Statistical analysis
Statistical analyses were performed using the STATA statistical analysis
software (StataCorp. 2017. Stata Statistical Software: Release 15.
College Station, TX: StataCorp LLC.) Proportional meta-analysis was
performed using the random effects model. In the random-effects model,
the observed difference between the proportions is not entirely
attributed to error in sampling. The pooled estimate describes the mean
of the distribution of the estimated parameters. The diagnostic adequacy
and the diagnostic yield of EBUS-TBNA/EUS-B-FNA were calculated for the
finally selected studies for meta-analysis. Proportions with 95%
confidence intervals (CI) were calculated for individual studies. The
data was subsequently pooled to estimate a summary proportion with 95%
CI representing the combined diagnostic adequacy/yield. Using Metaprop,
95% confidence intervals were calculated using the score statistic and
the exact binomial method. This method incorporates the Freeman-Tukey
double arcsine transformation of proportions. Within-study variability
is modelled using the binomial distribution using this method.