TABLE 2

Effect of Specialty on 279 Pediatricians’ Planned IV Antibiotic Durations Beyond 48 Hours in a Febrile Neonate With UTI

SpecialistnActual Average Additional Days, d (95% CI)Expected Effect of Specialist Versus General Pediatrician on Additional Daysa
Baseline Model, d (95% CI)bFull Model, d (95% CI)c
General pediatrician1292.4 (1.9 to 3.0)ReferentReferent
Neonatologist814.2 (3.0 to 5.4)+1.7 (+0.4 to +3.0)+2.2 (+0.8 to +3.7)
ID specialist522.7 (1.7 to 3.7)+0.3 (−0.8 to +1.4)+1.2 (−0.4 to +2.8)
Other specialist172.0 (0.7 to 3.3)−0.4 (−1.9 to +1.0)−0.5 (−1.8 to +0.8)
  • a Negative binomial regression model with the dependent variable of total days of IV antibiotics beyond 48 h.

  • b Baseline model included specialty, coded into the 4 categories listed in the first column.

  • c In addition to specialty, the full model included the following respondent characteristics from Table 1 as main effects. Nominal characteristics (sex, years in practice, location, proportion of clinical time managing inpatients, practice setting, and compensation) were coded with indicator variables. Annual experience with UTI was entered as an ordinal variable (<1 case = 0, 1–5 cases = 1, 6–12 cases = 2, and 12 or more cases = 3). In addition, backward selection of candidate interaction terms led to retention of sex by compensation model and sex by location.