Now, what is a cohort?

A cohort is a group of patients that share a common trait.

And begin to identify them and follow them up over time.

And these are difference between a cohort study, a cohort series and

a case-control series.

In a case-control series, remember, have a snapshot in common, make a diagnosis, and

we look at what happened up until that point.

Now with a cohort series, we are going to identify a trait, a group of people and

we're going to follow them forward in time.

And let's look at an example.

le Roux and colleagues, they included mother child pairs born

in a urban area of the Western Cape in South Africa.

And they followed these pairs up for a year and

they looked at the development of pneumonia.

So, the start of the study was a point in time and that was at birth.

They then follow these children and mothers up for a year, and they look for

the development of pneumonia.

If we were to reconstruct that as a case-control series,

we would go to the clinic and identify children with pneumonia.

Make a control group of children without pneumonia and

then look at data points up till the point.

So, you can clearly see the difference between a cohort series,

looking forward in time, and a case control series looking rearward in time.

Now, I want to bring up two very important definitions.

Retrospective and Prospective data analysis.

People use these terms differently.

Let's come up with one way to properly use them.

Now, retrospective data analysis is usually used in terms of case-control

series.

We identify, for instance, patients with a complication after surgery and

those without.

We go to the patients' files and we gather data in those files,

from those files up until the point at which those complications occurred.

Now, these usually come from freehand notes made by clinicians, so

those notes were never designed to contain specific data points, specific variables.

We've got to extract it from that, and

that is when you called retrospective data analysis.

We retrospectively get some data out of a patient's file.

Now, prospective is used much more commonly in terms of cohort studies.

We identify the patients for our trial, for our study I should say,

at the beginning of the study, so at birth.

We now have a form.

We've designed a format because those are the data points we want to put them on

the form, that is what we want.

So, the data that we gather every time we phone those patients,

every time we see them in the clinic, that's prospective data collection.

Now, it's not so clear cut, because in this day and age with computers,

we can have databases.

So, I can have a database in my ward,

whereby I collect data on certain operations prospectively.

Whether they're part of a trial or not, whether they're part of a study or not,

that is the data I collect.

That's clean data because there's little boxes that I want

the clinicians to fill in.

I can now do a case-control series with complications without looking at

that data.

That data was really collected prospectively.

I could also decide I want to look at children with pneumonia and

without pneumonia, and I have not included them at the start of the trial.

But at the end, I'm just going to grab the files from the clinic, and

I'm going to look at them again though from birth, and get that data from a file

that had that data entered without being part of a trial.

So, that would really be retrospective data collection of

a forward looking cohort trial.

So, you can see those things in both directions.

So, in short, those would be cohort trials prospective and

retrospective data analysis.

And you might imagine that these children, in our example,

are followed up for a year.

Just the last word, it needn't be that long a follow up to make a cohort trial.

I might just want to follow up patients for the first five days after surgery.

And see what happens to them, do complications occur.

Now, next up we're going look at the exciting world of experimental studies.