January 3, 2013 
Certainly, the push around Big Data should drive more companies to look into predictive algorithms. You already have Fair Issac (above), ScreenRx by Express Scripts, and RxAnte.
I’d always been under the impression that women were less adherent than men to prescription drugs. I’d heard several very logical reasons:
- As the caregiver, they often took care of their children, spouse, and/or other people first before being adherent themselves.
- They took more medications on average which is highly correlated with non-adherence.
But, some researchers recently told me that their data showed women to be more adherent. And, I then noticed that the infographic that Express Scripts put together showed something similar. (see zoomed in picture)
So which is it? I can find lots of research online to support females being less adherent. Here’s a few links:
- http://www.mdnews.com/news/2012_03/women-are-prescribed-a-greater-number-of-medications.aspx
- http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3251481/
- http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3251481/
BTW…If you want to see a good presentation on some adherence data from CVS Caremark from a few years ago, you can follow this link to a presentation that was given. Here’s a more recent one from URAC.
Enabling Healthy Decisions 
I just realized the one CVS Caremark link was broken. Here’s the deck that Jack Bruner presented a few years ago.
http://www.nebgh.org/pdf/presentations/bruner072210.ppt
Nice post, George. You raise an interesting question.
What we’ve found at Express Scripts is that no single attribute is all that predictive of adherence; instead, it’s the constellation of many of these factors working together that really provides actionable insight into which specific patients are likely to stop taking their medications in the future.
If you look only at gender, for example, women as a group are slightly more adherent than men. However, women with certain attributes (e.g., mid-20s, with children in the household) are some of our least adherent members. In the big picture, gender is only a small piece of the predictive puzzle.
What sets Express Scripts’ models apart from others in the industry is the detailed level of data we use. While other organizations may look at 6-10 personal attributes to get a ballpark predictor of adherence, we consider more than 400 factors that reflect the entire adherence equation – including the patient, the physician, the pharmacy, the disease and the drug.