Skip to main content

Table 4 Performance characteristics of classification and regression tree models for predicting community-acquired pneumonia with a bacterial pathogen detected in patients presenting with a lower respiratory tract infection at outpatient clinics in Tanzania

From: Clinical sign and biomarker-based algorithm to identify bacterial pneumonia among outpatients with lower respiratory tract infection in Tanzania

 

Prediction of community acquired pneumonia with a bacterial pathogen detected

 

Patients with a lower respiratory tract infection

Patients with radiological pneumonia

 

All variables

RR ≥ 32/min and PCT ≥ 2.9

RR ≥ 32/min and PCT ≥ 0.25 µg/l

RR ≥ 32/min and PCT ≥ 0.25 µg/l

Sensitivity

88%

88%

94%

94%

Specificity

87%

88%

82%

87%

Negative likelihood ratio

0.1

0.1

0.1

0.1

Positive likelihood ratio

6.8

7.5

5.1

7.1

Negative predictive value

98%

87%

99%

93%

Positive predictive value

56%

89%

48%

89%

  1. The identified algorithm was also tested to predict community-acquired pneumonia with a bacterial pathogen detected among patients with radiological pneumonia
  2. PCT procalcitonine, RR respiratory rate