Empirical prescribing of penicillin G/V reduces risk of readmission of hospitalized patients with community-acquired pneumonia in Norway: a retrospective observational study

Table 3 Multivariable regression analysis. Association between empirical antibiotic prescribing, various covariates and length of hospital stay (LOS) (n = 626) and 30-day readmission (n = 609), respectively

Variable	Length of hospital stay^a		30-day readmission^b
Variable	Adjusted coefficient	[95% CI]	Adjusted odds ratio	[95% CI]
Empirical antibiotic prescribing
Penicillin G/V in monotherapy	0	Ref.	1	Ref.
Penicillin G + gentamicin	0.68	[−0.10, 1.46]	1.39	[0.73, 2.66]
Other antibiotics	0.33	[−0.43, 1.09]	1.92	[1.08, 3.42]
Antibiotic use pre-hospitalization	−0.25	[− 0.95, 0.46]	1.15	[0.66, 1.99]
Female	0.08	[−0.53, 0.69]	0.50	[0.30, 0.82]
Age	0.04	[0.02, 0.07]	1.01	[0.99, 1.03]
Comorbidities
COPD	0.22	[−0.46, 0.86]	2.07	[1.26, 3.41]
Heart failure	1.40	[0.62, 2.17]	1.45	[0.83, 2.51]
Diabetes Mellitus I or II	0.38	[−0.52, 1.27]	1.22	[0.63, 2.37]
Nursing home resident	−0.98	[−2.05,2.17]	1.40	[0.64, 3.07]
Year admitted (2012 vs. 2010)	−0.42	[−0.73,-0.11]	0.85	[0.66, 1.08]
Hospital
Hospital A	0	Ref	1	Ref.
Hospital B	0.81	[−0,14, 1.76]	0.74	[0.37, 1.50]
Hospital C	1.10	[0.34, 1.87]	0.51	[0.29, 0.88]
Pathogens
None identified	0	Ref.	1	Ref.
S.pneumoniae	1.22	[0.10, 2.34]	0.96	[0.37, 2.47]
H.influenzae	2.13	[− 041, 4.67]	3.01	[0.65, 14.02]
M. or C. pneumoniae	0.75	[−0.98, 2.48]	0.39	[0.05, 3.22]
S.aureus	4.10	[1.24, 6.96]	5.24	[0.99, 27.6]
Other bacteria	5.19	[2.98, 7.40]	1.38	[0.26, 7.38]
Influenza virus A or B	1.86	[−0.83, 4.55]	–	–
Other respiratory viruses	0.44	[−1.56, 2.44]	0.85	[0.10, 6.98]
Two or more pathogens	1.40	[−1.30, 4.10]	–	–
CRB-65 score
0	0	Ref.	1	Ref.
1	−0.19	[− 1.24, 0.85]	0.82	[0.32, 2.06]
2	0.50	[−0.65, 1.64]	0.99	[0.37, 2.61]
3	1.92	[0.35, 3.50]	0.71	[0.19, 2.66]
4	1.63	[−2.85, 6.11]	–	–
Missing data	−0.27	[−1.55, 1.01]	0.78	[0.26, 2.39]

CI confidence interval, COPD Chronic obstructive pulmonary disease, CRB-65 Confusion, Respiration, Blood pressure and Age
^a Linear regression to explore impact of empirical antibiotic prescribing on length of hospital stay. Adjusted for antibiotic use pre-hospitalization, gender, age, comorbidities, nursing home resident, year admitted (2012 vs. 2010), hospital, pathogens and severity of infection (CRB-65)
^b Logistic regression to explore impact of empirical antibiotic prescribing on 30-day readmission. Adjusted for antibiotic use pre-hospitalization, gender, age, comorbidities, nursing home resident, year admitted (2012 vs. 2010), hospital, pathogens and severity of infection (CRB-65)

ISSN: 1471-2466