5 Reasons High-Cost Claimant Programs Only Catch 5% of Future Claimants

To hear the full conversation, watch the webinar.

Most employers already know that a small group of high-cost claimants drives a third of their healthcare spend, and most also already run a program meant to catch them. But Zoe Greenburg, Garner's Director of Product, joined me on our recent webinar to explain why those programs keep missing. Healthcare trend is heading toward 9.5%, and claims over $1 million have grown 70%.

When we polled the room, most attendees said high-cost claims drive between 40% and 80% of their total costs. Yet Greenburg's data shows that the industry-standard method for spotting future high-cost claimants only identifies about 5% of them, meaning the other 95% never hear from anyone until the money is already spent.

In our conversation, there seemed to be five reasons so many high-cost claimants slip through the cracks, starting with the data employers never get to see.

1. Employers can only see the full picture for about 1% of employees

Predicting who will become a high-cost claimant takes four kinds of data, and almost no employer has access to each one:

1. Current claims: come from the carrier or third-party administrator.
2. Historic claims: show past diagnoses, the doctors a member has seen, and their prescription patterns, but that record is often incomplete.
3. Search activity: what a member is looking for right now. For example, a search for a knee surgeon points to a procedure they are already considering.
4. Doctor quality: According to Greenburg "A doctor's practice pattern tends to be pretty consistent, and so the practice pattern they use with their previous patients actually gives us a signal for how they're probably going to treat current and future patients." In other words, a surgeon who skips conservative treatment with most patients will likely skip it with the next one.

Today, employers have all four for only about 1% of their employees. For everyone else, any prediction runs on a partial record at best.

2. Traditional models wait for certainty, but by then the window has closed

Even with complete data, most programs do not act until they are certain a member is going to become a high-cost claimant.

"Traditional models are looking for definitive, 100% certainty signals that a member is a high-cost claimant or is likely to become a high-cost claimant," Greenburg says, and by the time those signals show up, "the window for intervention and cost avoidance has mostly passed."

She showed what that looks like for one member. Maura has had lower back pain for a year, searches for it, and gets an MRI. Nothing about her spend would trip a traditional threshold. Then she books a neurosurgeon and gets a spinal stenosis diagnosis, and only now does a traditional program notice her. "Maura's already found a doctor," Greenburg says. "She might not be as compelled to change her mind because she already found a provider that she trusts." 

Greenburg traces Maura's full path on the webinar, including the complication rates that separate the surgeon Maura picked from the Top Provider nearby.

3. The people you most need to reach trust the medical system least

Even a program that spots a member early often runs into an even more difficult problem. Members who use a lot of healthcare trust their physicians and the medical system much less than the general public does, so they meet any proposed intervention with skepticism.

"It's unlikely that a patient struggling with a specific diagnosis will trust a generic piece of outreach or a technology tool they're unfamiliar with to help them choose their doctor," Greenburg says.

Those members ask a friend or a family member who had the same condition instead. Outreach only works when it proves it understands the person, and Greenburg showed the actual message thread her team sent Maura. It refers to her visits and her searches, and it tells her plainly that staying with the surgeon she picked could mean a longer recovery because of complications.

4. Only about 4% of employees use the tools designed to reach them

When we polled attendees on their plans for high-cost claimants, 57% said they were pursuing enhanced screening, condition-specific point solutions, or stop-loss coverage. Every one of them depends on an employee choosing to engage, and Greenburg's data shows only about 4% of employees engage with doctor search tools or point solutions. She expects the rate is even lower among members dealing with a complex condition.

"Legacy navigation tools lack personalization and data that's persuasive to a member who's dealing with a very complex and potentially scary situation," she says. A member in that spot goes where someone credible tells them to go, not to an app they have never opened.

5. Centers of Excellence start after the decision that matters

I asked Greenburg why an employer would choose this approach over a Center of Excellence, and her answer came down to timing. "If you're sending a patient to a Center of Excellence, you already believe that that person's going to need surgery," she says. So the decision a program most needs to influence, whether surgery happens at all, is already made.

A Center of Excellence also covers only the conditions where one exists, while most high-cost claimants are managing several conditions at once. Greenburg's data shows the majority also have common conditions like hypertension, musculoskeletal problems, high cholesterol, and diabetes.

"Solving for rare and expensive one-off interventions doesn't fully address the complex healthcare situations of many of these patients," she says. Getting someone to a best-performing doctor in their own community earlier in the journey can avoid the procedure altogether, and it works for any condition.

What all five have in common

What struck me listening to Greenburg is that none of these failures come from a lack of effort. Managing high-cost claims is the top healthcare priority for 90% of employers. The programs fail because they start after the member has decided and speak in a voice the member has no reason to trust.

Garner built Predictive Outreach to fix both problems. Its models track more than 400 data points across claims, searches, and demographics, flagging members before a definitive diagnosis, and a care navigation team then reaches out with a message built around that person's history. Today, the program identifies about one in three members who will go on to spend more than $50,000 in the coming year, about six times what traditional programs catch. As Greenburg puts it, "only by intervening early are we able to actually avoid costs before they happen."

Our webinar also covers how Garner builds a separate model for each type of cancer, what fully insured employers get out of the program at renewal, and how stop-loss carriers treat groups that have it in place.

Watch the full session to hear it in her own words.