Now we're going to move things up a gear. We're going to discuss Case- control studies now. In the previous lecture, I mentioned that there's actually more than one group. There's going to be a case Group and a control group and they�ve got to differ in some fundamental way. There�s got to be some variable in which they differ. And now begin to identify other sets of variables and we're going see, is there a difference in those variables between those two groups. So what we really do is we're choosing a spot in time and then we look rearwards. And what do we want to do? We want to identify differences in those variables and perhaps risk factors. Now, I mentioned I did the same operation on a set of patients over and over again and some developed a complication, some others did not. When they developed those complications that is our spot in time. We now look at data before those complications occur, so that is a rearward look. Now, let's consider an example. Yung and colleagues, they looked at factors associated with atopic disease in toddlers. Now, that's allergic disease. Now, the first sentence in the abstract says in this case-control study, so it's a beautiful example of a case-control study. So, they selected 206 toddlers. Toddlers visiting a clinic and they collected all sorts of data on a variety of variables but they had two main groups. The case group was toddlers with atopic disease and the control group without atopic disease. So that diagnosis was made. It was a point in time, and then now they looked at data before that diagnosis was made. So they looked at age, and gender, they looked at delivery variables, breastfeeding patterns, family history, etc. They took one at a time, one of these variables, and looked at differences between those two groups, the group with atopic disease and the group without. Wanted to see was there differences. So notice very carefully, it is a rearward look, and that is what makes a case-control study or series. Now there are problems, they don't stand out from cases perhaps, but there are problems with case-control series. Number one is bias. Now if you read that article, you'll see the authors say that this clinic was four people from a higher socioeconomic group, we got to ask the question, is the answers they came up with after the analysis, is that generalizable to a more general population. And inherent to these types of studies really is confounding. Now what is confounding? Confounding really is hidden causes. Say for instance, for argument's sake, they found a big difference in the breast feeding patterns between the two groups. We could now say there is a difference between the two groups in the breast feeding patterns, but that does not prove that breast feeding was the cause for the difference between children with and without atopic disease. So simply very important we can't exclude, most of the time, confounding factors in case-control series. Next up, we're going have a look at cross-sectional series.