Skip to main content

Table 3 95% Coverage probabilities and 95% confidence intervals for sensitivity estimates across 1000 simulations

From: Estimating sensitivity and specificity of diagnostic tests using latent class models that account for conditional dependence between tests: a simulation study

  

Culture

Model

 

GS Model

CIndep Model

CDP Model

CDN Model

CDPN Modela

Data generating mechanism

CIndep

0(0–0.3)

96.4(95.1–97.4)

97.4(96.3–98.3)

97.0(95.8–97.9)

96.8(95.6–97.8)

CDP

0(0–0.3)

94.7(93.2–96.0)

94.8(93.3–96.0)

93.5(91.8–94.9)

96.6(95.3–97.6)

CDN

0(0–0.3)

81.0(78.5–83.3)

79.1(76.5–81.5)

95.5(94.1–96.7)

99.2(98.5–99.6)

CDPN

0(0–0.3)

92.2(90.4–93.7)

84.2(81.8–86.4)

96.7(95.5–97.7)

99.5(98.9–99.8)

  

Serology

Model

 

GS Model

CIndep Model

CDP Model

CDN Model

CDPN Modela

Data generating mechanism

CIndep

96.2(95.5–96.7)

97.0(96.4–97.4)

96.8(96.3–97.3)

96.6(96.0–97.1)

96.8(96.2–97.3)

CDP

89.1(88.1–90.0)

80.4(79.1–81.6)

95.5(94.8–96.1)

79.4(78.1–80.6)

95.0(94.3–95.6)

CDN

95.4(94.7–96.0)

96.5(95.9–97.1)

96.5(95.8–97.0)

96.9(96.3–97.4)

97.2(96.6–97.7)

CDPN

88.4(87.4–89.4)

87.5(86.5–88.5)

97.4(96.8–97.8)

85.2(84.1–86.3)

97.0(96.4–97.5)

  1. Values in bold show scenarios where the upper limit of the confidence interval is less than 95%. Confidence intervals for coverage calculated using Jeffreys prior
  2. aTotal number of simulations summarised is not equal to 1000 for the CDPN model due to a number of simulations with convergence problems