Hello,
I am undertaking a systematic review of prevalence rates of individuals with ASD and doing a narrative synthesis of results as I anticipate there will be heterogeneity in the data. Because I am choosing this method of data synthesis, does this mean I won't have to do any specific calculations comparing mean prevalence rates across studies?
Thank you
By definition, narrative means telling a story, therefore no, no calculations would be required. However, if you are simply obtaining prevalence estimates, I would be surprised if at least some of them could not be combined in a meta-analysis. Software is available to do this (eg Stata's metaprop procedure).
Right, no calculations are required. However, I would also like to suggest you to conduct a meta-analysis using collected data. A meta-analysis paper is higher regarded than systematic reviews and by extension, you can do the analysis using your data.
At the point of designing the protocol for your systematic review, it would probably be worthwhile to plan for a meta-analysis ( as Adrian Esterman and Claire Su-Yeon Park pointed out, some computations may be possible. This would depend on a number of things: e.g. the quality of the studies, type and extent of heterogeniety, etc.).
If in the course of conducting the systematic review, you do find that there are indeed valid reasons NOT to combine the data (as you had anticipated), then it would be appropriate to mention these reasons in your write-up and proceed with a narrative review without any computations.
Depends. If everyone agrees then the review becomes a summary of the evidence and an analysis may be unnecessary. If the studies disagree then a litany of "this author said this and that author said that" can become tedious unless there is some objective way to form a conclusion. A meta-analysis may prove useful, though that depends on how many papers you are working with and the quality of the research.
- Badges
- Science topic