Are Predictive Analytics the Missing Link in Maximizing Your Claims Outcomes?

Predictive analytics can be the key to keeping minor injury claims from becoming costly creeping catastrophic claims. Through machine learning, these analytics can provide cost savings, faster claims closure and less litigation. But achieving these results means focusing on the right types of data, as it applies holistically to the injured worker.

“There are lots of reasons why this technology is underutilized in claims,” said Steve Peacock, Assistant Vice President – Claims at Safety National. “Building predictive analytics programs can be expensive, and the customization can be very time consuming. But the beauty of this tool is that it can identify red flags early enough for the right resources to be implemented, thus avoiding negative trends and minimizing claim exposures.”

Psychosocial Indicators & Comorbidities

Analytics that identify psychosocial indicators coupled with intervention strategies have been recognized to be very effective. “Psychosocial” refers to a patient’s psychological and social issues that can negatively impact or delay recovery from the work-related injury. Examples of psychosocial indicators include:

• Does the employee have an addictive personality?
• Is the employee experiencing family stress?
• Is the employee dissatisfied with their job?
• Does the employee have outstanding financial or legal issues?

Intervention strategies specifically designed to address the identified psychosocial factors are an essential component for changing outcomes for the better.

Information around patient comorbidities, such as obesity, diabetes, hypertension and/or overuse or misuse of opioids for pain management, can also be very useful in claims management. It is optimal to identify which comorbidities are present in a claim up front and then determine the best way to manage the claim differently when they are identified.

Diagnosis & Treatment History

Medical data can be used to predict potential issues based on the type of injury, current diagnosis and treatment, treatment history and prescriptions. Some systems can provide an alert when an actionable item occurs. For example, if an employer is notified when an employee is first prescribed opioids, they could monitor the drug-dispensing process for signs of it becoming problematic. Examples of medical diagnosis indicators include:

• Is there a potential for a pharmaceutical dependency?
• What is the standard course of treatment for this injury?
• What have past treatment outcomes for this injury looked like?
• Who are the treating physicians and what are their projections for future treatment?

If data shows that the current course of treatment has proved unsuccessful in the past, the best course of action may be to intervene, coordinating with a nurse case manager to recommend an alternative and share predicted concerns with treating physicians.

External Factors

Factors like personal demographics and litigation data are often forgotten when analyzing claims data. An analysis of the employee’s household and zip code can provide insight into potential high-dollar costs related to home modifications, home healthcare assistance and how accessible treatment is for the employee. Examples of external factors include:

• Will home accessibility be an issue for the employee?
• Will a family member be providing care, or will an attendant be required?
• Are their treatment disparities based on where the employee resides?
• Based on the case type, judge and district, what are the typical outcomes?

Litigation data can be vital to potential settlement strategies where necessary, especially with the rise of nuclear verdicts, where juries award the plaintiff millions or even billions of dollars. Settlement should never be taken off the table, particularly in cases where despite your efforts, there are red flags indicating that the claim can develop into a creeping catastrophic claim.