8. 8 Other types of studies
Although RCTs produce the highest quality of evidence other types of studies can sometimes be more appropriate.
- Observational cohort studies.
- Case-control studies.
- Cross-sectional studies.
- Case studies and case note reviews.
- Literature review and systematic literature review.
- Meta analyses.
Randomised controlled trial (RCT)
RCTs are usually experimental and prospective, and compare two or more groups.
- Randomisation makes sure each group will be similar at the start.
- The control group helps confirm whether any effect is real, rather than just being by chance, or from other external factors.
All new drugs are studied in RCTs before they can be approved.
Observational cohort studies
Cohort studies are usually observational and longitudinal.
They can be either prospective – to see what happens going forward over time – or retrospective, looking at medical records or databases.
Cohort studies might include all patients at a hospital or in a region or country. If a study needs to be very large, international cohorts can combine different national cohorts.
The difference to RCTs is that the whole group will be studied together and there is no active intervention. This means cohort studies produce different types of results to an RCT.
- They include a wider range of participants including people who would not normally enrol in an RCT.
- They can often run for much longer, sometimes for decades.
- They can collect data on a wide range of treatments and events, not just the limited intervention in an RCT.
- However, they only report what is observed. They can report associations between different factors but unlike an RCT they can’t prove that one thing leads to another.
- Results have to be interpreted carefully to try to alow for other things that might explain the results (called ‘confounding factors’).
These studies are usually observational and retrospective.
A group of people with a symptom or illness (cases) is compared to a similar group without the symptom or infection (controls). This is to find out what factors either caused the symptom or prevented it.
For example, a case control study could look at a group of gay men aged 18–30 who tested HIV positive and compare them to a similar group who stayed HIV negative.
The comparison might find social factors were strongly linked with becoming positive, for example, alcohol use or income or education etc.
These studies look for what happens at a simple time point.
For example looking at hospital records to see how many people smoke. Or what proportion of people living with HIV smoke.
These are usually quick studies to identify the scale of a problem.
Cross-sectional studies can identify prevalence of an illness (how many people have this at one time) but not the incidence of an illness (how many people will develop an illness over time).
The results are limited by not being followed in time. They don’t provide any information about what happens next – whether some people might get better and other get worse.
They can report associations – for example that smokers might have higher blood pressure – but they cannot prove that one thing causes another or whether an intervention improves health.
Case studies and case-note review
These examples are not a strong type of evidence. They are very early observations or research and collect data that might lead to larger more detailed studies.
A case study is where an individual patient report is included as evidence.
Even though all sorts of other factors could have caused what was seen, case studies can alert researchers, doctors and patients to something new.
A case-note review is where patient notes are reviewed retrospectively. The quality of the results are dependent on the quality of the data that was recorded.
Literature review and systematic literature review
A literature review can report collected results of selected studies. The process of selecting studies can make the results biased.
A systematic literature review has to include all relevant studies in the area being looked at. Even though these results should represent the whole range of outcomes it can still be biased by the type of studies that are published.
Publication bias includes, for example, that studies with positive results are more likely to be published that studies were nothing new is found. Positive studies are also likely to be published quicker. Studies that are not published can’t be included in literature reviews.
A meta analysis combines results from many different studies.
Although small studies might not find a result that is statistically significant. If ten similar studies are combined, the larger number of participants might be enough to show a real effect.
But combining different studies cannot just look at the results, it needs to adjust for the study differences, including if participants or the intervention was very different.
As with literature reviews, meta analyses are reliant on studies that have been published. This means that publication bias affects these types of studies too.
Last updated: 1 January 2023.