Wednesday, January 18, 2012

Predictive Modeling: Prioritizing Vision Over Execution

When Lou Gerstner took over a struggling IBM in the summer of 1993, he famously commented, “‘There’s been a  lot of speculation as to when I’m going to deliver a vision of IBM, and what I’d like to say to all of you is that the last thing IBM needs right now is a vision.  What IBM needs right now is a series of very tough-minded, market driven, highly effective strategies for each of its businesses.”
Well said.
I gave an interview last week and the topic of discussion was predictive modeling and its ability to help “fix” the problem of prescription drug abuse in work comp.  Could sophisticated software that identifies potentially problematic claims before the patient becomes tolerant, dependent, or addicted really help mitigate this difficult, multi-billion dollar issue? 
As the discussion unfolded, I couldn’t help but think of Mr. Gerstner.  The essence of his early days at IBM was a focus on execution.  He recognized that spending time and money on developing a “vision” created a built-in organizational excuse for lack of execution. When it comes to problem solving, the practical should always trump the theoretical. 
I absolutely believe there is a place in worker’s comp for predictive modeling.  Smart people are finding ever smarter ways to more effectively manage claims.  But when it comes to the issue of prescription drug over-utilization, let’s not let the theoretical trump the practical.  I talk to claims executives constantly and, for the most part, there exists a clear understanding regarding the identification of complicated claims and an intuition regarding claims that may end up being complicated. 
So we can talk about how predictive modeling can help.  We can even wonder aloud about whether, in the future, these models will be the best possible means by which to manage claims. 
But first, let’s put some prescription data in a spreadsheet, pick all the claims with Actiq, and get on the phone with the treating physician to figure out how to fix it. 
We don’t need a vision for that.  We just need to execute.
On Twitter @PRIUM1

No comments:

Post a Comment