A cross-sectional study was conducted. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) for cross-sectional checklist was followed.
Inclusion criteria
All medical journals that published clinical studies, systematic reviews/meta-analyses, and clinical practice guidelines, which were published in China including Mainland China, Hong Kong, Macau, and Taiwan were included. There was no language of publication restrictions.
Exclusion criteria
Journals that have ceased publication were excluded. Journals lacking official websites were also excluded.
Identification of Chinese medical journals
The Chinese Biomedical Literature Service System (CBM) [18], China National Knowledge Infrastructure (CNKI) [19], Wanfang Data [20], and VIP Chinese Medical Journal Databases [21] were systematically searched for the medical journals from mainland China. The Hong Kong Macau Periodicals Network [22], HKInChip [23], Macau Periodical Index [24], and Airiti Library [25] were systematically searched for the medical journals from Hong Kong and Macau. The Airiti Library was used to search for medical journals from Taiwan. All medical journals in the databases and networks listed under the heading “Journal Navigation” were examined. We determined whether the journal included clinical studies, systematic reviews/meta-analyses, and clinical practice guidelines by i) the classification of journals in the databases, ii) the texts on the introduction of journals and submission guidelines for authors, and iii) whether clinical studies, systematic reviews/meta-analyses, and clinical practice guidelines were included in last year’s issues. If the journal had a searchable table of contents, we used the keywords “case report”, “case series”, “observational study”, “cohort study”, “cross-sectional study”, “case–control study”, “controlled trial”, “clinical trial”, “clinical study”, “systematic review”, “meta-analysis”, and “clinical practice guideline” to search for the publication of clinical studies. The journals were searched and screened by two researchers (LZ and JC), confirmed by a third researcher (YD) from December 2020 to January 2021.
Extracting data
First, the basic characteristics of the eligible journals were extracted; these included the Chinese and English names of the journal, publication place, publication institution/publisher, subjects, languages, the journal impact factor (JIF), and the official website address of each journal. Subjects were classified by referring to the journal discipline navigation within the CNKI [19]. The JIF from Science Citation Index (SCI)/Science Citation Index Expanded (SCIE) through the Web of Science database [26] were extracted for the medical journals which are included in SCI/SCIE. The JIF from the journal citation reports in the Chinese Science Citation Index [27], which is currently the most complete citation database of Chinese journal articles on mainland China, was selected as complementary data.
Second, the endorsement of reporting guidelines [3,4,5,6,7,8,9,10, 28] was extracted by systematically searching the official websites of eligible journals. For example, whether the CONSORT statement and its extensions were mentioned in the instruction for authors, author guidelines, peer reviewer guidance, editorial policies, or other relevant directions for authors of a journal and its recommendation level were extracted using a standardized form. We assessed the level of reporting guideline endorsement with a modified 5-level evaluation tool, namely i) active strong, ii) active weak, iii) passive moderate, iv) passive weak, and v) none. For the evaluation tool, we added two conditions to “active weak” and “passive moderate” based on a reference [28].
A completed CONSORT [28, 29] checklist and/or a flow diagram with article submission was assessed as “active strong”. “Active weak” was assessed as the journal “encourages” or “should” reference or follow a specific guideline; priority publication if the manuscript follows a specific guideline. “Passive moderate” were assessed as adhering to “relevant” RGs; abstracts are required to follow a specific guideline. Preparing the manuscripts according to the International Committee of Medical Journal Editors (ICMJE) was assessed as “passive weak”. No mention of any reporting guidelines was assessed as “none”.
The endorsement types of other reporting guidelines, including PRISMA, STROBE, SPIRIT, STARD, TRIPOD, CARE, AGREE, and RIGHT, which are the basic reporting guidelines related to clinical studies, systematic reviews/meta-analyses, and clinical practice guidelines were also extracted according to the above evaluation criteria.
All the above information was extracted and assessed by two researchers (LZ and JC) during the period from January to August 2021. After the first extraction, 10% of records were double-checked by two researchers (YD and YM). When disagreements happened, the judgment was made by senior researchers (ZB and DM). Primary data sources (i.e., website pages) were recorded; relevant text describing guideline endorsement was extracted and coded into a standard data extraction sheet in Excel. All the original data has been submitted as an open-source data set on the Open Science Framework platform.
Statistical analyses
All data were collected and recorded in Microsoft Office Excel (Version 2016). Basic characteristics of included journals (clinical contents, language, publication place, journals indexed databases, and journal impact factor), endorsement type of reporting guidelines, were presented using descriptive statistics such as counts (n) and percentages (%). The bar and pie charts made by Excel 2016 were used to show the results of subjects, language, and journals indexed databases. A heat map was generated using Tableau (Version 2018.3.2) to present the number of journals in different publication places. Logistic regression was used to analyze the influencing factors of reporting guidelines endorsement. The factors associated with the endorsement of reporting guidelines were analyzed by logistic regression using SPSS (Version 25.0). The endorsement types of CONSORT statement including “active strong”, “active weak”, “passive moderate”, and “passive weak” were used as a positive outcome. The endorsement type of CONSORT, “none” was considered as a negative outcome. Whether the journal is included in the SCI/SCIE database, whether the journal is for traditional Chinese medicine (TCM), and whether the publication language of the journal includes English were included as the independent variables.