Metric | |||||
---|---|---|---|---|---|
Group metrics | Non-IDP (241) | IDP (56) | All | Unadjusted RR (95% CI) | p Value |
Days of therapy (DOT)/100 admissions | 282 | 173 | 261 | 0.61 (0.50 to 0.76) | <0.001 |
Defined daily doses (DDDs)/100 admissions | 340 | 227 | 320 | 0.67 (0.55 to 0.80) | <0.001 |
Broad-spectrum DOT/100 admissions | 169 | 98 | 156 | 0.58 (0.44 to 0.77) | <0.001 |
Length of therapy (LOT)/100 admissions | 210 | 129 | 195 | 0.61 (0.48 to 0.78) | <0.001 |
DDDs/100 bed days | 66 | 38 | 60 | 0.47 (0.31 to 0.70) | <0.001 |
Patient-level metrics | n (%) or med (IQR) | n (%) or med (IQR) | Adjusted OR /RR (95% CI) | p Value | |
Given an antibiotic* | 85 (35%) | 14 (25%) | 99 (33%) | 0.25 (0.07 to 0.84) | 0.03* |
DOT† | 0 (0–5) | 0 (0–0.5) | 0 (0–5) | 0.71 (0.54 to 0.93) | 0.01† |
DDDs† | 0 (0–5.25) | 0 (0–0.13) | 0 (0–4.5) | 0.67 (0.53 to 0.90) | 0.006 † |
LOT† | 0 (0–5) | 0 (0–0.5) | 0 (0–3) | 0.69 (0.52 to 0.94) | 0.02† |
DOT if started on an antibiotic | 6 (5–9) | 6 (3–7) | 6 (5–9) | – | – |
LOT if started on an antibiotic | 5 (4–7) | 4 (2–6.5) | 5 (3–7) | – | – |
Broad-spectrum*,‡ antibiotic given | 70 (29%) | 11 (20%) | 81 (27%) | 0.53 (0.19 to 1.52) | 0.24* |
Broad-spectrum antibiotic DOT† | 0 (0–2) | 0 (0–0) | 0 (0–1) | 0.75 (0.54 to 1.03) | 0.08† |
Broad:narrow ratio | 408/271 (1.51:1) | 55/42 (1.31:1) | 463/313 (1.48:1) | – | – |
Given intravenous antibiotic* | 52 (22%) | 8 (14%) | 60 (20%) | 0.39 (0.12 to 1.27) | 0.12* |
Intravenous DOT† | 0 (0–0) | 0 (0–0) | 0 (0–0) | 0.76 (0.49 to 1.18) | 0.23† |
Intravenous DOT if given intravenous | 2 (1–3) | 1 (1–4.5) | 2 (1–4) | – | – |
*Multivariable logistic regression model based on backwards elimination from IDP plus table 1 factors.
†From multivariable zero-inflated Poisson regression models based on backwards elimination from IDP plus table 1 factors (see Methods). Table 3 presents the full multivariable model for DOT; predictors were similar for DDDs and LOT with the only differences being that an admission at the weekend was additionally associated with greater DDDs and Charlson score was not associated with DDDs; and the only factors associated with greater LOT were male sex, sepsis and higher CRP on admission. Predictors for giving a broad spectrum antibiotic and being given an intravenous antibiotic were immunosuppression, higher CRP and WCC on admission, with predictors of higher broad spectrum DOT being fever, higher CRP, female sex, age, deprivation index and admission at night, and predictors of greater intravenous DOT were sepsis, higher CRP, deprivation index and female sex.
‡Broad-spectrum antibiotics: co-amoxiclav, ceftriaxone, meropenem, gentamicin, clindamycin, ceftazidime, ciprofloxacin, piperacillin/tazobactram. CRP, C reactive protein; DDD, defined daily dose; DOT, days of therapy; IDP, Infectious Diseases Physician; LOT, length of therapy; RR, rate ratio; WCC, white cell counts.