Eye on the exchanges: What predictive analytics tells us about drivers of health plan profitability

As insurers submit initial filings about how they plan to price their Affordable Care Act exchange policies next year and where they'll offer them, both current political uncertainty and past profitability will factor into in their decisions. 

But politics aside, what exactly are the factors that led some insurers to thrive in the individual marketplaces while others lost millions of dollars? That’s precisely what Syed Mehmud, an associate of the Society of Actuaries and senior consulting actuary at Wakely, set out to uncover.

For two years in a row, Mehmud has conducted a study to determine the factors that drive financial success—and struggles—on the exchanges. The most recent results are based on 2015 EDGE server data covering about 5 million lives, which insurers supplied in exchange for insights about how their exchange business is faring relative to their peers.

In order to glean insights from so much information, Mehmud and his colleagues wrote their own algorithm geared toward parsing large-scale medical data and created an adaptation of a popular predictive modeling technique called decision premodeling.

“What we wanted to do was to analyze the information and see if we can find any patterns, some common threads,” Mehmud told FierceHealthcare.

The patterns they did find may be surprising to some. Here’s a look:

  • Sicker patients aren’t necessarily driving higher costs. Thank the federal risk adjustment program for that, according to Mehmud. What makes the program so powerful is that it’s a concurrent model rather than a prospective model, he said, meaning it moves more money around among issuers. The greater the number of patients with costly conditions—meaning higher risk scores—a health plan has, the more money it gets from the risk adjustment program, and vice versa. Thus, “sometimes it is actually the healthier patients and the younger patients that are more unprofitable, because they have fewer medical conditions,” Mehmud said.
  • Not all metal levels are created equal. Plans with richer benefits, like gold and platinum, struggled in the individual marketplaces in 2015, Mehmud noted, but oftentimes thrived in the small-group market. That goes to show “that the dynamics of the individual and small-group markets are very different,” he noted.
  • Contracting plays a key role in profitability. In Mehmud’s project, he repriced claims to a percentage of the local Medicare rates in order to compare across the board what individual market insurers were paying for services. Not surprisingly, he found a considerable spectrum of contracting rates—and more profitable health plans tended to have negotiated better rates. While provider contracting doesn’t get as much "air play" as other factors, the study showed that it’s an “important ingredient in the success of an ACA individual market participant, and carriers should pay attention to that,” Mehmud said.
  • Command of data can make or break a health plan. Unlike in Medicare, where plans had some time to get used to risk adjustment as the program ramped up, in the commercial market, “a switch turned on,” Mehmud noted. So the carriers that weren’t paying good attention to their data did not fare well in 2014, and that pattern continued in 2015, with “tens of millions of dollars left at the table” among carriers that failed to adequately code for risk. Health plans, Mehmud concluded, “doesn’t necessarily win by making sure that their data is correct, but definitely lose if they don’t.”

One caveat to the patterns that Mehmud observed, however, is that they don’t apply to all insurers. Therefore, he recommended that insurers study their own data to determine strategy shifts tailored to their individual organizations.

“Every carrier has a different story to tell,” he said.