2.5 Data analysis methods
Median and range were used to describe continuous data, while frequency and percentage were used for summarizing categorical data. For individual items of the COS-STAD and COS-STAR, we summarized the frequency of “Yes” response for all included systematic reviews. We performed the Chi-square test, calculated odds ratio (OR) with 95% confidence intervals (95% CIs), and obtained a P-value for each item compared between group one (COSs with statistician or epidemiologist, funding, and with prior protocol) and group two (without statistician or epidemiologist, non-funding, and without protocol). Additionally, this method was used to compare the compliance for each item of COS-STAD between OG COSs and cancer COSs (the result of cancer COSs assessment was extracted from a previous study conducted by Elizabeth G).19 For each of the COS-STAR and COS-STAD checklist items, we calculated the mean compliance items and standard deviation (SD). The mean difference (MD) and 95% CI were calculated for each item to compare the overall compliance items between the groups. The MD value represents the difference in the mean total compliance items between group one and group two. Analyses were conducted using STATA (13.0; Stata Corporation, College Station, Texas, USA), and statistical significance was set at P < 0.05. The complete compliance rate of the methodological quality was calculated with the acquired number. Spearman correlation coefficient (r) was estimated to determine the linear association between average citations per year (dates from Web of Science) and the total complete compliance rates of COS-STAD and COS-STAR for each OG COS. The correlation analyses were conducted in IBM SPSS Statistics v. 24.0 (IBM Corp., Armonk, NY, USA), and statistical significance set at P < 0.05.
RESULT