Cropped_new_image

Insights from GNS Healthcare

Failing Ethel: Are Your Interventions Blind to Individual Value?

Posted by Colin Hill on Aug 26, 2015 2:00:00 PM

In this era of individualization and value in health care, why are so few health intervention dollars invested on the basis of their value to individuals?

People respond to – and benefit from – interventions differently. The pharmaceutical and life science communities have been acting on these differences for many years through personalized, or precision, medicine. GNS Healthcare has been at the forefront of this movement since 2000 and continues to work with a range of world-class partners to advance data-driven solutions that are improving care for millions of individuals around the world.

Today, GNS is bridging care management and precision medicine to inform health plans’ and other stakeholders’ accelerating investments in population health programs. Whether applied to home-grown or vendor-based interventions, the patented GNS machine learning technology enables new collaborations by analytics and care management teams. It reveals, in advance, whether a given intervention has the power to impact each individual’s health by the magnitude needed to improve their outcomes. These insights allow health plans to optimize intervention spending.

This value-based methodology turns traditional rules-based thinking on its head. To improve medication adherence, for example, 80 percent adherence is the typical rules-based intervention threshold. When many, many more data points are part of the equation, a different picture emerges. Some of the highest-value opportunities are hidden by a strict focus on rules.

Meet Larry, Nora and Ethel, three hypothetical members who, like almost half of Americans, have at least one chronic condition.

Failing_Ethel

Larry’s 44 percent adherence level is far below the 80 percent threshold. Nora’s adherence level is an alarmingly low 29 percent. Ethel, with an adherence level of 82 percent, is the only one above the rules-based threshold.

Both a rules-based and a value-based approach select Larry for an intervention, but that’s where the similarities end. This is because the underlying methodologies are so divergent.

The most sophisticated rules-based big data analytics programs examine large datasets to predict an outcome. Instead, GNS identifies the causes of predicted outcomes so that something can be done to prevent a poor outcome from becoming a reality.

The GNS patented machine learning platform simulates every possible combination and every possible outcome from large, heterogeneous datasets. This causal analysis creates new knowledge, often characterized by many, many trillions of data points.

We’ve already seen that the rules-based predictive modeling and value-based machine learning approaches can agree, as they do with Larry. So, how does machine learning make a difference in population health strategies and in the lives of Nora and Ethel?

Prevents wasteful spending on interventions that have little or no chance of helping people, such as Nora, improve their health. With a 29 percent adherence level, a rules-based predictive modeling program, of course, would flag Nora for intervention. The value-based approach, however, determines that it would take a 45 percent improvement in Nora’s adherence to prevent less than $200 in healthcare costs stemming from her low adherence level.

Reveals hidden opportunities to improve health, such as with Ethel. The value-based, machine learning-enabled approach reveals that, with just a 10 percent improvement in her already good 82 percent adherence level, Ethel can prevent an emergency room visit or a hospitalization predicted to cost more than $10,000. A rules-based approach, of course, does not select Ethel for intervention because her adherence level already exceeds its 80 percent threshold.

The rules-based approach fails people like Ethel by ignoring the potential value intervention would bring to her. Instead, it leads us to spend intervention dollars that have little to no chance of improving health -- on people like Nora. Adherence is important for Nora too, but health plans can no longer afford to be blind to the hidden value in their own data.

The opportunity for healthcare organizations to optimize intervention investments is enormous. Medication adherence is just one example. As a leader in the precision medicine movement for 15 years, GNS has pioneered big data-driven solutions in premature birth, multiple myeloma, Huntington’s disease, metabolic syndrome, diabetes, asthma, breast and colorectal cancer, multiple sclerosis, rheumatoid arthritis, childhood asthma, and others.

What effect does a value-based approach have on a health plan’s bottom-line? We’ll explore that question in our next blog post.

Colin Hill is chairman and CEO of GNS Healthcare. He can be reached at [email protected]

 New Call-to-action

Topics: Precision Medicine, Pop Health, Machine Learning, Health Plan, Payer, Medication Adherence, Care Management, Value Based Analytics

Subscribe to Email Updates

Recent Posts

New Call-to-action