Diminishing Returns: Improving Health vs. Selling Drugs

In the previous post I laid out the Inverse Benefit Law that Donald Light and I recently proposed. Another insightful article recently appeared that demonstrates further how the basic mechanisms work that cause a clash between drug marketing and public health:
http://www.jabfm.org/cgi/content/full/23/3/371

Dr. James W. Mold and colleagues from U. Oklahoma Dept. of Family and Preventive Medicine address a phenomenon that is becoming common as new guidelines are promulgated with management strategies for chronic diseases (often, today, written by authors or organizations with significant financial conflicts of interest with Pharma). If docs follow the guidelines rigorously, they will feel duty bound to attempt numerous risk reduction strategies for each patient. The authors construct a hypothetical illustrative case of a 65-year-old African American man diagnosed a year ago with Type 2 diabetes, whose blood pressure, hemoglobin A1C (a measure of diabetes control), and "bad" cholesterol are all higher than normal. According to standard guidelines, this patient needs at least 6 major medical or lifestyle interventions right now to reduce his risks of the major complications of diabetes such as heart attack, stroke, amputations, and kidney failure. (Given the current fetish for preventive medicine, many would be outraged at the idea that we'd skip even one of the 6.) If you read the guidelines, more often than not, these 6 interventions would all be listed as equally high priority.

Mold and colleagues tell us that the real situation is quite different and demands a much more nuanced approach. There are two reasons for this, one fairly obvious, the other less so.

The first reason is that if you look individually at the posible interventions, the 6 are not equal. They have very different effects on the reduction of risk. For instance, in our diabetic patient, taking a lo-dose aspirin daily offers twice the value (in terms of lowering the risk of a heart attack) as does lowering the "bad" cholesterol from 140 to 100.

The second reason is a matter of math. When you do intervention #1, you lower the patient's risk by a certain percentage. Let's for argument's sake pretend that all 6 interventions lower risks by the same percentage, which as we have just seen, they don't. So aspirin, let's say, reduces your risk of dying by 20%. So you add aspirin. Intervention #2 also reduces your risk by 20%, so we add that intervention. But your risk is no longer 100%; it's now down to 80% because of Intervention #1. So to calculate the new risk reduction, you need to take 20% of 80% which is only a 16% additional risk reduction. And so on each time you add a further intervention--hence the "law of diminishing returns" for risk reduction strategies.

An aside here-- what we are basically saying is that risk reduction is not additive. If you had 5 interventions, and each alone reduced your risk of dying by 20%, then if they were additive, and you did all 5, you'd become immortal. But we know the world doesn't work like that.

So back to pharmaceutical marketing. Dr. Mold and company basically are recommending that the smart primary care physician will go for the biggest bang for the buck as it were. For that hypothetical patient, if you focused on the top 3 interventions only, you'd get more than 80% of the real value of doing all 6. And if you focused on 3 instead of 6, it's much more likely that the patient would actually do them; you'd likely save a whole lot of money for both the patient and the payer; and you'd avoid the risk of medication side effects if the intervention involved a medicine. From both the individual patient-doc perspective, and the larger public health perspective, it's a smart move.

The drug industry, on the other hand, is much more likely to want docs to think that if the guidelines list 6 interventions, well doggone it, you had better do all 6. Recall that in our "diminishing turns" list for our patient with diabetes, the most effective intervention was the daily lo-dose aspirin, and the least was cholesterol lowering. To do the daily aspirin involves a cheap generic drug, that generates no profits to speak of for Pfizer or Merck or any of the big boys. (You don't see consumer ads on TV saying, "Ask your doctor if a daily generic aspirin tablet is right for you.") To get the man's cholsterol from 140 to 100 will probably take one if not two expensive statin drugs, many of which are still brand-name. Do you think the statin makers want us to think that cholesterol-lowering might be an expendible risk reduction in a smart patient-oriented strategy? Their marketing message to the doc is almost for sure to be that lowering cholesterol is every bit as important, if not more so, than any other risk reduction move.

A final tangent. Mold et al. give us pause not only about drug marketing; they also call into question the wisdom of most pay-for-performance strategies. In the abstract, the idea for paying the doc more for providing better quality care seems irrefutable. But if by "quality" we mean "meeting all the guideline targets" then we see how P4P can undermine smart practice. As the law of diminishing returns shows, you often do the best for your patient by getting close to, but not quite meeting the target--and adding more and more meds to be sure that you get to the target (whether cholesterol, or blood pressure, or blood sugar) might add very little extra benefit while greatly increasing the chances of adverse reactions. But the doc is heavily incentivized to try to meet the payment goal and not to think about what really most helps the patient.

Tip of the hat to Primary Care Medical Abstracts for including the Mold paper in their reviews.