Factors associated with tracheostomy-associated infection treatment: A multicenter observational study.

Abstract

To characterize factors that influence the decision to treat suspected pediatric bacterial tracheostomy-associated respiratory infections (bTRAINs; e.g., pneumonia, tracheitis). We conducted a multicenter, prospective cohort study of children with pre-existing tracheostomy hospitalized at six children's hospitals for a suspected bTRAIN (receipt of respiratory culture plus ≥1 doses of an antibiotic within 48 h). The primary predictor was respiratory culture growth categorized as Pseudomonas aeruginosa, P. aeruginosa + ≥1 other bacterium, other bacteria alone, or normal flora/no growth. Our primary outcome was bTRAIN treatment with a complete course of antibiotics as documented by the discharge team. We used logistic regression with generalized estimating equations to identify the association between our primary predictor and outcome and to identify demographic, clinical, and diagnostic testing factors associated with treatment. Of the 440 admissions among 289 patients meeting inclusion criteria, 307 (69.8%) had positive respiratory culture growth. Overall, 237 (53.9%) of admissions resulted in bTRAIN treatment. Relative to a negative culture, a culture positive for P. aeruginosa plus ≥1 other organism (adjusted odds ratio [aOR] 2.3; 95% confidence interval [CI] 1.02-5.0)] or ≥1 other organism alone (aOR: 2.8; 95% CI: 1.4-5.6)] was associated with treatment. Several clinical and diagnostic testing (respiratory Gram-stain and chest radiograph) findings were also associated with treatment. Positive respiratory viral testing was associated with reduced odds of treatment (aOR: 0.5; 95% CI: 0.2-0.9). Positive respiratory cultures as well as clinical indicators of acute illness and nonculture test results were associated with bTRAIN treatment. Clinicians may be more comfortable withholding antibiotics when a virus is identified during testing.