By Luke Rabin, Director of Product Development, Creo Wellness
How conversant is your organization with the tech trends that are re-shaping the wellness industry? Stories about big data, the Cloud and the Internet of Things (IoT) pepper industry publications, but the adoption of machine learning technology marks the spot where the rubber truly hits the road.
According to Gartner’s Top 10 Strategic Technology Trends for 2016 report, the machine learning trend is an “extension and accelerator of digital business” that will be part of every industry by 2025.
The trend, the report continues, will reach beyond IT to influence corporate business models and completely reshape industry practices. Suffice it to say, this is huge.
Big Data, pouring in from a myriad of sources, provides the raw fuel for disruption of the health care sector. Taken alone, however, the deluge of data lacks value. Alchemy occurs when machine learning, the ability to use complex algorithms to uncover relationships between disconnected data, makes predictions on incomplete data, and in turn empowers machines and systems to become autonomous actors. The health care sector lags behind other industries in adopting an algorithmic approach, but several factors indicate the time for adoption is here.
Government Policy: The implementation of the Health Information Technology for Economic and Clinical Health (HITECH) Act in 2011 incentivized medical institutions to adopt electronic health records. A snapshot of IT adoption provided by HealthIT.gov shows:
- 95 percent of all eligible and Critical Access hospitals have demonstrated meaningful use of certified health IT systems
- 56 percent of all office-based physicians have demonstrated meaningful use of certified health IT
Unlike the datasets generated by huge public health studies of the past (i.e. NHANES or the Nurses’ Study), which took years to collect and analyze, clinical practices now generate enormous datasets as a byproduct of business-as-usual.
Technological Innovation and Consumer Adoption: In 2015, two enormous, consumer-facing tech canaries entered the health care tech coalmine, signaling health care as the new tech goldmine. Apple introduced ResearchKit, an open-source framework that offers an IT Rosetta stone for use by researchers and developers to create and share apps that facilitate the process of medical research. Google’s Baseline Project uses mobile technology to collect exhaustive health and genetic data from a single controlled set of healthy subjects in order to “mine” the practices and genetic patterns that promote optimal health. The entry of these giants into the healthcare sector signals an explosion of applications, wearable tech and new uses for mobile devices in the sector.
The Personalization of Data: Perhaps the most disruptive quality of machine learning is the ability to personalize as well as generalize. Data collection through EMR, lab results, mobile- and wearable-devices, presents a highly individualized and almost infinite pool of information ripe for analysis and action. Like Netflix, which observes user behavior in order to serve viewers increasingly targeted viewing selections, machine learning algorithms have begun to interact with different health care ecosystems and are responding with decisions and insights that drive value. Machine learning applications also aggregate individual data to create population “snapshots” that are much more reflective of an actual group than generic “reference data.” Look for machine learning to massively reshape corporate wellness programs as well as decision-making by Accountable Care Organizations.
Machine Learning in Healthcare
To better envision this future, let’s begin with several examples of machine learning already at work in health care:
Text2Move: In the field of diabetes management and prevention, the dynamic messaging program Text2Move hinges on machine learning technology. The algorithm uses four different data sets (self-reported motivation, continuous activity data, rudimentary location data and local weather data) to deliver personalized messages to each user. Research conducted using the TTM program showed a decrease in diabetes marker HbA1C equivalent to that achieved by anti-diabetic agents like metformin. The texting system is active for sunscreen use and smoking cessation as well.
Jintronix: This computer software package uses Microsoft Kinect motion-tracking technology and visual feedback to create in-home or rehab-based physical therapy. Machine learning and a gamification component promote patient adherence, while real-time objective data generated from patient-interaction is transmitted to their clinician.
Ovia: Designed to track numerous variables about a woman’s lifestyle and reproductive cycle, the algorithms driving Ovuline’s baby-making application Ovia gain power as a woman continues to use the application, and the app uses that information to provide an increasingly accurate prediction of her “fertile window.” In addition, Ovia uses machine learning to target coaching and content selected specifically for the user. Interestingly, Ovuline has partnered with Blue Cross Blue Shield to connect users with information about benefits.
Deriving the Real Benefits of Machine Learning: Bespoke vs. Off-the-Shelf
Machine learning, as hinted at previously, derives power from the ability to personalize as well as generalize, and as such can create synergy in many different ways. Using the Ovia example, the constant high-volume touches the app facilitates with the user creates value for the user, who is served a selection of information and activities tailored to their preferences and hopefully achieves pregnancy sooner than she would otherwise. The app creates value for BCBS because it gains brand equity through a personalized relationship with their clients while at the same time promoting a healthier population base.
The ability to measure objective improvements in health as well as cost-savings is also built into algorithmic businesses. As platforms continue to evolve, machine learning delivers what promises to be the biggest driver in the digitization of health care.
The BCBS partnership with Ovuline hints at further, profound implications that personalization carries for organizational healthcare cost and savings. The recent passage of the Affordable Care Act (ACA) changes the landscape of healthcare delivery in ways that are only beginning to become clear. First, the ACA shifts reimbursement incentives away from volume and toward value, and healthcare systems are now responsible for delivering quality outcomes at lower costs. Second, that same ACA is shifting consumer priorities, because high-deductible plans now comprise about 25% of all new health plans. This means being healthier will translate directly to consumer savings. Employers, too, are more interested than ever in controlling employee healthcare and pressure has increased for human resources and benefits programs to adopt data-driven, evidence-based practices for employee wellness and benefits.
One thing is certain; the kinds of questions the savviest organizations will be asking about wellness will be informed by the use of machine learning. Imagine, for instance, the power of machine learning in the sector of corporate wellness, a field that is dogged by the current state of lab testing (finger-stick) and lack of user engagement. Could comprehensive screenings identify more than just high cholesterol and high blood sugar, and instead provide meaningful information about a broader scope of risk? Could the results of health testing facilitate unique and custom-measured health information designed to motivate through specificity? Is there a platform that can leverage and learn interaction with participants and deliver, like Netflix, increasingly targeted content? Will this platform be able to demonstrate which components of a program are getting the most traction from participants? These are the questions that savvy business executives will be asking in the near future, if not now. Machine learning will have the answers.
Creo was founded by an innovative, cross-disciplinary team of laboratory diagnostic, business finance and health care management veterans. Contact us to learn how our advanced blood testing, precision analytics, health coaching and technology build employee plans that are personalized and effective—delivering real health improvements with measurable reductions in long-term health plan costs to employers.