CARPE DIEM – PEM Prediction of Pneumonia Complications

June 17, 2021

Written by Clay Smith

Spoon Feed
Clinician gestalt leaves much to be desired in determining which children with community acquired pneumonia (CAP) will develop subsequent complications.

Why does this matter?
No prediction tools are available to risk stratify pediatric patients with CAP, meaning clinicians rely on their gestalt in determining the probability complication will occur. How good is clinician gestalt in predicting CAP complications?

CARPE DIEM – seize the opportunity to develop a clinical decision rule
This was a prospective study of 634 pediatric ED patients with radiographically confirmed CAP. Pneumonia complications occurred in 37 children (5.8%). PEM clinicians at Cincinnati Children’s went on the record as to their gestalt that the children would develop complications, “such as respiratory failure, empyema or effusion, lung abscess or necrosis, metastatic infection, sepsis or septic shock, or death.” After the ED workup, they prospectively assigned a probability of CAP complications: <1%, 1-5%, 5-10%, 10-25%, 25-50%, 50-75%, or 75-100%. Physician gestalt was lackluster, with overall accuracy (AUC) 0.747. Less experienced physicians performed even worse, with an accuracy of 0.693. At the extremes of low or high, diagnostic accuracy was good, 94.6% sensitivity, 99.1% NPV at the <1% level and mid to high 90s% specificity, mid-30s% PPV at the >10% level. The problem is, the majority of kids were ranked in the 1-10% range. In this range, sensitivity was in the 30s to mid-50s%, specificity 75-90%. This is an area in which a clinical decision tool may help improve physician gestalt.

Source
Predictive Value of Clinician “Gestalt” in Pediatric Community-Acquired Pneumonia. Pediatrics. 2021 May;147(5):e2020041582. doi: 10.1542/peds.2020-041582.