Diagnostic styles of attending pediatricians, residents, and nurses in evaluating febrile children.

In order to compare the diagnostic styles of attending pediatricians (A), residents (R), and nurses (N) in judging the degree of illness of a febrile child prior to physical examination, we studied the frequency with which they used 139 different history variables and 186 observation variables in evaluating 162 consecutive children less than or equal to 24 months of age with temperature greater than or equal to 38.3 C. Because individual history and observation variables were not used with sufficient frequency, comparison of diagnostic styles based on the individual variables was not possible. The individual history and observation variables were then combined into 25 history and 30 observation categories respectively. One observation category, "Eyes," was used with sufficient frequency to allow comparison of A, R, and N diagnostic styles. In the Eyes category, there were 11 variables describing visual response to stimuli (e.g., looking at the observer) and 10 variables describing appearance of the eyes (e.g., glassy, shiny). Attendings used variables describing visual response to stimuli significantly more frequently than R or N, which may indicate a greater facility in evaluating age-appropriate behavior and/or may be a technique to increase the sensitivity of clinical judgments. These data demonstrate the variety of diagnostic styles of A, R, and N. There is a need to define valid and reliable observation data to identify serious illness in febrile children.