What standard ways of evaluating existing studies are available?
In my opinion, the guidelines set forth by the Cochrane Collaboration set the gold standard for conducting a systematic review. (See below.)
J.P.T. Higgins and S. Green (eds). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011. Available at http://handbook.cochrane.org/.
The main problem with the Cochrane handbook is section 9.5.4 as that method (random effects model) probably should be abandoned. See www.epigear.com for details under "MetaXL"
Your question and inclusion criteria would decide what to include and exclude.
However, if you are talking about evaluating the quality of included studies, there are scoring systems available such as Jadad scale, Newcastle-Ottawa scale.
Dear Colleague
Quality appraisal checklists like STROBE, STARD, PRISMA, CONSORT and so on could be used in case of your research questions with some modifications.
Thank you for the very helpful answers. Amongst the tools listed (Jadad, Newcastle-Ottawa, STROBE, STARD, PRISMA, and CONSORT) are they interchangeable? Is one better than the others? Does it depend on what type of meta-analysis is being done? What determines when each should be used?
Hello Everyone
Guidelines such as STROBE, PRISMA, CONSORT, QUORUM, MOOSE and others provide standards for the REPORTING of different types of studies- to borrow a phrase from Dr. Brand (and numerous other journal editors), "In essence, these guidelines tell investigators and authors what information is required to ensure readers (and reviewers) can properly evaluate the study." (Clin Orthop Relat Res. 2009 Jun; 467(6): 1393–1394.). The Jadad scale, Cochrane Risk of Bias instrument and Newcastle-Ottawa scale provide an indication of the methodological quality of a study.
Hi
In depend on your study type that you want to review, the tools maybe different. for example for RCTs and CCTs Cochran ( REVMAN5.3) very useful but maybe for cross sectional ones not. Many journals use to follow CONSORT and PRISMA guidelines. Really as a researcher you have to justify between this methods for choosing the best one for your own study and just remember your choice may be the best for your research and not for the other ones.
Best
The decision for including or excluding publications, respectively studies in a systematic review (SR) or metaanalysis (MA) depends primarily on your definition of your research question. Depending on your question you have to set a list of inclusion and exclusion criteria which publications - of those you will receive by your systematic literature search - you will add to your SR / MA.
As Gorden L Warren mentioned guidelines from the Cochrane Collaboration set the gold standard for conducting a systematic review. The other standards mentioned (like STROBE, STARD, PRISMA, CONSORT) will not help you in anyway. They set standards for publishing research. - You can use them to assess the quality of the publications you will include or exlude (according to these standards). But they give no criteria whether you could include papers in your MA; this depends on your research question.
It will also depend on your research question and the studies published in that area. While RCT's are preferred, you may only have observational studies published in your area of interest. Once you have the sense of the literature, you will then choose the quality assessment tool to match with the type of studies you are summarizing in your review.
It's important to define the inclusion criteria, however while reading a meta analysis I would focus on tbe exclusion criteria more than on inclusion criteria. Make sure your exclusion criteria is rationale-based and thus it may help you avoid bias.
Hello everyone. What amaizing forum!! I think that inclusion criteria might be defined according to the PICO acronym. For quality assessment or risk of bias, I can suggest the PEDro scale for randomized controlled trials. It is also important to distinguish that methodological tools such as STROBE, PRISMA, CONSORT are not useful neither for quality assessment nor risk of bias, they only allow us to verify the quality of reporting.
Sincerely, Jose.
The criteria are defined by the authors. If you'd like to do a meta-analysis of observational studies MINORS is also a good choice, I think it is more comprehensive than Newcastle-Ottawa scale.
http://www.ncbi.nlm.nih.gov/pubmed/12956787
Criteria to include studies I'm a systematic review are specified a priori by the authors of the metanalysis. They should include the PICOS (for reviews of interventions): Patients/participants; interventions; comparisons; outcomes; and study design, for instance randomized controlled trials and cluster trials.
Quality of evidence may be assessed by GRADE.
Best regards, Matteo
I think by following PICO, you can construct your title first. Then after searching literature systematically, try to find out whether the articles of interest answer your PICO. If yes, can include for the time being, however, need to assess critically later whether these articles are truly eligible to be included or not (whether the data representation is complete and clear).
Good Luck!
Regards,
Md. Asiful
I agree with you all. Do not forget that the type of article or design is a Systematic Review. Meta analysis is the quantitative analysis of Systematic Reviews and based on the heterogeneity of the included studies one can simply decide if that sys rev. needs a meta analysis or not. You must always try to include high quality studies to reduce metabias. You can for example use JADAD score to evaluate the bias of the studies you want to include so you can be assure the quality of your Systematic Review. Sometimes the evidence is so vague that a meta analysis is not the best option to adress the investigation question in a numerical way. So I recommend that first, analyze the heterogeneity of the studies so you can be sure that a quantitative analysis of the data is needed.
Can we use PRISMA and some sections of COCHRANE together for an intervention review paper?
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis