Leading policy-makers and scholars explain how market forces, deregulation, and consumer choice can work to improve health care for all Americans.


Shotgun Medicine, Flimsy Evidence
Dr. Eric Topol, Baltimore Sun, 4-3-07

Topol argues that most of the medicine we use today is poorly targeted at the people who stand to benefit the most. We can only overcome this problem, he thinks, by improving our understanding of the interaction between drugs and individual biology.

In medicine, there's a concept called "numbers needed to treat," or NNT—an estimate of how many patients use a drug or treatment compared with how many benefit from it. For treatments considered "effective," the NNT is typically between 30 and 80. That is, 30 to 80 people have to use a specific drug or device before one person will benefit. Doctors and researchers simply don't know which individuals will respond to what. So we treat as many people as possible in an attempt to benefit as many people as possible. But that can also backfire.

One of the most "effective" treatments in medicine is the implantable cardioverter defibrillator. This device, smaller than an iPod Nano, can sense a life-threatening abnormality in the heart's electrical rhythm and rapidly deliver a shock to restore normal rhythm. There is only one problem. Out of every 100 patients who get this device permanently implanted, only 12 to 15 ever need it and receive an appropriate electrical discharge. The remaining 85 percent or more have a device that costs more than $50,000 to put in, carries a risk of infection and might "fire" inappropriately. More than 200,000 such devices are implanted in this country every year.

On the pharmaceutical side, statins are among the world's top-selling drugs, with more than $15 billion per year in prescriptions. These medicines lower the blood's "bad" cholesterol level. However, for every 100 patients who take a statin, only eight to 10 derive any real benefit in terms of reducing the risk of heart attack, stroke or death. The remaining 90 percent or so are taking an expensive medicine—at a cost of $3 to $4 per day—with potential side effects. They gain the psychological comfort of a reduced cholesterol level but no real health value.

Likewise, the second most commonly prescribed medicine in our country is Plavix, taken to avoid blood clots. About 30 percent of patients do not respond to the common dosage and are left vulnerable to developing potentially lethal blood clots.

Part of the problem is that we don't test medicines in the same way that we prescribe them. Almost all medicines in common use have never been tested in the elderly population, whose metabolism and kidney function are drastically reduced. And most clinical drug trials are performed in low-risk populations that don't reflect the real world, where patients often have multiple medical problems.

How can we evolve toward a more intelligent delivery of medical care?

The best prospect is the "ome." This simple term collectively stands for the genome, proteome, epigenome, transcriptome and metabolome. Basically, in aggregate, the "omes" define an individual's biology at the gene, protein and metabolic level.

We are not far from having practical, affordable and complete human genome resequencing. In recent months and even weeks, key genetic markers associated with diabetes, blindness, lupus, inflammatory bowel disease and autism have been identified.

Although we all recognize the profound inefficiencies in overall health care delivery, we largely have taken for granted the waste in medicine itself. Now we are at the brink of an exciting discovery phase for biomedical research—one that will radically reboot medicine. In the years ahead, doctors will tailor a specific therapy or prevention to particular biologic vulnerability or need, one human being at a time.

Project FDA.
home   spotlight   commentary   research   events   news   about   contact   links   archives
Copyright Manhattan Institute for Policy Research
52 Vanderbilt Avenue
New York, NY 10017
(212) 599-7000