Pre-Emptive Medication Adherence. Is Pre-Emptive DTC Next?

Computers are intruding into our lives more and more these days. I'm not talking necessarily about the Internet, but about automated programs that different industries use to do things like initiate stock market trades without human intervention, which is a story I saw last night on 60 Minutes. It's reputed that one such trade sent the market in a momentary tailspin that could have lead to a monetary crisis. According experts, such automated trades are undermining the general public's trust in the stock market. I for one have taken a lot of my retirement money out of stocks.

Today I read about how Express Scripts, a pharmacy benefit management (PBM) company, is now able to accurately predict up to a year in advance which patients are most at risk of falling off their physician-prescribed drug therapy -- and to intervene in customized ways to improve those patients' adherence. To do this, Express Scripts uses "a set of proprietary computer models" that analyzes personal data of patients in its database. The data includes such things as prescription history, whether the patient has kids living at home, etc.

"Previous industry attempts to predict therapy adherence were hampered by both the types and quantity of data available," said David Tomala, director of advanced analytics at Express Scripts. "Our tens of millions of members, hundreds of millions of annual prescriptions, and advanced understanding of human behavior were key to 'cracking the code' on therapy adherence. We are now the first pharmacy benefit manager to be able to -- with high fidelity -- discriminate in advance and intervene in an effective manner. This approach addresses adherence problems among those patients who need our help the most. Improved adherence is the hallmark of better quality care, healthier patients, and reduced overall medical costs" (see press release).

Of course, it is nothing new for PBMs to mine their patient data and find those patients who have not filled prescriptions.  According to the Wall Street Journal, "the new efforts are broader, and can focus on apparently healthy people. They use models developed from enormous troves of medical and other data. These are then applied to each patient's own claims information."

PBMs often get paid by pharmaceutical companies to identify patients who are not refilling their prescriptions and to send them prescription refill reminders by mail or call them. Increasing adherence can greatly help pharmaceutical bottom lines and PBMs also benefit when more prescriptions are filled. Conventional wisdom -- and maybe even some research data -- suggests that patients will also benefit.

But computer models often result in unintended consequences as evidenced by automated stock market trades. What could be the equivalent breakdown in the health market? Obviously, there's the privacy issue: "Ethics researchers say such efforts can raise privacy and other concerns if people don't deliberately grant permission for such use of their data, as well as potentially usurp the role of doctors, who know patients best," notes the Wall Street Journal.

"It undermines the trust an individual has in their physician," said Mark A. Rothstein, a bioethics professor at the University of Louisville," who was quoted in the WSJ article.

It could also undermine the trust in the pharmaceutical industry, I suppose.

One further thought. If these computer programs can identify future health problems in otherwise healthy people, will there be pre-emptive pharma-sponsored notices sent out to patients advising them to see their doctors? As consumers provide more and more private information to pharmaceutical companies, I can even imagine pre-emptive direct-to-consumer advertising!