Will Patient Reported Data Augment Claims Based Models?

On the one hand, it seems fairly obvious that patient reported data (use of OTCs, exercise, food intake) is important in understanding their healthcare.  On the other hand, the historical bias has been to use historical claims to predict future costs.  At a minimum, I think that studies around tools like PAM (Patient Activation Measure) have shown that patient reported information is important in understanding their literacy and attitudes on healthcare.  This data is critical in designing effective healthcare engagement programs.  [One of the reasons that Silverlink has stressed our focus on using data for segmentation and personalization for years.] 

That’s why I found one of the latest studies by Kaiser to be really important.  They used both claims data and patient reported data to evaluate inpatient admission rates and costs.  And, as explained below, this data increased the predictive power of their model. 

The research determined that self-reported information about being in poorer health was a key determinant in predicting higher inpatient admissions and for being in the top tier for costs. Higher admission rates and costs were associated with patients who self-reported:

• Lower score for general self-rated health
• Yes to “do you need help with one or more activities of daily living?”
• Yes to “do you have a bothersome health condition?”

The addition of this self-reported information to a claims history model explained an additional 2.8 percent of variance in admissions and 4 percent in cost.