Meta-analyses are usually the final step of systematic reviews and provide robust scientific evidence (1,2). The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) Statement is a useful tool to improve the reporting of meta-analyses (3). There are few data available on the current topics and characteristics of meta-analyses in dermatology or on the clarity and comprehensiveness of the information they report (4-8). This case study aims to describe the meta-analyses published in the British Journal of Dermatology (BJD) and assess whether the data they contain are reported fully according to the PRISMA Statement. We conducted a descriptive study of all meta-analyses indexed in PubMed between January 1, 2010, and December 31, 2014. To identify the manuscripts, we used PubMed to search all articles published in the BJD classified by MEDLINE as “meta-analysis.” We excluded all trials that did not involve human patients and all secondary analyses of previously published data. Study topics were not mutually exclusive. Two review authors collected the data, resolving any disagreements through discussion. We extracted the following data from each included study: country of origin, number of authors, research topic, funding (yes/no), type of funding, type of population, number and type of databases searched, initial and final number of studies, study type, and number of patients included. In addition, we applied the PRISMA 27-item checklist to all manuscripts. Twenty-seven meta-analyses were published during the study period. Germany was the main country where the studies were conducted (25.9%). The most common topic was psoriasis (n=11; 40.7%). The mean number of authors was 4.8. Thirteen (48.1%) documents received funding. Only five (18.5%) studies included the pediatric population. The mean number of databases consulted was four. All meta-analyses used the PubMed database; 18 (66.7%) used Cochrane and 17 (63.0%) used Embase. The mean number of included studies in the analyses was 31.1. The main study design was case-control (48.1%), followed by cohort (40.7%) and randomized controlled trials (33.3%) (Table 1). Most documents completed the 27 items on the PRISMA checklist correctly. In total, 51.9% of the reports were identified as both systematic reviews and meta-analyses and 48.1% as meta-analyses only. A structured summary was provided in 40.7% of cases. All the publications described the rationale in the introduction, and most (85.2%) correctly explained the objectives (including at least four of five sub-items). Regarding methods, only 18.5% of the studies specified a protocol or registration number. Eligibility criteria and information sources were explained in all the documents. Most of the documents fulfilled the criteria listed under search (96.3%), data collection process (88.9%), data items(85.5%), and summary measures (88.9%). In Methods and Results, the publications included the following items in the respectively indicated proportions: study selection (100.0% and 95.3%), risk of bias in individual studies (70.4% and 63.0%), synthesis of results (92.6 and 100.0%), risk of bias across studies (66.7 and 70.4%), and additional analyses (70.4 and 81.5%). All studies described the study characteristics, and 96.3% presented the results of individual studies. The summary of evidence, limitations, and conclusions were provided in all the publications. Similarly, all the documents detailed sources of funding (Table 2). To sum up, we observed that most of the studies in the BJD correctly recorded the items included in the PRISMA checklist. The meta-analyses mainly concerned psoriasis, which is consistent with a previous study on randomized trials in BJD finding that psoriasis was the main topic (5). A limitation of our study was that the meta-analyses included were published in a single journal, so the results are not generalizable to all dermatology journals. However, this is a novel study of PRISMA checklist compliance in dermatology meta-analyses, and we consider it important to validate and build on these results through larger studies including more journals specializing in dermatology.

References

PubMed