Exploring reporting quality of systematic reviews and Meta-analyses on nursing interventions in patients with Alzheimer's disease before and after PRISMA introduction.

backgroundSystematic reviews (SRs) and meta-analyses (MAs) are distillation of current best available evidence, but are potentially prone to bias. The bias of SRs and MAs comes from sampling bias, selection bias and within study bias. So, their reporting quality is especially important as it may directly influence their utility for clinicians, nurses, patients and policy makers. The SRs and MAs on nursing interventions in patients with Alzheimer's disease (AD) have been increasingly published over the past decade, but the reporting quality of article has not been evaluated after the introduction of Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) Statement.

methodsAccording to the inclusion and exclusion criteria, we searched the databases including PubMed, EMBASE and The Cochrane Library from inception through October 16th 2018. Two reviewers independently selected articles and extracted data. The PRISMA checklist was adopted to evaluate reporting quality. Comparisons were made between studies published before (2001-2009) and after (2011-2018) its introduction.

resultsA total of 77 eligible articles, 18 (23.4%) were published before the PRISMA Statement and 59 (76.6%) were published afterwards. There was higher score after publication of the PRISMA Statement than before (20.83 ± 3.78 vs 17.11 ± 4.56, P <  0.05). There was an improvement in the following items after the PRISMA statement was released (P <  0.05): title (item 1, 50.0% vs 74.6%, OR = 3.10, 95CI%: 1.00-9.61), search (item8, 27.8% vs 57.6%,OR = 3.25, 95CI%: 1.14-9.28), study selection (item 9, 44.4% vs 81.4%,OR = 6.28, 95CI%: 1.93-20.37), Data collection process (item 10, 50.0% vs 76.3%,OR = 3.45, 95CI%:1.10-10.84), risk of bias in individual studies (item 12, 50.0% vs 83.1%, OR = 5.78, 95CI%:1.71-19.52), risk of bias across studies (item15, 5.6% vs 28.8%,OR = 3.60, 95CI%:1.04-12.43), study characteristics (item 18, 77.8% vs 98.3%, OR = 28.13, 95CI%:3.35-236.19), risk of bias with studies (item 19, 50.0% vs 83.1%, OR = 5.78, 95CI%:1.71-19.52), results in individual studies (item 20, 72.2% vs 94.9%, OR = 11.09, 95CI%:1.99-61.82), conclusions (item 26, 77.8% vs 98.3%, OR = 28.13, 95CI%:3.35-236.19). After controlling for the confounding factors, there were higher PRISMA score for systematic reviews including meta-analyses, protocol or registration, can't answer of RCT, journal source of SCI (Science Citation Index), manuscript length > 13 page and funding support.

conclusionSince the publication of the PRISMA Statement, there has been an improvement in the quality of reporting of SRs and MAs on nursing interventions in patients with AD. More endorsement by journals of the report guideline for SRs/MAs may improve articles reporting quality, and the dissemination of reliable evidence to nurses. We recommend authors, readers, reviewers, and editors to become more acquainted with and to more strictly adhere to the PRISMA checklist.