Table 3

Independent predictors of days on therapy (DOT) in a zero-inflated Poisson model

FactorAdjusted OR (95% CI)p Value
Association with starting an antibioticManagement under IDP0.32 (0.10 to 0.99)0.05
Presentation with frailty syndrome3.87 (1.41 to 10.67)0.01
Received community antibiotics10.57 (3.91 to 28.60)<0.001
CRP on admission (per 1 mg/dL higher)1.02 (1.02 to 1.03)<0.001
Immunosuppression49.12 (2.85 to 847.89)0.007
White cell count on admission (per 1^9/L higher)1.22 (1.11 to 1.34)<0.001
FactorAdjusted RR (95% CI)
Association with greater days on therapy (DOT) per admissionManagement under IDP0.71 (0.54 to 0.93)0.01
Female0.77 (0.65 to 0.92)0.003
Age (per 10 years older)1.08 (1.02 to 1.13)0.005
Charlson score (per 1 unit higher)0.94 (0.90 to 0.99)0.01
Sepsis1.38 (1.15 to 1.65)0.001
Night admission1.40 (1.18 to 1.67)<0.001
CRP on admission (per 10 mg/dL higher)1.02 (1.00 to 1.03)0.02
Presentation with frailty syndrome0.77 (0.62 to 0.97)0.02
  • Sepsis and fever perfectly predicted antibiotic use and so therefore could not be included in the zero-component of the model. No other factors in table 1 were independently associated.

  • CRP, C reactive protein; IDP, infectious diseases physician; RR, rate ratio.