Economists Do It With Models

Warning: “graphic” content…

Bookmark and Share
Truth in Advertising, Diarrhea, Er, Alli Edition…

July 31st, 2008 · 8 Comments
Advertising · Behavioral Econ

Now that I have your attention with the title of this post…

As someone who is quite accustomed to the academic lifestyle, I found it quite painful to have to be at work at 8:30 this morning. I was up early because this week is the start of the Midcareer Summer Program here at the Harvard Kennedy School of Government, and the instructors were supposed to show up to one of the orientation meetings to introduce ourselves to students. (If you are curious, there is a 1 year midcareer master’s program at the Kennedy School, and, since these students have been out of school for a while, they arrive a month or so before the fall semester starts and go through boot camps in economics and math.) I listened to a speech given by Frank Hartmann, the faculty chair of the Summer Program, in which he said that it is important for even generalists to know a bit about economics and be familiar with quantitative methods. He argued that, even though these students would not likely be directly using these skills in their future endeavors, it was still important to have them in order to not have to take what various specialists in these disciplines say at face value. In other words, they should be able to question and challenge rather than just smile and nod. I agree…but I digress. I use this anecdote to motivate one of my favorite examples of what organizations can get away with when people aren’t thinking critically.

Consider alli, the weight-loss supplement marketed by GlaxoSmithKline. (See for an overview.) According to the company, alli works in the intestines by attaching to enzymes that break down fat and thus preventing its absorption into the body. GSK claims that alli blocks the absorption of about 25% of fat consumed. However, alli is not a miracle pill: GSK is very quick to point out that there are very unpleasant side effects (too gross to go into here) that occur if an individual doesn’t restrict his diet to 15 grams of fat per meal.

GSK claims that the results of its clinical studies show that people who take alli lose 50 percent more weight than people who don’t. They quote data points in the range of 5-10 pounds, saying for example that people who would lose 10 pounds without alli would lose 15 pounds with it, and that people who would lose 5 pounds without alli would lose 7.5 pounds with it. GSK implies that this difference is due to the functioning of the drug itself. (Note that clinical trials of pharmaceuticals are supposed to get at this causality, since people are randomized into treatment and control groups and all other factors are held constant as much as possible.)

Now, let’s do some math. If a person is restricted to 15 grams of fat per meal, then alli would block at most 3.75 grams of fat per meal. Assuming 3 meals per day, this would imply a maximum blockage of 11.25 grams of fat per day. Fat contains 9 calories per gram, so the maximum calorie blockage is about 100 per day. A decrease of 100 calories a day translates to a loss of about 1 pound per month, which does not sound so impressive. Apparently I’m not the only one that thinks so, since the following is posted in the FAQ on the alli web site:

“The math behind alli does not seem to add up. If the drug in the alli capsule inhibits the absorption of about 25% of the fat that I eat, and I follow the plans, it would take me around 5 weeks to lose 1 pound. How does that equate to 50% more weight loss than dieting alone?
It is important to remember that alli capsules are not intended to do all of the work for you or to replace diet and exercise. alli is indicated to be taken with a reduced-calorie and low-fat diet.

Clinical studies have proven that people who followed a reduced-calorie, low-fat diet while taking alli capsules experienced 50% more weight loss than those who just followed the diet. alli capsules (orlistat 60 mg) prevent absorption of about 25% of the fat you eat from being absorbed and cause a calorie deficit in your diet. The actual amount of fat blocked by alli is dependent on the total amount of fat consumed. As discussed in the educational materials that are provided in the alli starter pack, individuals will have different total caloric daily needs based on their weight and activity levels.

In addition, individuals who used alli reported experiencing greater motivation. They were more likely to follow and maintain a low-fat diet while on alli. In addition, they were more likely to exercise.”

I think that GSK did a nice job of skirting the question. I also think that the advertisements and marketing points for alli are misleading on 2 levels:

1. The “50% more” assertion doesn’t even make sense. This would imply that the people who would have lost more weight in a baseline scenario get more benefit from alli. But aren’t these people the ones who are likely to be restricting their diet more and likely eating less than the 15 grams of fat per meal allowed? Based on the stated functioning of the drug, these people would benefit less from alli, not more. GSK is quoting a percentage based on the result of one, or at most several, studies without actually establishing a functional relationship.

2. The clinical trials, which are the basis for the claims, do not actually give a valid comparison between the treatment and control groups. Why not? Presumably, the placebo did not give a subject unpleasant side effects if he went over his 15 grams of fat. Because the placebo was inert and the treatment and control groups were given the same instructions and the same warnings (since subjects are not supposed to know whether or not they have the placebo), placebo subjects were able to go over the 15 grams allowed and falsely believe that they hadn’t, based on the information that they didn’t experience the side effects. Worse yet, subjects could figure out that they were given the placebo, which would invalidate the double-blind aspect of the study.

In fact, GSK recently updated its web site to say the following:

“The second, and probably most significant, way alli works is that it teaches you to stop, think and make healthier food choices. You may be able to lie to yourself about how much fat you did or did not eat, but you can’t lie with alli. Treatment effects, though usually avoidable, can be an effective deterrent.”

So even the company acknowledges that its product is, for the most part, equivalent to someone followng you around with a baseball bat and smacking you with it every time you eat something fatty. I know people that would likely do that for free! If GSK wanted to be up front with potential customers, they would be promoting this aspect of the product a lot more, perhaps even more than the actual fat blocking function. I am curious as to why the FDA doesn’t require the product to be marketed in this way.

From a behavioral economics perspective, I can explain fairly easily why GSK doesn’t want to market the product in this way: people tend to overestimate their own self-control, and thus wouldn’t see the need for the drug as a commitment device. Then again, people do sign up for Christmas Clubs and the like, so maybe the baseball bat approach wouldn’t be such a bad idea.

(For those of you that don’t know, a Christmas Club is a savings institution that essentially works as a commitment device- you put money into an account throughout the year and the club doesn’t let you take it out until some specificed period before Christmas. Since the accounts typically don’t pay interest, people are essentially paying for the “baseball bat”. Also, here’s an interesting article on mental accounting that I found during my search for Christmas Clubs.)

Tags: Advertising · Behavioral Econ

8 responses so far ↓

Leave a Comment