Skip to main content

Table 2 Small-study effects and time trends

From: Small-study effects and time trends in diagnostic test accuracy meta-analyses: a meta-epidemiological study

  Accuracy measurea Relative increaseb(95% CI)Pvalue T2 vs. T1c(95% CI)Pvalue T3 vs. T1c(95% CI)Pvalue
Number of diseased Sensitivity 1.11 (0.98 to 1.26) P = 0.09 1.08 (0.87 to 1.34) P = 0.50 1.22 (0.99 to 1.51) P = 0.06
Number of non-diseased Specificity 1.00 (0.99 to 1.02) P = 0.49 1.05 (0.83 to 1.33) P = 0.66 0.97 (0.73 to 1.28) P = 0.82
Sample size DOR 1.01 (1.00 to 1.03) P = 0.07 1.15 (0.94 to 1.40) P = 0.16 1.26 (0.96 to 1.64) P = 0.09
Time since first publication Sensitivity 0.89 (0.80 to 0.99) P = 0.04 0.61 (0.11 to 3.39) P = 0.57 0.59 (0.13 to 2.67) P = 0.49
Time since first publication Specificity 1.04 (0.90 to 1.19) P = 0.60 1.07 (0.85 to 1.35) P = 0.55 1.00 (0.76 to 1.32) P = 0.99
Time since first publication DOR 0.94 (0.80 to 1.10) P = 0.42 0.94 (0.71 to 1.25) P = 0.68 0.93 (0.71 to 1.21) P = 0.57
  1. aThe analyses were performed on the natural logarithm of the DOR and on the logit scale for sensitivity and specificity; brelative increase for sensitivity, specificity, and DOR is reported per increase in 100 diseased, non-diseased, or total participants, respectively. For time since first publication, the relative increase is reported per 5 year increase; cT1 is the lowest tertile of sample size or time since first publication, T3 the highest. DOR, diagnostic odds ratio.