Analysis Overview
We’re working on this section. For now, have a look at:
- Free sample size calculator for cluster trials, intra-cluster correlation coefficients and a flowchart to help with decisions about cluster trials are available from Aberdeen’s Health Services Research Unit.
Top 5 Analysis Tips
01.
Make sure that the data collected in the trial are indeed analysed and published. There’s plenty of evidence that a lot of trial data languishes in drawers and hard drives.
02.
Think about how to handle any missing data (or, better still, how to avoid it in the first place). A 2014 review in Trials highlights problems with how handling of missing data was reported.
03.
Thinking about a cluster trial? Have a look at the resources available from Aberdeen’s Health Services Research Unit
04.
More to come: keep checking the site…
05.
More to come: keep checking the site…
More About Analysis
Before the results of a clinical trial are reported, statisticians need to analyse the data to allow conclusions to be drawn on the effect of an intervention. At the core of trial analysis is the statistical analysis plan; this plan ensures that the analyses to evaluate all planned study hypotheses are conducted in a scientifically valid manner, and that all decisions are documented. The analysis of clinical trial data encompasses a large number of related topics including; the choice and definition of analysis sets, the choice of an appropriate statistical model for the type of data being studied, handling of multiple comparisons or endpoints, and appropriate data presentation. Complications such as missing data, participant noncompliance and deviations from protocol must also be taken into account. Trial data are not routinely re-analysed (though this may be changing) so it is important that analyses are correct before results of the study are reported.
GET IN TOUCH / ASK A QUESTION
Looking to get in touch with us directly ? Just fill out the short form below and we’ll get back to you.
