There are number of study designs that are used to evaluate diagnostic tests and decision rules. The list below shows you some of the designs you may come across.
Cohort studies (where you take a group of subjects, record their exposures and follow them up to see who develops the outcome) are widely used to evaluate diagnostic tests.
These studies can be biased if subjects are lost to follow-up, particularly if the reason they were lost was that they developed the outcome (for example, if they died of a pulmonary embolism!).
Cross-sectional studies are sometimes used in evaluation. These studies don’t follow patients up, but rather collect data about exposure and outcome at the same time.
The evidence they provide is not as strong as other studies, but they are cheaper and easier to carry out.
They can also be more biased, as all the data is usually collected by the same person.
Randomised controlled trials are sometimes used to compare two diagnostic tests against a gold standard.
This is where you take a group of patients and randomly allocate them to a diagnostic strategy. Both groups are followed up to see who develops the outcome.
These studies are however expensive and difficult to do.
Case-control studies are rarely used for diagnostic tests.
This where you take a group of people (cases) who have already developed the outcome (for example, deep venous thrombosis) and a group who have not developed the outcome, and compare their exposures.