. 2026 Jun 22;17:1868789. doi: 10.3389/fphar.2026.1868789

TABLE 2.

Summary of the main algorithms used for signal detection.

Method	Formula	Threshold
ROR	$ROR = \frac{(a / c)}{(b / d)} = \frac{ad}{bc}$ $SE (lnROR) = \sqrt{(\frac{1}{a} + \frac{1}{b} + \frac{1}{c} + \frac{1}{d})}$	a≥3 and 95% CI (lower limit) > 1
PRR	$PRR = \frac{a / (a + b)}{c / (c + d)}$ $SE (lnPRR) = \sqrt{\frac{1}{a} ‐ \frac{1}{a + b} + \frac{1}{c} ‐ \frac{1}{c + d}}$ $95 % CI = e^{\ln (PRR) \pm 1.96 \sqrt{\frac{1}{a} ‐ \frac{1}{a + b} + \frac{1}{c} ‐ \frac{1}{c + d}}}$	a≥3 and PRR ≥2, χ² ≥ 4
BCPNN	$IC = \log_{2} \frac{p (x, y)}{p (x) p (y)} = \log_{2} \frac{a (a + b + c + d)}{(a + b) (a + c)} E (IC) = \log_{2} \frac{(a + γ 11) (a + b + c + d + α) (a + b + c + d + β)}{(a + b + c + d + γ) (a + b + α 1) (a + c + β 1)}$ $V (IC) = \frac{1}{{(\ln 2)}^{2}} \{[\frac{(a + b + c + d) ‐ a + γ ‐ γ 11}{(a + γ 11) (1 + a + b + c + d + γ)}] + [\frac{(a + b + c + d) ‐ (a + b) + α ‐ α 1}{(a + b + α 1) (1 + a + b + c + d + α)}] + [\frac{(a + b + c + d) ‐ (a + c) + β ‐ β 1}{(a + c + β 1) (1 + a + b + c + d + β)}]\}$ $γ = γ 11 \frac{(a + b + c + d + α) (a + b + c + d + β)}{(a + b + α 1) (a + c + β 1)}$ $IC ‐ 2 SD = E (IC) ‐ 2 \sqrt{V (IC)}$	IC025 > 0
EBGM	$EBGM = \frac{a (a + b + c + d)}{(a + c) (a + b)}$ $95 % CI = e^{\ln (EBGM) \pm 1.96 \sqrt{\frac{1}{a} + \frac{1}{b} + \frac{1}{c} + \frac{1}{d}}}$	EBGM05 > 2

Abbreviations: ROR, reporting odds ratio; PRR, proportional reporting ratio; BCPNN, bayesian confidence propagation neural network; MGPS, multi-item gamma Poisson shrinker; IC, information component; EBGM, empirical Bayes geometric mean; CI, confidence interval; χ2, chi-squared; IC025, lower limit of 95% two-sided CI, of the IC; EBGM05, lower limit of 95% one-sided CI, of EBGM.