Table 5. Diagnostic accuracy at each specific cut-point in distinguishing mild TBI patients with positive CT findings from negative CT findings and in distinguishing patients who required neurosurgical intervention from patients who did not.

	At & above cut-point		Below cut-point		Sensitivity (95%CI)	Specificity (95%CI)
	n	(%)	n	(%)
Positive CT findings
Model-based score (cut-point 2)	≥2		<2
Positive CT findings	242	(99.2)	2	(0.8)	99.2	8.2
Negative CT findings	845	(91.9)	75	(8.2)	(97.1,99.9)	(6.5,10.1)
Clinical-based score (cut-point 2)	≥2		<2
Positive CT findings	239	(98.0)	5	(2.1)	98.0	16.2
Negative CT findings	771	(83.8)	149	(16.2)	(95.3,99.3)	(13.9,18.7)
Canadian CT Head Rule	Presence		Absence
Positive CT findings	214	(87.7)	30	(12.3)	87.7	22.7
Negative CT findings	711	(77.3)	209	(22.7)	(82.9,91.5)	(20.0,25.6)
Surgical intervention required
Model-based score (cut-point 7)	≥7		<7
Surgical intervention required	57	(100)	0	(0)	100	33.2
No surgical intervention required	739	(66.8)	368	(33.2)	(93.7,100)	(30.5,36.1)
Clinical-based score (cut-point 3)	≥3		<3
Surgical intervention required	57	(100)	0	(0)	100	28.3
No surgical intervention required	794	(71.7)	313	(28.3)	(93.7)	(25.6,31.0)
Canadian CT Head Rule	Presence		Absence
Surgical intervention required	54	(94.7)	3	(5.3)	94.7	21.3
No surgical intervention required	871	(78.7)	236	(21.3)	(85.4,98.9)	(18.9,23.9)
CT ordering proportion
Model-based score (cut-point 7)	796	(68.4)
Clinical-based score (cut-point 3)	851	(73.1)
Canadian CT Head Rule	925	(79.5)

Comparison of CT ordering proportions between our newly-derived models and the classic CCHR.

Abbreviations: CI, confidence interval; CT, computed tomography.