Let’s begin with an obvious comedy. Way back in the day when Saturday Night Live was new and fresh, they did a skit that portrayed a General briefing a Soldier on a dangerous mission that required lots of instruction and equipment.
General: This is a laser-guided electronic gaphook. This will help you get over the fuel dump selector fence perimeter. It’s a beautiful software item.
Soldier takes the gaphook.
General: This is the digispan target fighter. Team Lab One’s newest piece of personnel elimination equipment. You’ll love it.
Soldier takes the digispan target fighter.
General: And, uh, take that, and take that pill.
Soldier takes the pill!
General: Okay, now . . . the chopper will drop you here, in Quadrant B, and, uh . . . That pill, by the way, is an L-Pill. Kills within 20 seconds of time of consumption. Poison lab came up with it. Take it if they try to capture you.
Rim shot! Be here all week!
Now. To the science of Take That Pill which is just as funny without being comedy!
Unless you are unfortunate enough to have a rare and properly diagnosed disorder, vitamin supplements are only the means to produce more expensive urine. In just the latest meta analysis, we know that vitamins (and anti-oxidants) produce no mortality advantage, not even with silly statistical significance and this from a sample in excess of 290,000 cases! Myuong et al. scoured the research lit and located 50 randomized controlled trials, then meta analyzed the individual outcomes to determine an RR of 1.00 with a 95% confidence interval of 0.98 to 1.02. The big news with this meta is the old news that at least since 2003, meta analysis of vitamin experiments has shown no effect, yet researchers still keep doing vitamin trials.
Consider this fun two experiment package. Other Guys were recruited to pilot test a future vitamin study by taking either the actual vitamin or a placebo. They were randomly assigned to either condition and told which pill They were taking. Of course, all pills were placebo, so the active persuasion variable is whether the doctor told the Other Guy it was a vitamin or a placebo rather than the actual pill contents. The cover story here involved that pilot study to gather information about how the pill tasted and was tolerated before the main study. Thus, all the Other Guy are influenced only through words. The two experiments varied on health behaviors the Other Guys performed during the experiment.
We examined the effect of dietary supplementation on the relative desirability of exercise and hedonic activities (Experiment 1), the choice between a buffet and a healthful, organic meal (Experiment 1), and walking for health (Experiment 2). Table 1 shows the means for the behaviors by the pill conditions.
All of the differences between the vitamin and placebo groups are Medium Windowpanes, about 35/65 effects, so they would be obvious to a careful observer and you wouldn’t need to count on your fingers (or toes!) to determine the difference. The funny thing to note about these differences, however, is the direction of the effect. People who thought they were taking the placebo showed the health behavior! Not the vitamin group!
The researchers had anticipated this outcome and thought it would arise because Other Guys getting the (fake) vitamin would feel more invulnerable. To that end, the researchers provided a survey to measure just that perception. Sure enough, they found that people taking the faux vitamin rated themselves as less vulnerable to health problems. The researchers then ran a standard path model test on the three variables of pill type, vulnerability, and health behavior. While the pill manipulation alone had an impact on the health behaviors, all of the action came from that perceived invulnerability.
So. We know from repeated meta analytic summaries of RCTs that vitamins have no effect on mortality, but just the belief that you are taking one is enough to make you feel invulnerable and therefore show fewer healthy behaviors related to diet and exercise! Is this a great persuasion pill or what?
Physicians can stupidly recommend it and make money on the recommendation and then later on make more money on the patient who is no healthier for spending the money and taking the pill because the pill has no effect except to motivate a false sense of invulnerability which in turn motivates more couch sitting and large serving soda drinking!
All Bad Science Is Persuasive!
Myung S, Ju W, Cho B, Oh S, Park SM, Koo B, et al. Efficacy of vitamin and antioxidant supplements in prevention of cardiovascular disease: systematic review and meta-analysis of randomised controlled trials. BMJ 2013;346:f10
Wen-Bin Chiou, Chao-Chin Yang, and Chin-Sheng Wan. (2011). Ironic Effects of Dietary Supplementation: Illusory Invulnerability Created by Taking Dietary Supplements Licenses Health-Risk Behaviors. Psychological Science published online 15 July 2011.
P.S. Think about that meta outcome of 1.00 with a CI of 0.98 to 1.02. That’s 50 different RCTs with nearly 300,000 people. Think of all that effort and cost to run 50 randomized and controlled trials on that many people. And combined, the results are zero. Zero, with a tight little confidence interval because it is nearly 300,000 cases. If you were teaching stats and wanted a real life example of a random distribution, you could use these data. Or else a random number generator.