00:01
Hello and welcome back to epidemiology. In
today's lecture we're going to get into the
meat and potatoes of population epidemiology,
that is the design of observational studies,
so much of my personal consulting work involves
advising people on how to design studies and
how to critique their existing designs, so
I think you'll find this quite interesting.
00:22
After today's lecture, you'll be familiar
with the main observational study designs,
it's not an exhaustive list, but we get to
the main ones. You will also know the difference
between prospective and retrospective designs
and you’ll know why sometimes, we like
to match subjects across study designs and
study groups. You'll also know why some designs
are more appropriate than others and that
I want for you to be able to do is to be able
to tell when to use certain designs in what
contexts.
00:51
So first let's go over how our various designs
splay across the research universe.
00:57
You'll recall that we tend to have either qualitative
or quantitative research. Qualitative is what
the social scientists tend to do and quantitative
involves numbers and statistics. Amongst quantitative
designs we have descriptive or analytical
approaches. Under descriptive we are describing
a scenario or a sample, who gets the disease,
when do they get it, where do they get it,
who, what, where, when essentially. But the
analytical approach involves an association
between two or more variables and we either
have an observational scenario or an experimental
scenario. The difference is that in experiments
you the investigator are manipulating a variable,
are causing a group of subjects to have an
experience that they otherwise would not have
had, whereas the observational scenario you're
letting the universe unfold as it will and
you're just observing. Amongst the observational
designs, we have the case control, the cohort
and cross-sectional and that's what we're
going to talk about in this lecture. The experimental
approaches tend to be intervention, clinical
trials, RCTs and so forth. So again today,
we are focusing on the observational designs,
the case control, the cohort and the cross-sectional.
02:10
Now how do we rank these various designs,
in the sense of most difficult to least difficult?
Well the least difficult is probably the cross-sectional
design, a survey is cross-sectional, that's
when I want to know what's happening right
now. Next up is the cohort study, it's expensive
and difficult, but it's intuitive and easily
understood. The case control is next, it's
not as intuitive and a little less expensive,
but is not easy to do. Lastly the experiment
is always the most difficult, the most expensive
and we tend to reserve that for things we
really care about.
02:46
So we're going to work through a particular
scenario and that scenario is the question,
is there an association between smoking and
lung cancer. Smoking is our risk factor, our
exposure, lung cancer is the outcome that
we care about. Now let's talk about how we