There is only one systematic review without numerical meta-analysis [11].
We verified the numerical precision of Meta-Analyst by comparing its output with that from standard meta-analysis routines in Stata over a large database of 11,803 meta-analyses of binary outcome data, and 6,881 meta-analyses of continuous outcome data from the Cochrane Library of Systematic Reviews.
It contains 6 sheets with primarily worksheet formulas and 10 sheets with various kinds of pre-calculated graphical and numerical results from meta-analytical tests.
If this information was not given in sufficient numerical detail in the meta-analysis, we sought to complement it by extracting information from the meta-analysis forest plot using software for image digitisation (Engauge Digitizer 4.1).
Although this numerical requirement was met, meta-analysis was prevented by heterogeneous units of analysis between studies.
Studies where the data was represented in graphical format only, or if there was insufficient raw numerical data to perform meta-analysis, study authors were contacted [ 21- 25]; we successfully obtained further information only for three of these studies [ 21- 23].
To compare the performance of the meta-heuristics, random numerical instances are evaluated by four criteria involving mean ideal distance, spread of non-dominance solution, the number of Pareto solutions, and CPU time.
Meta-analysis uses numerical tools to pool data and to estimate a summary effect size for the comparison of two interventions from a set of randomised controlled trials identified in a systematic review.
Two papers were excluded because relative fitness measures were illustrated graphically, without providing the numerical measures necessary for the meta-analysis (Björkholm et al. 2001; Huitric et al. 2010).
Numerical tests confirms that the proposed meta-heuristic is able to generate proper Pareto solutions considering all of the objectives for decision maker.
Researchers also need to better report their studies in terms of describing the intervention (to allow replication), the study methodological features (to allow judgement of validity), and the numerical and statistical results (to allow meta-analyzing the results across studies).
