Skip to main content

Table 4 Accuracy measures for recursive partitioning and expert opinion algorithms

From: Developing model-based algorithms to identify screening colonoscopies using administrative health databases

Algorithm

Reference standard

Sensitivity%

Specificity%

PPV a%

NPV b%

(95% CI)

(95% CI)

(95% CI)

(95% CI)

Recursive partitioning with latent class outcome

Latent class indication

84.5

63.3

65.4

83.3

(81.5-87.5)

(59.7-67.0)

(61.9-68.9)

(80.0-86.5)

Recursive partitioning with endoscopist outcome

Endoscopist indication

85.1

62.2

66.5

82.6

(82.2-88.0)

(58.5-66.0)

(63.1-70.0)

(79.2-85.9)

El-Serag

Latent class indication

49.2

82.0

69.1

66.4

(45.1-53.4)

(79.1-84.9)

(64.6-73.7)

(63.3-69.6)

El-Serag

Endoscopist indication

49.3

83.0

71.9

65.0

(45.2-53.4)

(80.2-85.9)

(67.5-76.3)

(61.8-68.3)

  1. a Positive predicative value.
  2. b Negative predictive value.