Another thing that epidemiological studies have convincingly established is that wealth is associated with less heart disease and better health, at least in developed countries. Studies have not been able to establish why this is so, but this too is part of the problem of the healthy user and a possible confounding factor in the history of hormone therapy and many other associations. these epidemiologists are trying to study. George Davey Smith, who began his career studying how socioeconomic status is associated with health, says this research teaches one thing is that woes “come together.” Poverty is a disaster, and the poor are less educated than the rich; they smoke more and weigh more; they are more likely to have high blood pressure and other risk factors for heart disease, to eat what is affordable rather than what experts tell them healthy, to have poor medical care, and to live in environments with more pollutants, noise and stress. Ideally, epidemiologists will carefully measure the wealth and education of their subjects, and then use statistical methods to adjust for the effect of these influences – multiple regression analysis, for example, as such a method is called – but, as Avorn says, she “doesn’t always work as well as we would like.

Nursing investigators argued that differences in socioeconomic status cannot explain the associations they observe with HRT because all of their subjects are registered nurses and therefore this “controls” for variations in wealth and education. . Skeptics respond that even if all RNs had the same education and income, which is not necessarily the case, their socioeconomic status will be determined by whether they are married, how many children they have. have and their husband’s income. “All you have to do is watch the nurses,” says Petitti. “Some are married to CEOs of companies and others are unmarried and still live with their parents. It cannot be true that there is no socio-economic distribution among nurses. Stampfer says that since the publication of the results of the Women’s Health Initiative in 2002, investigators of the Nursing Health Study have gone back to their data to examine socioeconomic status “to the extent that we could.” – examining measures that could indirectly reflect wealth and social class. “It doesn’t seem plausible” that socioeconomic status could explain the association they observed, he said. But nurse investigators never released this analysis, so skeptics remained skeptical.

Compliance bias

An even more subtle component of healthy user bias needs to be addressed. This is the effect of conformity or adhesion. Quite simply, people who comply with their doctor’s orders when they receive a prescription are different and healthier than those who do not. This difference may ultimately be unquantifiable. The compliance effect is another plausible explanation for many of the beneficial associations that epidemiologists commonly report, meaning that this alone is reason to wonder if much of what we hear about what constitutes a diet and a healthy lifestyle is poorly designed.

The lesson comes from an ambitious clinical trial called the Coronary Drug Project, which was started in the 1970s to test whether any of five different drugs could prevent heart attacks. The subjects were some 8,500 middle-aged men with known heart problems. Two-thirds of them were randomized to take one of the five drugs and the other third a placebo. Because one of the drugs, clofibrate, lowered cholesterol levels, researchers had high hopes that it would ward off heart disease. But when the results were compiled after five years, the clofibrate showed no beneficial effects. The researchers then considered the possibility that clofibrate appeared to fail only because the subjects did not faithfully take their prescriptions.

It turned out that the men who reported taking more than 80 percent of the prescribed pills fared significantly better than those who did not. Only 15% of these loyal “adherents” died, compared with nearly 25% of what project researchers called “poor adherents”. This could have been taken as a reason to believe that clofibrate actually cut heart disease deaths nearly in half, but the researchers then looked at the men who faithfully took their placebos. And these men, too, seemed to gain by scrupulously respecting their prescription: only 15% of them died against 28% who were less conscientious. “So taking the placebo faithfully reduces the death rate by a factor of two,” said David Freedman, professor of statistics at the University of California, Berkeley. “How can that be? Well, people who take their placebo regularly are just different from the rest. The rest is a bit speculative. Maybe they take better care of themselves in general. But that conformity effect is a pretty big effect.

The moral of the story, says Freedman, is that any time epidemiologists compare people who faithfully engage in an activity with those who don’t – whether it’s taking prescription drugs or taking vitamins. , exercise regularly, or eat what they consider to be a healthy diet – researchers must account for this compliance effect or they will most likely infer the wrong answer. They will conclude that this behavior, whatever it is, prevents disease and saves lives, when in reality all they are doing is comparing two different types of people who are, in fact, incomparable.

This phenomenon is a particularly compelling explanation for why the Nurses’ Health Study and other cohort studies have found a benefit of HRT among current users of the drugs, but not necessarily among former users. By distinguishing between women who have never used HRT, those who have used it but then quit, and current users (who were the only ones for whom consistent benefit appeared), these observational studies may have inadvertently focused their focus specifically on, as Jerry Avorn puts it, the “Girl Scouts of the group, the compliant continuing users, who probably do a lot of other preventative things as well. ”