Table 3.

Accuracy of Demirjian’s classification of the lower third molar in estimating age of young Ugandans

Age cut-off (years)	Tooth	No.	Sensitivity (%)	Specificity (%)	AUC (95% CI)	PPV	NPV	+LR	−LR
12	38	496	92.0	80.3	0.86 (0.81 to 0.91)	0.97	0.63	4.7	0.1
	48	494	92.9	80.3	0.87 (0.82 to 0.91)	0.97	0.66	4.7	0.1
14	38	496	97.4	83.2	0.90 (0.87 to 0.93)	0.93	0.93	5.8	0.0
	48	494	97.4	82.9	0.90 (0.87 to 0.93)	0.93	0.93	5.7	0.0
16	38	496	68.8	97.4	0.83 (0.80 to 0.86)	0.97	0.73	26.7	0.3
	48	494	68.7	96.9	0.83 (0.80 to 0.86)	0.96	0.73	22.5	0.3
18	38	496	88.0	87.8	0.88 (0.85 to 0.91)	0.79	0.94	7.2	0.1
	48	494	85.9	86.7	0.86 (0.83 to 0.90)	0.77	0.92	6.5	0.2

No. is number of observations. The difference between the test sample number, 501 and number of teeth on each side (Right—494 and left—496) is accounted for by missing teeth or those that had not yet commenced mineralisation. PPV, NPV, +LRF and −LR are positive predictive value, negative predictive value, positive likelihood ratio and negative likelihood ratio, respectively