356 - Non-specific Pediatric Point of Care Lung Ultrasound Factors that Influence Medical Decision Making in the Context of Clinical Uncertainty for Pneumonia

Poster Number: 356
Publication Number: 356.205

Thomas Lee, Children's Hospital Los Angeles, Pasadena, CA, United States; Patricia T. Acharya, Children's Hospital Los Angeles, Los Angeles, CA, United States; Todd P. Chang, Children's Hospital Los Angeles, Los Angeles, CA, United States; Marsha A. Elkhunovich, Keck School of Medicine of the University of Southern California, Children's Hospital Los Angeles, Studio City, CA, United States

Background: Pediatric emergency medicine (PEM) providers use lung point of care ultrasound (POCUS) to help guide medical decision making in respiratory illnesses. Interpretation of lung findings positive for pneumonia (PNA) on POCUS can vary, with little consensus on non-specific factors due to no easily obtainable gold standard.

Objective: The aims of this study were: 1) to determine how pediatric providers use non-specific POCUS findings to diagnose PNA with script concordance testing (SCT), 2) to evaluate their accuracy in detecting findings on lung POCUS, and 3) to identify provider factors that influence the tendency to diagnose PNA.

Design/Methods: We drafted 24 clinical vignettes with POCUS cine clips using SCT methodology. The primary outcome variables were provider impression for PNA after ultrasound (US) review using a 5-point Likert scale and an SCT score measuring concordance to the reference panel. The secondary outcome variables were accuracy in detecting lung POCUS findings against the author panel, and tendency a provider would diagnose PNA, measured as a sum of post-US Likert answers. Spearman’s Rank was used for aim 1 with a Bonferroni correction. Mann Whitney U tests were used for aims 2 and 3.

A “reference panel” (n=10) and “general participants” (n=14) were recruited. The reference panel was chosen for their POCUS expertise, and general participants were PEM providers recruited through the P2 network, a pediatric POCUS listserv, and the authors’ home institution.

Results: Large and medium consolidations and confluent b-lines had the strongest correlation with a high Likert to diagnose PNA and give antibiotics (Rho 0.37, 0.18, 0.20, p < 0.001). Effusion and pleural line abnormalities were also correlated (Rho 0.125, p < 0.003, Rho 0.126, p < 0.002). Focal multiple b-lines and small ( < 0.5cm) consolidations were not significantly correlated with a PNA. Accuracy for US factors ranged from 68-81% against the author panel. Training, expertise, US usage, and demographics did not significantly affect post-US SCT scores, accuracy, or tendency to diagnose PNA.Conclusion(s): This study showed a novel use of SCT to assess how lung POCUS factors affect medical decision making in the context of clinical uncertainty. While some non-specific factors that were not previously conclusive for bacterial PNA were well correlated, others remained equivocal.