What if “electronic health records” could become “electronic health oracles” – not just recording the past, but helping to predict and influence the future?
Thanks to a proliferation of data, “electronic health oracles” are well on their way to reality.
In his book Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, Eric Siegel details how technology can be built to pore over millions of records (data) to predict the future behavior of individuals in ways that drive better decisions. The examples astound. From retailer Target identifying customers most likely to become pregnant to banks averting theft through fraud, colleges retaining students on the verge of dropping out and the Obama campaign mobilizing persuadable voters (and ignoring the rest).
The implications for improving health are profound. Predictive analytics allow the use of electronic health records and other data sources to consistently and efficiently:
- Target areas with environmental factors like air quality and food deserts that predict vulnerability to specific conditions
- Raise awareness among people with risk factors about how to reduce their vulnerability
- Tailor guidance to avert health problems before they start
- Give timely, personalized information to patients who have specific conditions
- Reach out to people due for check ups and tests with condition-specific information
Data Teach Us
The success of prediction depends on the quality of data used in the process and the predictive models that are applied to make sense of that data. Tech companies like ours essentially task computers to learn from any given data set. By encoding combinations of data points, patterns are revealed that explain or correlate to specific results: if xyz, then abc.
A recent study found pediatric asthma patients ages 10 to 17 improved their health when they received daily texts with questions about their symptoms and information about how to manage their asthma. The results are encouraging, but communication could be improved even more with a deeper look at the data. What difference does air quality make? Do messages during certain times of day improve asthma control? Does wording make a difference? Do boys and girls or different age groups react differently?
Data from social media sources like Facebook and Twitter, as well as wearable sensors and fitness apps can also provide key direction for delivering more personalized, precise health care.
In fact, recent research revealed that Facebook provides a pretty accurate reflection of obesity rates around the country. Researcher John Brownstein, Ph.D., of the Boston Children’s Hospital Informatics Program said the findings suggest huge opportunities for population health management:
Online social networks like Facebook represent a new high-value, low-cost data stream for looking at health at a population level … The tight correlation between Facebook users’ interests and obesity data suggest that this kind of social network analysis could help generate real-time estimates of obesity levels in an area, help target public health campaigns that would promote healthy behavior change, and assess the success of those campaigns.
The more data, the more challenging the predictive model is to create, but the greater the payoff. The “Health Information Exchanges” (HIEs) and “Health Insurance Exchanges” (HIXs) set up under the new federal health care law will generate galactic amounts of data. HIEs represent the electronic sharing of health-related information among organizations. HIXs are new organizations set up in each state to provide competitive “one-stop shopping” for affordable health insurance coverage. Bringing these two data streams together could tell us a lot about how to improve individual and population health.
The success of the exchanges will depend on the systemic capture of health data that can answer critical questions about quality. The URAC STAR Data System – a technology platform for accrediting health plans participating in state health insurance exchanges – will capture data from insurance claims, pharmacy transactions, and provider charts. Each will tell an important part of the quality story.
The data portal Pantheon created for URAC will track health plan performance across 38 quality measures, including cancer and diabetes screenings, obesity and tobacco use prevention and measures related to depression and heart disease. Eventually predictive analytics could unlock a host of ways to get better results in health care by following results for specific conditions and providers. For example, we could learn what works best to effectively address cardiovascular disease and obesity.
For those toiling in the underappreciated field of “population health,” it’s an exciting time. Few realize that 80 percent of population health can be attributed to personal biological factors like genetics and social determinants of health, such as education, housing, safety and sanitation, with healthcare accounting for only about 20 percent of the influences on population health.
With data and predictive analytics in hand, healthcare can be much more proactive and precise, no longer waiting for patients to either become conscientious about their health or so sick they need costly interventions.
Imagine NASA-inspired mission control centers signaling “hot spots” for obesity, heart disease, diabetes, depression and more. With more real-time, personalized data from a range of sources, generic prevention messages like “stop smoking,” “exercise” and “lose weight” become personally relevant.
With the advent and use of “big data” – population health management is fast becoming the hottest field in health care and key to solving our most intractable problems.