TABLE 5

Negative binomial mixed effects model estimates of the duration of intravenous (i.v.) antibiotic treatment (n=364, years of observation=2805)

	Unadjusted coefficient	IRR	95% CI (IRR)	p-value	Adjusted coefficient^#	IRR	95% CI (IRR)	p-value
Intercept	1.76	5.83	3.97–8.56	<0.001	1.48	4.38	2.82–6.79	<0.001
Time	0.15	1.16	1.07–1.26	<0.001*	0.15	1.16	1.07–1.26	<0.001*
CFTR modulator	−1.28	0.28	0.19–0.40	<0.001*	−1.28	0.28	0.19–0.40	<0.001*
Time : CFTR modulator	0.16	1.18	0.93–1.49	0.170	0.17	1.19	0.94–1.50	0.153

Interpretation: coefficients are on the log-scale. Incidence rate ratios (IRRs) are transformed back to the original scale. The IRR of the intercept represents the mean duration of received i.v. antibiotics (in days) of the study population at the time of CFTR modulator initiation (baseline). The IRR of time shows the relative annual change in the duration of i.v. antibiotics before CFTR modulator treatment. The IRR of CFTR modulator reflects the acute change in the duration of i.v. antibiotics in the first year after CFTR modulator initiation, whereas time : CFTR modulator treatment indicates the relative change of i.v. antibiotic treatment in the years after modulator initiation compared to the annual trend before CFTR modulator use. CFTR: cystic fibrosis transmembrane conductance regulator. ^#: coefficients and IRRs were adjusted for sex and age at baseline; *: significance level p<0.05.