Q: How does one objectively assess 'study quality' for a systematic review?

Detailed Question -

How does one objectively assess 'study quality' for a systematic review? And what data should we extract from the different selected studies for a meta-analysis? My understanding of a systematic review is that data is collected from different studies, statistics performed on these studies, and results interpreted. But even similar studies would have variations in study design, mild variations in the population studied, different data collected, different statitical tools applied? What data should we collect for our statistician?

1 Answer to this question

Systematic reviews and meta-analyses use PRISMA guidelines (http://prisma-statement.org/). Generally, the following steps are involved:

1. A research question is formulated

2. Comprehensive data is collected from various databases using PICOS (participants, interventions, comparisons, outcomes, and study design). The type of data collected depends on what type of studies the researchers are looking at  (randomized controlled studies, observational studies etc.) and MeSH keywords are used for searching databases. A librarian can help in narrowing down the search.

3. The collected data are critically evaluated using the inclusion and exclusion criteria. This needs to be defined by the researchers and depends on the research question. This is an important part of the process and is generally done by two people to exclude bias

4. The data from the selected studies is  extracted and tabulated – information such as study design, authors, publication date, intervention & outcomes of the study etc.

5. Finally, using software such as the following, statistical evaluations can be done and conclusions are drawn.

For a comprehensive tutorial on how to conduct Systematic review and Meta-analysis, visit http://training.cochrane.org/about and click on “Learning resources”.