for consumer data is to fill in the gaps if clinical data is unavailable or inadequate, as is often may be. As people change employers or health plans, the majority of clinical data typically does not transfer. One client had a dearth of information for a population segment that was largely overlooked before the application of consumer data. The analysis of lifestyle and demographic information revealed that 14 percent of those consumers were obese and/or diabetic and would be receptive to targeted interventions, giving the client a head start on activating that population.

Combining consumer and clinical data

While consumer data is imperative, it is not absolute. The combination of all available data – both consumer and clinical – create the most accurate predictive models. Said another way, the data generated by EMR/EHR systems and benefits administration systems remains vital. But, by adding data about consumers and consumer actions – which requires a complementary type of consumer activation system, and by applying machine learning to these consolidated data sets, it is now possible to coordinate provider and consumer efforts to drive behavior change and improve outcomes at scale.

Bottom line, the absence of consumer data in healthcare is equivalent to ignoring the most untapped change agents of healthcare– consumers – altogether. There’s empirical evidence that the combination of all data types is a powerful means for advancing targeting and consumer activation. Since population health is the summation of individual health, it would be irresponsible to discount consumer data and personalized consumer activation. Likewise, it would be irresponsible for traditional health organizations to assume they can readily and acquire and apply consumer data. (How many years has it taken for them to predictively apply the data they generate?) It takes years to amass and license the consumer data, train the machines and build the models. Partnerships are the optimal approach to consumer-data health endeavors rather than each organization expending the resources and time required to build the second tube.

Clinical data and expertise are vital, but the only way the healthcare industry will truly fix itself is to understand consumers at an individual level – by leveraging information about every aspect of their lives – to create personalized experiences that ultimately drive behavior change and improve outcomes.