Abstract
W
The subjects covered by this meta-analysis are broad. Although all articles that have only reported the frequency of antibiotic resistance were not included in this meta-analysis, those that have estimated the rate of antibiotic resistance but had other aims have been included, which could lead to bias. The study's aim should be narrowed to report the pooled prevalence of antibiotic resistance, and entire articles that have reported the antibiotic resistance rate should also be included. Inclusion criteria should be optimized for this aim.
As a search strategy, they conducted a systematic literature search. The combinational search was conducted applying the following strategy (in the MEDLINE for example): “((Biofilm formation) OR Biofilm production) AND Pseudomonas aeruginosa) AND Prevalence) AND Iran)).” There are no MeSH terms in the search syntax, and thus, the search results do not include the MeSH terms such as P. aeruginosa. In addition, due to the absence of quotation marks for compound words such as “biofilm formation,” the combined keywords have been searched separately. For example, the articles that included “biofilm” or “formation” keyword are included in search results.
According to the article, the analysis was performed based on all the antibiotic resistance test results, regardless of which laboratory standard was used to interrupt antimicrobial resistance breakpoints such as CLSI or EUCST. In contrast, in all antibiotics, the resistance rate heterogeneity is significantly high. It is necessary to find the source of heterogeneity if possible. One of the inclusion criteria should be the use of the standard guidelines. The subgroup analysis by guideline category or antimicrobial susceptibility test method could also help find the sources of heterogeneity.
The Egger tests and funnel plots were used to investigate the publication bias and small study effects. The Eggers test result was significant, according to which the null hypothesis is rejected; the null hypothesis is there is no small study effect. Also, the funnel plot showed the occurrence of publication bias and, therefore, the meta-analysis was performed in the presence of publication bias. As a solution, some methods have been described to improve the estimation of effect size, such as Trim and fill 2 and Fail-safe N. 3 However, these methods can be useful only if the sample size cannot get any larger. Nevertheless, in this meta-analysis, a more extensive search to include articles beyond the included areas and find out about gray articles or unpublished data would be a better choice.
Footnotes
Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received.