| Тойм: | <p>Systematic reviews are increasingly important in evidence-based medicine. Their validity
is dependent on the underlying data, thus any decisions about the eligibility of trials based
on where they are published, for example excluding trials from the grey literature, could
_bias their results and conclusions. However, the identification and inclusion of trials
published in the grey literature can be time consuming and difficult. This thesis assesses
the actual and potential impact of including trials reported in the grey literature in
systematic reviews of healthcare interventions. A review of existing knowledge in this
area was first carried out. Four substantial projects were then conducted.</p>
<p>The first of these projects assessed the actual impact of grey literature in meta-analyses
published in the <em>Cochrane Database of Systematic Reviews</em>. This showed that published
trials may produce larger treatment effects than trials identified in the grey literature (grey
trials). There was no difference in methodological quality between published and grey
trials; this was often unclear or not reported in both sources. The second project assessed
the potential impact of grey literature in systematic reviews. Practical issues in locating
and including grey trials were identified, namely the quantity and quality of information
available. The third project built on these findings, assessing the quality of trials reported
in conference proceedings and whether this improved over time. Important limitations
were identified including a lack of numerical data and the preliminary nature of the results
presented; these appear to be worse in recent years. The final project assessed how trial
information reported in a conference abstract differs from its full publication. Major
changes between abstract and full publication were identified, in particular details of trial
results and participants included and analysed in the trial. It reinforces the preliminary
nature of some trials reported in conference proceedings.</p>
<p>This thesis confirms that the exclusion of trials reported only in the grey literature remains
a threat to the validity and reliability of systematic reviews. Systematic reviewers must,
therefore, search for and include trials from the published and grey literature. Failure to
do so may result in the results and conclusions of that review being biased and unreliable.</p>
|
|---|