Predictions Aren’t Fate, They’re a Foundation for Decision-Making

In Philip K. Dick’s 1956 novella, The Minority Report, “precognition” has advanced to a point where criminals can be identified before they commit a crime, reducing felonies by 99.8% and making violent crime a thing of the past. However, when police commissioner John Anderton — a man who has dedicated his life to preventing crime and protecting the citizenry — is predicted to be a murderer, the logical fallacy inherent in precognition becomes personal.

Can you punish an individual for something predicted, that they haven’t yet done?

Predicting the future has long been fodder for science fiction, but recent leaps in technology and artificial intelligence (AI) are changing the fantasy of “precognition” into reality.

AI, with the help of machine learning, is now able to churn through massive amounts of data at lightning speeds. In the management of workers’ comp and auto no-fault claims, we can use that information to predict likely outcomes for injured persons with incredible accuracy, very early in the life of the claim. Our predictive analytics use factors from treatment — including surgeries, medications, and dosages — along with patient-specific factors like age, weight, and comorbidities, and even population trending information, such as the individual’s zip code. All of this information, analyzed together, enables us to predict whether a claim will likely become high-risk, high-cost, and long-term.

While these predictive capabilities can be extremely valuable when managing a claim, many high-risk claims do not exhibit aberrant behavior at an early stage; that makes early intervention near-impossible. With limits to the amount of claim information that can be analyzed, and contributing factors that are uncontrollable in claims management, payers have few options to mitigate risk beyond increased oversight. It’s impossible to correct behavior that has not yet occurred.

We’re seeing that AI models are continuing to evolve to not only predict risk, but to calculate the exact moment in the life of a claim when appropriate intervention can drive maximum value. This includes ensuring the best care for the injured person, shortening the total claim duration, and avoiding unnecessary costs. By applying analytic models throughout the claim lifecycle and using machine learning to continually improve the accuracy of the models, claims professionals and payers can be alerted just at the moment when changes can be implemented, enabling the highest likelihood of overall success.

However, even with technological advances, human claim oversight is critical. The claims professional needs to understand the specifics of the injured person’s injury or illness and factors influencing recovery. Payers should work closely with their PBM partner to develop alerts and interventions that drive the best care for their injured population and align with their business goals. Then, they can define and tailor their program to streamline workflows and minimize “busy work,” so their claims professionals can consistently focus on what matters most — effective and compassionate claims management.

In our partnerships with payers, we use proprietary predictive analytics to analyze their claims and identify the highest areas of clinical risk. Using AI and client-specific clinical analysis, we then configure clinical intervention programs to target high-risk claims and key risk factors to help payers meet their program goals.

We also work closely with payers to implement intervention strategies that work within their existing workflows: alerting specific staff based on the type of risk, working with external intervention partners, or using Optum internal clinical expertise to intervene on the payer’s behalf.

AI is revolutionizing the way we identify and evaluate risk. But injured person care requires a careful balance of science and art. At Optum, we’re pairing the massive advantages of sophisticated risk prediction with continual evaluation of the whole person, using reason and compassion. As we continue to move forward with AI advances, we’ll keep in mind the words of Philip K. Dick, “Predictions may guide us, but they do not define us … The power to change the future lies within us; let’s use it wisely.”