The present protocol has been registered within the Open Science Framework (registration: https://osf.io/aj2df) and is being reported in accordance with the reporting guidance provided in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) statement [12] (see checklist in Additional file 1). Any amendments made to this protocol when conducting the study will be outlined in the Open Science Framework and reported in the final manuscript
Search strategy and study selection
We will leverage a pre-existing sample of child health RCTs published in 2007 (n = 300) [11] used by our team in previous study of reporting quality of pediatric RCTs to answer our review question: What is the magnitude and direction of change in P value and Bayesian analysis reported in RCTs in child health research published over 10 years, if any? Details of the search strategy and study selection methods for the sample are available in our previous publications [11, 13]. We will replicate these methods to identify a comparable sample of child health RCTs published in 2017. The final sample will include 600 child-health RCTs, 300 published in each of 2007 and 2017.
To identify a sample of studies published in 2017, a research librarian will execute an updated literature search in the Cochrane Central Register of Controlled Trials (see Additional file 2). The Cochrane Central Register of Controlled Trials includes randomized and quasi-randomized controlled trials indexed in MEDLINE and EMBASE, hand-searched results, gray literature sources, and Cochrane Review Groups Specialized Registers of trials [14]. All retrieved records will be imported into EndNote (v. X9, Clarivate Analytics, Philadelphia, PA, USA) and exported to an Excel (v. 2016, Microsoft Corporation, Redmond, WA, USA) workbook for screening. We will randomly order the citations using the random numbers generator in Excel. Next, one reviewer will screen the titles and abstracts to identify the first 300 child health RCTs. These should be easily identifiable by title and abstract; however, in the unlikely (per experience) event that a record is deemed ineligible during data extraction, we will substitute it with the next relevant record. We will include the first 300 eligible citations from the randomly ordered list to make the sample size consistent with the previous publications [11, 12].
Eligible studies will include RCTs in health research conducted among individuals aged 21 years and below [15]. We will employ identical selection criteria used in the 2007 and 2012 samples to maintain consistency and comparability with earlier findings [11]. Literature will be limited to published full-text articles in the English language. There will be no restriction on settings in which the study was conducted, intervention, comparator, or the type of outcome.
Data extraction
We will adopt part of the data extraction form from the 2007 and 2012 studies [11], with some additions to gain the information on P values and Bayesian analysis. We will pilot test the form using three studies from 2007 and 2017 for completeness and accuracy. Data will be extracted by a single reviewer using Excel (v. 2016, Microsoft Corporation, Redmond, WA, USA), with verification by a second reviewer. Disagreements will be resolved by discussion between reviewers or by involving another reviewer when necessary. We will extract data on characteristics of the publication, study design, intervention, control, trial conduct, study sample, sample size, hypothesis, primary objective, diagnostic criteria, recruitment strategies, funding, data monitoring committee (DMC), and specific statistical attributes of frequentist and Bayesian analysis/methods that are related to the primary outcome (see Additional file 3). We will extract data for the primary outcome, and if not clearly stated, we will use the objective outcome (e.g., mortality, hospitalization), the outcome used to calculate sample size, or the first outcome reported in the results. We will also use trial registers and published protocols (when cited in the publication) to supplement data extraction. When not cited in the publications, we will search for trial registers in the International Clinical Trials Registry Platform and the Google databases. We will not appraise the risk of bias of the included studies.
Data analysis
We will present summary characteristics and results of all trials in a tabular form. We will consider analyzing the data using Stata (v. 16.1; StataCorp, College Station, Texas, United States) or R [16] and JAGS statistical software [17]. The analysis of extracted data will be mainly descriptive, using counts and percentages for categorical data, and means and/or medians (with standard deviations and/or ranges) for continuous data. We will compare extracted data from the 2017 sample with 300 RCTs published in 2007 to assess 10-year change in the reporting of P value and Bayesian analysis. The difference between the two periods will be assessed using both the frequentist and Bayesian methods. We will present the proportion (%) of studies reporting P value and Bayesian analysis in 2007 and 2017 in graphical forms. We will also present specific characteristics of studies, which used Bayesian analysis in tabular forms, if any. We will present the clustering around P values of significance, if observed in the samples.