What are special populations? A look at women and minorities
Now, what’s the definition of special populations? I think this term refers to a population that has distinct features. So we can describe any population as a special population as long it’s a subset different from the general population. One of the common classifications is based on ethnicity. This will be a major part of my talk because it’s important. There’s also gender. And I’ll show you examples that suggest that sometimes geographic location is important. And then there’s age, which is also very important.
So, now the question is, why study special populations? Let’s look at two examples.
I said women might be seen as belonging to a special population. So let’s consider a study from Australia that evaluated the outcome of early-stage colon cancer with chemotherapy treatment: There was not much difference in the length of time male patients on chemotherapy and those on no chemotherapy treatment remained alive, but in female patients, there was a difference. If the researchers had looked only at male participants, the results of the trial would be misleading.
For the second example, I’ll talk about ethnicity. In one study, children with acute lymphoblastic leukemia were all treated in the same manner in a clinical trial. So any differences in outcome were not due to differences in therapy. The Caucasian children did better than the others—the Hispanic and African-American children. It makes you wonder whether there was something else in the different ethnic groups that contributed to this difference in results. Again, if the researchers had studied only Caucasian children in the study, the results would be, yeah, we have a pretty good treatment here. And if they had studied only African-American children, they’d say, hmm . . . this is not so good. So, again, if researchers use a homogeneous patient population, they can come up with results that are skewed because our real population is heterogeneous. We have to make sure that there is inclusion of different populations in these studies.
So when we talk about special populations, this is not politics; this is scientific. There is good, sound common sense telling us that if we are going to get one treatment for everybody in a diverse population, then we actually have to make sure that, in the study, we treat the people we are going to give that drug to later. These are not racial quotas, this is not affirmative action–it is good science.
So that is the background. Let’s look at some more examples. Two professors, one at the University of Arizona in Tucson, and the other at the University of Chicago did a study. I think they looked at 11 journals—about 261 clinical trials between 1990 and 2000—and they asked several questions. One was, “How many of the studies will let you know that the researchers treated patients of different ages and genders and ethnicity?” What they found was that about 70 percent of the studies mentioned age, so you could tell whether the participants were younger or older. And some studies were breast cancer studies, so they mentioned menopausal status. Education and gender were also mentioned. But race and ethnicity was only reported in nine studies.
These professors found that 90 percent of the clinical trials failed to describe ethnicity. If you look at these studies, you don’t know whether there were different ethnic groups in them or not. You don’t know whether it was all Caucasians, whether it was all Asians, African-Americans—you can’t tell in 90 percent of the studies. The professors also found that when women and minorities were included in these studies, it was just by chance. The clinical trial designers didn’t set out to say, “This study is going to come up with results that affect the whole population. Let’s actually try to get a representative sample of the population.” No, it just happened that way.
In the few studies that reported specific analysis in minorities and women and so on, a number of times, these studies found different outcomes for different populations—people did better or worse depending on whether they were men, women, Africans, Europeans, Asians and so on. But even in the few studies that had these different outcomes, the trials' conclusions really didn’t address ethnicity or gender. From what we’ve discussed so far, it’s pretty obvious that if you don’t include participants from diverse enough populations, you are going to come to conclusions that may not apply to everybody. So this is an important issue.