BOSTON—Blue Cross Blue Shield of Nebraska set out to use its analytics tools to identify those patients who consistently drove high costs but didn’t show up on any other analytics efforts. Susan Beaton, senior director nurse care management and clinical policy, shared the organization’s experience during a session at the Big Data & Healthcare Analytics Forum.

Although nurses spend most of their time navigating the healthcare environment, they only reach about 1 percent of the population—people in acute crisis. “Those people are more than happy to engage with nurses,” said Beaton.

But, the high spenders were difficult to track. These were patients who didn’t go to the ED six times and weren’t on high-cost drugs but over time their spending was still high. “They were almost invisible to us,” she said.

The organization worked with its vendor to build a population-based strategy to figure out how to identify these high-cost members. They wanted to analyze the population, develop an approach and an intervention. “We looked at all the data to figure out to predict the person who is going to be a persistent and consistently high spender. Claims data only gets you so far. We needed to figure out how to incorporate the live, real-time data we had at our fingertips.”

Analysis revealed that they couldn’t reach 32 percent of their population. They knew the high spenders had to exhibit traits that make them persistent utilizers of the system. They already were reaching those who had malignant and catastrophic conditions as well as those with multiple chronic conditions. Beaton said they learned that the traditional practice of breaking down the population by condition wasn’t working because it is behaviors that drive utilizers of healthcare.

BCBS of Nebraska launched a pilot to make data meaningful and actionable to nurses. That required redesigning existing care management programs. This took longer than it did for their vendor to pull all the data, which Beaton said was probably the biggest hurdle.

One patient from the pilot had 42 different interactions with the healthcare system within six months. She went to work every day and saw her doctor as scheduled. This proved to be a huge barrier because these people thought they were healthy. “We knew we had to help people change their behavior and if they continued to really need all those services we had to figure out how to drive them to the right care at the right cost.”

Beaton’s team started almost three years ago with 516 members and $35 million. The organization’s leadership waited a year before they let her expand the program to 800 members and $55 million. That year of no intervention cost another $20 million, she said. “We have an obligation to help them control those costs.”

It takes a special skillset to reach out to this population, they learned. They retrained and rehired several different types of nurses to get those with motivational interviewing skills and behavior change training. Beaton had to prove to leadership that the predictive model was worth it by mapping the data to the 1.6 million member database that had received traditional care management outreach.

The model could predict at a higher percentage rate for those chronic users of care services. They run it monthly, she said, and give the information in a dashboard to nurses. They developed the BlueCares outreach program, which goes beyond phone calls to a two-way mobile messaging solution to send out reminders and check in on patients. As soon as a nurse sees that a patient went to the emergency room, she can send a message.