An international assessment of a web-based diagnostic tool in critically ill children.

Improving diagnostic accuracy is essential. The extent of diagnostic uncertainty at patient admission is not well described in critically ill children. Therefore, we studied the extent that pediatric trainee diagnostic performance could be improved with the aid of a computerized diagnostic tool. Data regarding patient admissions to five Pediatric Intensive Care Units were collected. Information included patients' clinical details, admitting team's diagnostic workup and discharge diagnosis. An attending physician assessed each case independently and suggested additional diagnostic possibilities. Diagnostic accuracy was calculated using the discharge diagnosis as the gold standard. 206 out of 927 patients (22.2%) admitted to the PICUs did not have an established diagnosis at admission. The trainee teams considered a median of three diagnoses in their workup (IQR 3-5) and made an accurate diagnosis in 89.4% cases (95% CI 84.6%-94.2%). Diagnostic accuracy improved to 92.5% with use of the diagnostic tool alone, and to 95% with the addition of attending physicians' diagnostic suggestions. We conclude that a modest proportion of admissions to these PICUs were characterized by diagnostic uncertainty during initial assessment. Although there was a relatively high accuracy rate of initial assessment in our clinical setting, it was further improved by both the diagnostic tool and the physicians' diagnostic suggestions. It is plausible that the tool's utility would be even greater in clinical settings with less expertise in critical illness assessment, such as community hospitals, or emergency departments of non-training institutions. The role of diagnostic aids in the care of critically ill children merits further study.further study.