Table 3.

Bivariate and multivariable negative binomial regression model of variables associated with receipt of at least 1 inappropriate antimicrobial dose among patients receiving antimicrobials in 2 chronic outpatient hemodialysis units

	Bivariate analysis			Multivariable analysis
Variable	IRR	95% CI	P value	IRR	95% CI	P value
Age, each additional year	0.99	0.97–1.02	.58
Male sex	0.57	0.27–1.21	.14
Race-ethnicity
White	ref	—	—
Black	1.44	0.65–3.18	.37
Other	1.26	0.56–2.86	.58
Primary indication for hemodialysis, diabetes mellitus (vs other)	0.80	0.39–1.63	.54
Tunneled catheter for access (vs arteriovenous access)	1.72	0.87–3.40	.12	2.23	1.16–4.29	.02
Unit B	2.30	1.97–2.63	.03	2.67	1.34–5.35	.005
Daytime shift (vs evening shift)	2.05	0.49–8.49	.32
Duration of hemodialysis at time of enrollment, each additional year	1.08	0.98–1.20	.13	1.11	1.01–1.23	.03
Charlson Comorbidity Index score >4	1.17	0.58–2.36	.66
Requires wheelchair or bedbound	0.99	0.45–2.19	.98
Chronic wound present on enrollment	0.48	0.08–2.79	.42
Diabetes mellitus, type 1 or 2	1.42	0.73–2.79	.30
Immunosuppressive medication at time of enrollment*	0.72	0.26–1.99	.53
History of prior organ transplant	2.85	0.96–8.43	.06
Malignancy	1.92	0.62–5.98	.26
History of ≥1 multidrug-resistant organism in 12 mo preceding enrollment	0.78	0.28–2.18	.64
Hospitalizations in the 12 mo preceding enrollment
No hospitalizations	ref	—	—
1–2 hospitalizations	1.04	0.44–2.47	.92
≥3 hospitalizations	0.87	0.36–2.07	.75

CI, confidence interval; IRR, incident rate ratio; ref, reference.

^*Immunosuppressive medications include corticosteroids (prednisone dose equivalent of ≥20 mg/d), immunomodulators, and cancer chemotherapeutic treatment.