Randomised controlled trials, such as those
conducted at CTU, provide the best evidence about the effects of
treatments for different disease or conditions. However, if a
number of similar trials exist, then it is better to review and
summarise the evidence from all of these trials, rather rely on
just one.
Traditional ‘narrative’ reviews that simply
describe trials and their results are not the best way of
summarising the results of similar trials. This is because there is
no standard or way of doing these reviews and frequently they are
based on only some of the relevant trials. For example, they may
only include trials that are well known or that have particularly
striking results. Therefore, this sort of review can give an
incomplete and sometimes inaccurate impression of the effects of
treatments.
In contrast, systematic reviews:
- Use clear, standard and reproducible methods and
specify these in advance
- Say which types of trials they will include and
exclude and why
- Use a comprehensive search strategy to identify
all relevant trials, whatever their results
- Assess the quality of the individual trials
- Present the results of the trials in a clear and
structured way.
Therefore, they provide a much more objective
and reliable way of reviewing all the relevant evidence from trials
about the effects of treatments.
Meta-analysis is a means of quantitatively combining the results
of research studies to provide overall summary statistics.
Moderate improvements are often the best that can be hoped of
new medical treatments. These can be both clinically worthwhile and
for common diseases extremely important in terms of public health.
Many trials have recruited too few participants to detect these
differences. Thus, many will have been inconclusive while by chance
alone a few may have demonstrated statistically significant
positive or negative results. Combining the results of suitably
similar trials in a meta-analysis may give sufficient statistical
power to reach a clear and more reliable answer.
A good quality meta-analysis should always be done in the
context of a systematic review. The data used in these analyses can
be extracted from published reports or obtained as summary data or
as the "raw" individual patient data from those responsible for the
trial.
