Insurers vs. Drug Dangers: Cheap but Ineffective?

Health insurers are hoping to ride to the government's rescue on drug safety.

With various pharmaceutical scandals still in the headlines (Vioxx, Avandia and Vytorin chief among them), drug safety has been Topic A in certain corners of the healthcare world, especially with the Food and Drug Administration still coming to grips with how it will deploy the new safety-monitoring powers it received last year. Now comes word that health plans themselves will soon lend a hand by data-mining their vast collections of medical-claims and prescription-drug information to catch possible side effects as quickly as possible, the WSJ reports.

Well, not that soon. But by early next year, Wellpoint plans to be scanning through the health records of about half its 35 million members to identify drug problems. (Why half? No idea - maybe they'll be the only ones whose electronic records are sufficiently up-to-date to yield decent data.) UnitedHealth and other insurers are also looking into similar programs.

But would it work?
The WSJ article is pretty bullish on the prospect, noting that Kaiser Permanente's own internal patient reviews led it to begin backing away from use of Vioxx a year before Merck withdrew it because of heart-attack danger. Early passes through its own data convinced Wellpoint that it would have identified problems with both Vioxx and an anti-cholesterol statin from Bayer called Baycol within about four months of launch -- but nothing related to heart risk associated with the diabetes drug Avandia.

Is that good, bad or indifferent? At the moment, it's kind of difficult to say -- and that's one reason this sort of plan makes me just a little nervous. While the WSJ quotes FDA official Janet Woodcock saying that "[w]e're bringing together all these different groups in a network so that we can ask them to look at the same question at the same time," it's far from clear whether FDA and other experts will have access to the raw data or just analytic summaries provided by the insurers. (Wellpoint's own press release has nothing on that point either.)

Not only would the latter make it much harder to draw reasonable conclusions across different health plans, it would also give outsiders no way to double-check the insurers' calculations -- something of a priority when public health is at issue. Oddly, the WSJ is silent as to whether FDA also plans to tap the single, huge Medicare database for this sort of information as well, which would provide at least a publicly available backup to these private databases.

There is a season, churn, churn, churn
Finally, there's another problem with relying on health insurers: Churn. I haven't been able to find any decent numbers suggesting exactly how many customers switch into and out of health plans every year, but I suspect the number is fairly substantial. (If anyone's seen data suggesting how long the average non-Medicare recipient remains with one health plan, I'd greatly appreciate a pointer.) That's an issue here because if drug problems take time to crop up -- say a year or two, or five -- churn among the insurers is going to fragment the data so much that a danger "signal" may never be seen.

So it's important to be clear that FDA has basically decided to safeguard the health of the American people on the cheap, which is fully in keeping with the general ethos of the Bush administration. However, it also suggests that the insurers may well eventually miss a major drug-safety scandal, forcing us to start a new drug-surveillance scheme from scratch -- wishing along the way that FDA had just done it right the first time.

(Photo of a German open-pit mine from Flickr user ReneS via Creative Commons)