Premera Blue Cross is betting that new technology can predict future health problems, thereby giving patients and medical providers the ability to react before potential issues become severe.
The large Seattle-based health insurer in February announced a partnership with Cardinal Analytx Solutions to use Cardinal’s Cost Bloom predictive modeling program to identify Premera members with the highest probability of needing high-cost health care over the coming year.
It’s generally understood in the health insurance industry that a small portion of any population accounts for the vast majority of health care costs associated with that population. According to a Premera paper on the program, research at Stanford University found that approximately 10 percent of an insured population typically accounts for 70 percent of the costs.
The challenge is in predicting who will be in that 10 percent pool in a given year, Premera Data and Analytics Director Colt Courtright said in an interview.
That’s because, according to Cardinal Analytx, 60 percent of the high-cost pool changes year-to-year, and most analytics programs focus on identifying and providing managed care to only the 40 percent long-term portion of the high-cost pool.
Cardinal Analytx Solutions is a Palo Alto, California-based data analytics firm.
“Really, what we’re trying to do is help people avoid those (high health care) costs in the first place. The challenge using classical statistics is that it was impossible to find a large portion of that 10 percent and be able to predict who they might be, so you couldn’t really intervene,” Courtright said. “You couldn’t offer support programs; you couldn’t perform outreach; you couldn’t encourage provider visits.”
The key to the program is employing artificial intelligence with the ability to parse out much more subtle indicators of future high-cost health care users. When a potential high-cost individual is flagged, Premera can then notify that person and suggest preventative or early treatment methods.
The artificial intelligence can identify patterns in members’ use of medications, or a constellation of health conditions and discern if a social support program, for instance, could improve the condition before a major procedure or other intensive care is required, Courtright said.
Cardinal Analytx estimates the Cost Bloom program can result in 15 percent savings across an insured group over two years.
The condition forecasting is an addition to Premera’s existing clinical care management and care coordination programs. When a member is identified as someone who is likely to need high-cost care in the next year a case manager can reach out through those programs, according to Courtright.
He also said the predictive modeling works using data insurance companies have traditionally gathered; however, the artificial intelligence analysis of that data is driven by the interaction of more than 50,000 data points or variables processed through a predictive algorithm, according to Premera.
“It’s less about a specific data point as these are attributes that are often combined across data points,” Courtright said.
Premera Blue Cross Blue Shield Alaska is the lone insurer in Alaska’s individual health insurance market.
A reinsurance program first started by the State of Alaska in 2016 and then approved by the federal Centers for Medicare and Medicaid Services in 2017 has allowed Premera to reduce its individual market insurance rates by 26 percent in 2018 and 6.5 percent in 2019.
The Alaska Reinsurance Program is in the middle of receiving $332 million in CMS grants over five years to support the program.
Premera returned $25 million in reinsurance money to the State of Alaska in late 2017 after the company determined there had been a significant reduction in the use of medical services by members in the individual market.
Elwood Brehmer can be reached at firstname.lastname@example.org.