If you are focused on service line growth, population health management, or profitability, having a method to hone in on target populations with precision is absolutely paramount. This can only be done using statistics.
Smart targeting relies on the creation of a statistical model, which is essentially a mathematical equation derived from a sophisticated analysis of a target population. Often, this is accomplished by overlaying a target patient group you desire to "clone" with hundreds of household- and individual-level variables. These variables are demographic, psychographic, and behavioral in nature, as well as proprietary variables generated based on your patient data. The model then serves to locate other individuals in your service area who match the target patient group with respect to these variables.
Simply put, creating a model requires smart people and sophisticated software.
Statistical models can be built to meet a variety of marketing objectives. Want to shift your overall payer mix towards commercial? There’s a model for that. Want to push your ED abusers/misusers towards PCPs and urgent care? There’s a model for that. Want to grow a specific service line while focusing on profitability? There’s a model for that, too.
So where do you start? If you're not equipped with the team or tools to perform these sophisticated modeling techniques, the first step is finding the right partner.
Five Criteria for Finding a Modeling Partner
A good modeling organization will do five things:
- They will discuss your pain points and work with you to identify clear marketing objectives. Clear objectives yield valuable, actionable models.
- They will thoroughly explain their overall targeting strategy and how it relates to your marketing objectives.
- They will share insights gained from hours upon hours in the trenches with your data.
- The modeler will do all the scoring and provide you the actual list. This ensures nothing is lost in translation throughout the process, from the creation of the model to choosing the households it has prioritized for you.
- The modeler will have done all this with the utmost integrity for protecting your patient information and complied with all HIPAA guidelines for handling such data.
And a Few Bits of Advice
Here are a few thoughts to keep in mind as you explore modeling options:
- A profile or a set of selection criteria (e.g. women age 45+, income $75K+) is not a model. A model understands that these variables might have weighted significance among a host of other variables, while a profile does not.
- Ask your modeler for case studies and results. There are good models and bad models; good modelers and bad modelers.
- Don’t skimp. By its nature, a good modeling exercise is a relatively significant investment, but well worth every penny in results.