Costs of hospital-acquired Clostridium difficile infections: an analysis on the effect of time-dependent exposures using routine and surveillance data

Table 3 Sensitivity analysis: Cox regressions

	(1)	(2)
	Time-varying CDI exposure, no time-to-exposure stratification	Time-fixed CDI exposure, time-to-exposure stratification
Hazard ratio	0.669*** (0.0366)	0.741*** (0.0348)
N	52,679	23,052

Each cell gives the effect of CDI infection on the time to discharge (alive or due to death) from different Cox proportional hazard regressions. A hazard ratio < 1 indicates an increased length of stay due to CDI exposure. Due to convergence issues in the regression model, the amount of controls per case is not restricted to four in this case. All regressions use a within-main-diagnosis estimation by stratifying the regression and control for comorbidities according to Resch (2008), for age, age², and for sex. Standard errors in parentheses
* p < 0.1, ** p < 0.05, *** p < 0.01

ISSN: 1478-7547