This week Jane Galt and I have had a stimulating discussion - Discourse with Jane Galt. Many interesting issues have arisen, and today I would like to focus on knowledge.

The chief of Cardiology discussed Congestive Heart Failure (CHF) at this week's Medical Grand Rounds. During his presentation, he discussed a variety of medical devices available for managing severe CHF. He made an important observation when he pointed out the the device manufacturers had no interest in funding studies which carefully delineate which patients should benefit from a particular expensive device. They would rather show efficacy, and do not apparently mind if physicians implant excess devices (two examples include automatic implantable cardioverter defibrillators (AICD) and atrial synchronous biventricular pacing (ASBP)). Each of these treatments cost approximately $30,000 per patient. We know that in carefully designed studies and carefully selected patients these devices work, improve quantity and often quality of life. We do not know the proper indications for the devices. For ASBP in particular, data suggest that not all patients benefit. Cardiologist would benefit from studies which examine predictors of efficacy. The device manufacturer will not fund these studies, and given our current regulations they have no such obligation. Economic advisors would tell them not to limit the potential market, and the right efficacy study would limit their market. Thus, the economic incentives for society (use these expensive devices only in those patients likely to benefit) clash with the economic incentives for the device manufacturer (sell as many devices as feasible).

A naive response comes to mind. Let the NIH fund the study. But the NIH (actually in this case the NHLBI) will probably not fund that study, stating that they have higher priorities for their research dollars. Medicare rarely funds such studies - their bureaucracy does not seem to understand the importance of efficacy studies.

Let me switch to a pharmaceutical example. Adult onset diabetes mellitus causes more kidney failure than any other disease in the United States (and probably the world). We have learned much about the onset and progression of kidney disease in diabetic patients. We know that very small amounts of protein in the urine predict eventual kidney failure. We have learned that we can both decrease the amount of protein in the urine (without treatment these small amounts become grams of protein) and delay or even prevent the onset of kidney failure.

Recently published studies (for those interested, I have a slide series available from a talk I gave on this subject last year - Update in Nephrology) have documented both a decrease in urine protein and delayed progression of kidney disease. The studies that I cite in that talk all used a class of antihypertensives called angiotensin receptor blockers (ARBs). Of interest, earlier research in patients with childhood type diabetes used angiotensin converting enzyme inhibitors (ACE-Is). The firms that produce ARBs funded the recent studies. They have not, and likely will not fund studies to compare ARBs and ACE-Is. The ACE-I manufacturers will not fund any studies, because those drugs are nearing their patent expiration (at least 2 of that class have available generics, and that number will increase soon). One would expect that the ACE-Is should work as well as the ARBs, but how can we find out? One could easily design that study, but such studies are very expensive. No manufacturer has a financial incentive to fund the desired study, and the NIH apparently will not fund such a study.

We need a new mechanism to insure that we fund important clinical studies. The current system works only when it benefits the manufacture or the issue is so large that the NIH funds the work.

Therefore, I make this modest proposal. We should charge a research fee to device manufacturers and pharmaceutical manufacturers. I have not worked out whether a fee or a research tax makes more sense. We would then have moneys to fund efficacy studies. An expert clinical panel would prioritize proposed studies, and fund them in order until that year's moneys expire. This would allow us to do the right studies.

I suppose that this idea has many flaws. It seems too simple to work. What do you think? How important are efficacy studies? Can we fund the right ones?