studies.

Now, these are really a snapshot in time.

I've mentioned one of the examples before and those are surveys.

We've all received a request to fill in a survey in our email inbox, and

the authors of that might of wanted us to be part of a cross-sectional study.

Another example,

a very good example of cross-sectional studies is epidemiological surveys.

We just want to know how many patients at a certain time has a certain disease.

More exciting things about the cross-sectional study though is,

it can be incorporated into other study designs.

And you'll see that quite often.

Now, let's consider an example, Lawrenson and colleague in 2013.

They really sent out a questionnaire to healthcare professionals.

Those who are optometrists and ophthalmologists.

And I want quote from the abstract that says,

conduct a cross-sectional survey of current practices.

So, what did they do?

They just wanted to know what advice patients were given, and

those were patients with age related macular degeneration.

So, that's indeed a snapshot of time.

They send out the survey and ask what advice are you giving right now.

Now, one thing you can do with these results is form many, many subgroups.

Say for instance, not done in this study.

But you could say let's divide the participants or

people who filled in the survey in to the different areas of the town,

that side of the river and that side of the river.

Once we've cut these two groups,

we can now compare what were the answers to the other questions.

And so, for almost all of those questions you can form two little groups

depending on what the answer, and evaluate the other variables.

And so, there's so many answers you can come up with in this sort of study.

I mentioned it's also very common to do

epidemiological studies in the form of a cross-sectional study.

We're really looking at prevalence.

Another example I wanted to tell you about is Sartorius and colleagues

they just looked, right here in South Africa at the determinants of obesity in South Africa.

And they did those from data, from very large surveys.

But again, it's a snapshot in time.

Now, unfortunately, there are problems with this kind of analysis.

First of all, there is a response bias.

Think of sending out a survey.

There's a certain subset of people who will respond to a survey and

those who won't, that might skew the data.

There might be something inherent about the subset who do decide to respond

versus those that don’t.

You've got to be very careful about that.

Also, cause and effect can't really be identified, or

separated from each other.

We might get from the different sides of town, different results.

But was living on different sides of town really the cause for

those differences in answers?

So, why do cross-sectional studies?

Well, number one is very quick.

If you think how quickly you can fill in the survey.

If you're the researcher, you ask some people to fill in the survey for

you, your data's collected right there, it can be very quick.

It also can be very cheap.

Imagine just sending it out via email, there's really very little cost to that.

So, you can get some powerful answers very quickly and

very cheaply from cross-sectional studies.

Now, in the next lecture, I'm going to talk about cohort series.

It'll be the last type of observation studies we're going to talk about.

Really interesting stuff.