Quality assessment in meta-analysis

Abstract

Background: An important characteristic of meta-analysis is that the results are determined both by the
management of the meta-analysis process and by the features of studies included. The scientific rigor of
potential primary studies varies considerably and the common objection to meta-analytic summaries is that
they combine results from studies of different quality. Researchers began to develop quality scales for
experimental studies, however now the interest of researchers is also focusing on observational studies.
Since 1980, when Chalmers developed the first quality scale to assess primary studies included in meta-
analysis, more than 100 scales have been developed, which vary dramatically in the quality and quantity of
the items included. No standard lists of items exist, and the used quality scales lack empirically-supported
components.
Methods: Two of the most important and diffuse quality scales for experimental studies, Jadad system and
Chalmers’ scale, and a quality scale used for observational studies, developed by Angelillo et al., are
described and compared.
Conclusion: The fallibility of meta-analysis is not surprising, considering the various bias that may be
introduced by the processes of locating and selecting studies, including publication bias, language bias and
citation bias. Quality assessment of the studies offers an estimate of the likelihood that their results will
express the truth