Stata versions

The strength of the association between CCND1 G870A polymorphism and cervical cancer was evaluated by ORs with 95% CI, comprised with allele contrast (A vs. G), codominant model (GA vs. GG, AA vs. GG), dominant model (GA + AA vs. GG), and recessive model (AA vs. GG + GA). The HWE of the control group was assessed, and a P value of less than 0.05 was considered significant disequilibrium. Stratified analyses were used to evaluate ethnicity, study design, and genotyping type technique. Heterogeneity was explored with the use of a chi-squared test, and the quantity of heterogeneity was measured by the I² statistic. I² values of 25%, 50%, and 75% represent low, moderate, and high heterogeneity, respectively. The OR of each model was estimated by using the fixed-effects model (Mantel–Haenszel method) when I² ≤ 50%; otherwise, the random-effects model (DerSimonian and Laird method) was used. Meta-regression analyses were performed to assess potential covariates that can predict intertribal heterogeneity. Publication bias was assessed on the basis of modified Egger’s bias test and Begg’s funnel plot. Statistical analysis was performed by using STATA versions 11.0 (Stata Corporation, College Station, TX). Two-sided P value (P < 0.05) was considered statistically significant.

We first examined missingness of data; all survey items had less than 2% missing data. We used mean imputation for missing responses to GHSQ items that were part of informal or formal help-seeking. We summarized key variables using descriptive statistics. We analyzed sociodemographic and clinical characteristics of participants who were Facebook users and nonusers using 2-sample t tests and Pearson Chi-square tests. To compare help-seeking intentions from different GHSQ source categories, we used multilevel mixed-effects linear regressions. We performed subgroup analyses to determine whether the pattern of help-seeking intentions differed when the sample was restricted to various groups of participants, including frequent Facebook users (defined as visiting at least daily), active Facebook users (defined as sharing, posting, or commenting on Facebook at least daily), and users of social media platforms other than Facebook. Analyses were performed using Stata versions 14.2 and 15.0 (StataCorp LLC), and 2-sided statistical significance was defined as P<.05.