There was a joke in the Soviet Union that asked - How can anyone predict the future when it's so hard to predict the past? While it was referring to the revisionist historical practices of the Communist Party, there is some truth about how the past changes.
Changing historical data presents a challenge to analytics.
For example, in health insurance there is a something called Coordination of Benefits (COB). It is a complicated set of rules that determine who pays what portion of a claim when a person has more than one health insurance coverage at the same time. Most health insurance companies are very effective at processing COB rules at the time a claim is submitted.
But what if the past changes? The existence of other coverage may not have been known when a claim was paid. At is that at a later date, the health plan learns of the other insurance and must apply the change retroactively.
- Claim systems are not 100% effective at finding all impacted changes. Claim systems are designed for throughput, not historical analysis.
- There are costs to reprocess a claim.
- The doctor or hospital will also incur costs reprocessing a claim and possibly cash flow concerns.
Advanced analytic techniques are applied to historical data to identify recoverable claims that may have been missed.
Machine Learning scores the likelihood a person may have other coverage. You can focus on members with a better chance of cost recoveries.
Over time, Lia will be able to predict the possibility a new member has other coverage.
To learn more on how Lia can help, contact us.
