Table 3.

Internal and external validation classification performance at different risk thresholds.

Risk thresholda	Sensitivityb (95% CI)	Specificityc (95% CI)	PPVd (95% CI)	NPVe (95% CI)
Internal validation
≥10%	0.93 (0.88–0.96)	0.39 (0.33–0.45)	0.47 (0.41–0.52)	0.91 (0.84–0.95)
≥20%	0.85 (0.79–0.90)	0.57 (0.51–0.62)	0.53 (0.47–0.59)	0.87 (0.81–0.91)
≥35%	0.73 (0.66–0.80)	0.72 (0.67–0.77)	0.60 (0.53–0.67)	0.82 (077–0.87)
≥50%	0.63 (0.55–0.70)	0.84 (0.79–0.87)	0.69 (0.61–0.76)	0.80 (0.75–0.84)
External validation
≥10%	0.81 (0.66–0.91)	0.55 (0.38–0.70)	0.67 (0.52–0.79)	0.72 (0.52–0.86)
≥20%	0.73 (0.57–0.85)	0.76 (0.59–0.87)	0.77 (0.61–0.88)	0.71 (0.55–0.84)
≥35%	0.65 (0.49–0.78)	0.88 (0.73–0.95)	0.86 (0.69–0.94)	0.69 (0.54–0.81)
≥50%	0.54 (0.38–0.69)	0.94 (0.80–0.98)	0.91 (0.72–0.97)	0.65 (0.50–0.77)

^aModel scores equal to or above this value are classified as suicidal.

^bSensitivity = true positives divided by sum of true positives and false negatives.

^cSpecificity = true negatives divided by sum of true negatives and false positives.

^dPositive predictive value = true positives divided by sum of true positives and false positives.

^eNegative predictive value = true negatives divided by sum of true negatives and false negatives.