Table 5 Diagnostic accuracy for detecting referent temperature above \(100.0\,^{\circ }{\mathrm{F}}\) using different NCIT screening thresholds (\(T^\star _{\text {IR}}\)) of \(91.1\,^{\circ }{\mathrm{F}}\) (left), \(93.9\,^{\circ }{\mathrm{F}}\) (middle), and \(96.7\,^{\circ }{\mathrm{F}}\) (right).

From: Proper use of noncontact infrared thermometry for temperature screening during COVID-19

 

\(T^\star _{\text {IR}}=91.1\,^{\circ }{\mathrm{F}}\)

\(T^\star _{\text {IR}}=93.9\,^{\circ }{\mathrm{F}}\)

\(T^\star _{\text {IR}}=96.7\,^{\circ }{\mathrm{F}}\)

Diagnostic accuracy of NCIT at 1 fta (Synthesized STRIDE cohort)

SP\(^{\mathrm{b}}\) (%)

17.1

80

99.6

SE (%)

98.2

60.1

6.56

FPR (%)

82.9

20

0.39

FNR (%)

1.76

39.9

93.4

PPV (%)

3.99

9.54

37.1

NPV (%)

99.6

98.3

96.8

TN (n)

95, 620

446, 976

556, 458

FP (n)

463,017

111,661

2179

FN (n)

343

7809

18, 230

TP (n)

19, 241

11, 776

1285

  1. \(^{\mathrm{a}}T^\star _{\text {IR}}\triangleq T^\star +\Delta \mu + \delta T,\ \Delta \mu = -6.15\,^{\circ }{\mathrm{F}},\ T^\star = 100.0\,^{\circ }{\mathrm{F}}\).
  2. \(^{\mathrm{b}}\) Specificity (SP), Sensitivity (SE), False Positive Rate (FPR), False Negative Rate (FNR), Positive Predictive Value (PPV), Negative Predictive Value (NPV), True Negative (TN), False Positive (FP), False Negative (FN), True Positive (TP).
Back to article page