Improving The Decision-Making Process Through Behavioral Analytics
By Dave Hom, Chief Evangelist, SCIO Health Analytics
If you were to look at the answers people provide when taking a written driver exam, you would see them say they should drive at or below the speed limit, slow down when the traffic light turns yellow, come to a complete stop at a stop sign, and not pass a school bus with its stop sign extended. Yet, as 41 million annual moving violations for speeding alone will attest, those answers may be more convenient than truthful.
This goes to prove the old adage “actions speak louder than words.” No matter what people say they will do — or what well-meaning experts believe they will do — there is no better indicator of what actions people will take in any given situation than the actual behaviors those people display when faced with a decision.
This is an important lesson for any organization to understand, especially now that we have the means to go beyond intuition, experience, and gut instinct. But it has been particularly beneficial for hospitals and healthcare organizations as they work on the patient engagement strategies that are a critical component of the transition to a care system measured on outcomes (value-based care) as opposed to services rendered (fee-for-service).
Under this new reimbursement method, driven by the Centers for Medicare and Medicaid Services (CMS) for Medicare and Medicaid members, the challenge isn’t what care to prescribe. It’s getting the patient to follow instructions such as eat more grains and vegetables, exercise more to manage their health, and so on. That way, when they get older, they can see their son or daughter walk down the aisle.
Because no matter how brilliant or logical a strategy may seem in the abstract, if it doesn’t translate into the desired behaviors from the people being targeted it’s not going to have the desired effect. Here’s an example of how behavioral analytics are being used in healthcare to improve patient (and population) health, reduce hospital events including admissions, readmissions, and ER utilization, increase patient satisfaction, and lower the cost of care.
Why Patient Engagement Has Become Important
Giving patients a specified plan of care and encouraging them to follow it is nothing new. What’s different is it has become crucial to physicians’ and hospitals’ compensation. CMS has talked about basing up to 60 percent of its reimbursement to physicians on patient outcomes, including paying a maximum dollar amount for knee surgeries to both the surgeon and to the hospital.
Under the old fee-for-service model, if patients failed to follow care instructions (such as making a follow-up appointment with their primary care physician or filling and taking prescribed medications) after being released from the hospital or emergency department, it would negatively impact their health. This lack of action could potentially lead to readmissions, especially among patients with chronic conditions. Yet from a purely financial perspective, readmissions meant more revenue for providers so the incentives to keep patients out of the hospital didn’t exactly line up with the economic model.
Under value-based care, however, readmissions or deteriorations in health are costly to the provider as well as the patient. With incentives now becoming aligned with performance, it is now in everyone’s best interests for patients to be actively engaged in their own care. This is where behavioral analytics become critical.
Measuring Actions To Improve Outcomes
As with any industry, when healthcare organizations build consumer engagement strategies they typically start with a few assumptions. For example, if they are developing a program to reduce pregnancy or sexually transmitted diseases in teens, the logical assumption would be to build an app for smartphones.
According to Pew Research Center, 73 percent of teens age 13-17 have access to smartphones, and 91 percent use the Internet on a mobile device with 92 percent of them using it at least once a day. That’s some pretty strong evidence on the surface.
Yet the reality may turn out to be entirely different. It could be teens in the target audience would prefer to receive this information in another way, such as in-person consultations, an email newsletter or a hard copy pamphlet they can pick up and read later, or having the parents use the information to educate them.
Before making significant investment in building the app and rolling it out to the entire population, a better plan would be to create a pilot and test it against other, more traditional methods of disseminating information. As part of the test, the organization can try including different incentives for taking actions. For example, if participants review a set of information and take and pass an online test, they can earn a reward. Analytics around the rate of redemption will show how important rewards are to this population.
The organization can refine this concept further by offering different types of rewards (such as direct cash payment, gift cards or points toward prizes) to the test groups to determine which are more effective at driving the desired behavior. Gaming has become a unique and popular way to nudge the patient to take a particular action as well.
The more understanding there is around the target population and what motivates it, the more effective the organization can be in designing programs that lead to the desired behaviors — and lead to downstream improvements in financial performance.
Behavioral analytics can also be used to improve engagement (and guide resources) in an existing program. Here’s how that might look in a diabetes management program.
The organization gathers information regarding the types of programs it is offering, such as care management (a nurse or other care manager contacting and guiding patients proactively), automated identification and notification of care gaps (such as a missing test), and group counseling sessions. It also collects data around self-study information such as how to improve your diet and the value of intense exercise such as 15 minutes of jumping jacks or squats in your kitchen. It then compares the usage rates for each of its program elements, along with the effect the element has had on individual and population health (and subsequently on in-patient stays, readmissions, and emergency department visits).
Armed with this behavioral data, the healthcare organization can determine which interventions are the most effective (and therefore worth greater investment) as well as those that are no longer worthy of continued investment, freeing up budget to develop new options or enhance existing ones.
By coupling this information with analytics showing the impactability (likelihood that a particular intervention will have a significant impact on patient or population outcomes) and intervenability (likelihood that the patient or population will become actively engaged in their own care), the healthcare organization can further refine its programs and maximize its chances of delivering the right care to the right patient at the right time.
A good example of intervenability is examining the effect having to make a left-hand turn against three lanes of traffic has on whether patients in Florida will pick up medications at the pharmacy. Elderly drivers, which Florida has a lot of, don’t like making left-hand turns in general, much less against three lanes of traffic. Routing them to another pharmacy, or better yet building the pharmacy in a location that avoids such issues, is likely to increase the percentage of patients taking their medications thereby improving health outcomes.
Improving Decisions Through ‘What If?’
Clearly measuring behaviors when determining how to engage consumers more effectively can have a significant impact on success (and avoid costly missteps). Why aren’t more organizations doing it?
The reason is that up until now they haven’t had the analytics to do it. In healthcare, the first generation of analytics was focused primarily on clinical outcomes. In other words, showing what happened in the past.
Next-generation analytics using medical and prescription claims provide the opportunity not only to dig deeper to determine which programs/offerings/strategies drove those outcomes, but also to start asking “what if we did things differently?” These predictive analytics are helping hospitals and healthcare organizations take advantage of the data they have about patient behaviors to improve their engagement strategies. Which then helps them create an even greater impact on both patient and population health.
Decisions With Precision
As anyone who has rolled a stop sign, driven on a painted median or tried to beat the stoplight by hitting the gas when the light turns yellow can attest, what people say they want or will do isn’t always the same as the actions they actually take.
Behavioral analytics can help organizations ensure the strategies they’re following for consumer engagement are delivering the performance they’re expecting, or help them make course corrections in a timely and cost-effective manner. They deliver one more vote of confidence that the organization can succeed in achieving its goals.
About The Author
David Hom is Chief Evangelist at SCIO Health Analytics®, an organization dedicated to using healthcare analytics to improve clinical outcomes, operational performance and business results. He is an internationally-recognized expert in the field of consumer engagement through programs such as Value Based Benefits and Employee Wellness. With more than 25 years in healthcare, Dave has been a visionary at SCIO®, building SCIO’s leading products on behavioral economics applications for many health plans and a technology-enabled solution to engage members based on the gap value and impact on avoidable hospital events. He can be reached at dhom@sciohealthanalytics.com.