Skip to main content
. 2012 Jan 23;12:9. doi: 10.1186/1471-2431-12-9

Table 4.

Test Characteristics of Selected HC Percentile Thresholds for Diagnosing Children with IEC, MGCM, or BESS

A B C D E G H I K L M N
Threshold Number in source population Number diagnosed with IEC, MGCM, or BESS Number above threshold Number above threshold with IEC, MGCM, or BESS Sensitivity
E/C
Specificity
(B-C-(D-E))/(B-C)
Positive predictive value
E/D
Likelihood ratio positive
G/(1-H)
Likelihood ratio negative
(1-G)/H
Number Needed to Screen
B/E
Number Needed to Test
D/E

Above CDC 95th 74,428 324 12,325 221 68% 84% 1.8% 4.2 0.4 337 56
Above WHO 95th 74,428 324 18,528 242 75% 75% 1.3% 3.0 0.3 308 77
Above PCN 95th 74,428 324 7,694 193 60% 90% 2.5% 5.9 0.4 386 40
Above CDC 97th 74,428 324 8,373 203 63% 89% 2.4% 5.7 0.4 367 41
Above WHO 97th 74,428 324 13,275 225 69% 82% 1.7% 3.9 0.4 331 59
Above PCN 97th 74,428 324 4,532 167 52% 94% 3.7% 8.8 0.5 446 27
Above CDC 99.6th 74,428 324 2,030 129 40% 97% 6.4% 15.5 0.6 577 16
Above WHO 99.6th 74,428 324 3,438 155 48% 96% 4.5% 10.8 0.5 480 22
Above PCN 99.6th 74,428 324 711 69 21% 99% 9.7% 24.6 0.8 1,079 10
Crossed 2 IMPL-CDC 64,015 321 29,206 223 69% 54% 0.8% 1.5 0.6 287 131
Crossed 2 IMPL-WHO 64,015 321 22,462 162 50% 65% 0.7% 1.4 0.8 395 139
Crossed 2 IMPL-PCN 64,015 321 13,831 156 49% 79% 1.1% 2.3 0.7 410 89
Crossed 4 IMPL-CDC 64,015 321 5,727 103 32% 91% 1.8% 3.6 0.7 622 56
Crossed 4 IMPL-WHO 64,015 321 4,372 66 21% 93% 1.5% 3.0 0.9 970 66
Crossed 4 IMPL-PCN 64,015 321 1,703 55 17% 97% 3.3% 6.7 0.8 1,143 30
Crossed 6 IMPL-CDC 64,015 321 490 17 5% 99% 3.5% 7.1 1.0 3,766 29
Crossed 6 IMPL-WHO 64,015 321 556 17 5% 99% 3.1% 6.3 1.0 3,766 33
Crossed 6 IMPL-PCN 64,015 321 130 10 3% > 99% 7.7% 16.5 1.0 6,402 13

IEC (intracranial expansive condition); MGCM (metabolic or genetic condition associated with macrocephaly); BESS (benign enlargement of the subarachnoid spaces); CDC (Centers for Disease Control head circumference growth curves); WHO (World Health Organization head circumference growth curves); PCN (primary care network head circumference growth curves); IMPL (increasing multiple percentile lines). The negative predictive value (C-(D-E))/(C-D) was 99.9% for all thresholds.