Skip to main content

Table 1 Clinician predictions

From: How good is our diagnostic intuition? Clinician prediction of bacteremia in critically ill children

Provider Predictions Blood Culture Result Sensitivity (95% CI) Specificity (95% CI) Positive Predictive Value (95% CI) Negative Predictive Value (95% CI)
Positive (n) Negative (n)
Overall Prediction of Culture Result Positive (n) 41 107 82% (0.68,0.91) 69% (0.64, 0.74) 28% (0.21, 0.36) 96% (0.93, 0.98)
Negative (n) 9 241
RN Prediction of Culture Result Positive (n) 9 29 90% (0.54,0.99) 61% (0.49, 0.72) 24% (0.12, 0.41) 98% (0.87, 0.99)
Negative (n) 1 45
FLP Prediction of Culture Result Positive (n) 6 26 86% (0.42, 0.99) 67% (0.55, 0.77) 19% (0.08, 0.37) 98% (0.89, 0.99)
Negative (n) 1 52
Fellow Prediction of Culture Result Positive (n) 17 30 81% (0.57, 0.94) 75% (0.66, 0.82) 36% (0.23, 0.52) 96% (0.89, 0.99)
Negative (n) 4 90
Attending Prediction of Culture Result Positive (n) 9 22 75% (0.43, 0.93) 71% (0.59, 0.81) 29% (0.15, 0.48) 95% (0.84, 0.99)
Negative (n) 3 54
  1. Blood culture results compared to clinician prediction and characteristics of clinician prediction, separated by provider type. This table illustrates a series of classic “2 × 2” tables to calculate test characteristics sensitivity, specificity, positive and negative predictive values. For example, among RNs, the sensitivity was 90% or 9 of 10 positive cultures were correctly predicted to be positive. Abbreviations: RN registered nurse, FLP front line provider (in our unit, a resident or nurse practitioner), n number of observations, CI confidence interval