New ways of segmenting patients lead to better care

  | November 20, 2017

A cyborg and an ostrich walk into a doctor's office. That's neither a joke nor a sci-fi movie. It's a growing trend in population health: Understanding behavioral patterns of patients beyond gender, age, and disease state.

In the case of the cyborg and the ostrich, both are healthy women in their late 50s. The first is an educated woman with a knee replacement. She skips preventive screenings because of her busy schedule, but is quick to see a specialist. So, her primary care physician knows that to get her in for a mammogram, she will need extra reminders – probably via text messages as the cyborg manages her calendar on her smartphone.

The ostrich, on the other hand, sees her primary care doctor regularly, but chronic anxiety means she avoids mammograms and other recommended screenings. Closing the gaps in her care will require in-person conversations with her physician and care team.

These patient personas, developed by Decision Point Healthcare Solutions, are just one new way to segment patient populations — an evolving trend at the cutting edge of population health.

As health systems take on broader risk, conventional approaches to assessing risk no longer suffice. Instead, patients' highly variable lives and behaviors define their risk states. And providers say improved segmentation — however it's done — can improve quality metrics, target resources, and drive better outcomes.

Here's why: The top and the bottom of any set of quality benchmarks may sit a stubborn 15 points apart. That “flat part of the curve" is where segmentation makes a difference, says Troy Long, M.D., executive consultant on population healthcare for PMG Healthcare Solutions, a physician-led consulting firm created by Colorado Permanente Medical Group.

“Moving that a smidgen — based on whatever criteria you use — is really important to do," says Long, who is also an internist at CPMG. "If you've been at the flat part of the curve for a while, then you need to start segmenting. Providers need to know patients' personas and tailor care accordingly."

Beyond conventional categories

Regardless of the method, new approaches to segmentation represent the broader pivot to patient-centered care. At Decision Point Healthcare Solutions, segmentation is based on everything but disease state, says Andrew Whitman, VP of customer success.

“It's a combination of the following distinct dimensions: Predicted behavior, barriers to engagement, persona, and channel preference," he says.

The Deloitte Center for Healthcare Solutions proposes segmentation based on psychographic methods of consumer marketing: how patients view themselves and how they navigate the healthcare system. The company identifies six categories of patients: “content and compliant," “sick and savvy," “casual and cautious," “online and onboard," “shop and save," and “out and about."

Applying those segments to, say, the notion of treating cancer in women could lead to different approaches to engagement — and even different treatment plans. The "casual and cautious" don't see the need for mammograms, while "out and about" patients seek alternative treatments, often without the knowledge of their doctors. A patient population thick with "out and abouts," then, might welcome the onsite integrative therapies that Massachusetts General Hospital offers its oncology patients.

Building bridges to better access

Where fine-tuned patient segmentation might prove most useful, practitioners say, is in designing care management for chronic illness. “Diabetes is rich for intervention," says Arlene Ash, Ph.D., chief of the Division of Biostatistics and Health Services Research at the University of Massachusetts Medical School, “because the difference between what happens when it's well-managed versus poorly managed is so great."

Segmentation, for example, can identify diabetic patients with housing issues — and who therefore may not consistently refrigerate their insulin, reducing its efficacy.

“We're trying to find out any issues patients might have, and how we could link them to community resources," says Lisa Swaim, a care coordinator with the Mission Health Partners accountable care organization in North Carolina. “Are they having any housing issues? Are they able to afford their medications? Looking at the whole patient, we're able to provide wrap-around care."

Novant Health, a multi-state health system, segments its 4 million patients according to income, education, and the number of children in a household, as well as clinical data. Its proprietary segmentation categories guide care managers in targeting resources to help patients overcome barriers to care,” says Jennifer Garber, manager of public relations for Novant. 

Removing barriers for cultural subgroups

Defining cultural subgroups is another growing segmentation strategy. Take the issue of colorectal screening, a key quality metric and component of preventive care. In many cases, says PMG Healthcare Solutions' Long, “the male Hispanic population will not do colorectal screening. But with a Spanish-speaking patient, you will always get better results with a Spanish-speaking doctor."

Acknowledging that problems with patient engagement can sometimes best be segmented along lines of ethnicity, nationality or race “can be very uncomfortable for people," Long says.

But recognizing cultural differences, and managing them with sensitivity, can lead to improved outcomes. Esperanza Health Centers, a federally qualified health center in Chicago, switched to the fecal immunochemical test, or FIT kit, for at-home testing to reduce disparities in screening among its Hispanic patient population. That approach has increased colorectal screen rates to more than 70 percent of patients requiring the test while lowering the cost of care –all arguments that cultural segmentation can result in better care. 

John Rossheim is a Rhode Island-based journalist.

New ways of segmenting patients lead to better care