The healthcare costs of antimicrobial resistance in Lebanon: a multi-centre prospective cohort study from the payer perspective

Table 3 Regression analyses of excess hospital charges and excess LOS attributable to antibiotic-resistant bacteria

	Unstandardized beta	95% Confidence Interval		R²	p-value
		Upper	Lower
Community-associated infection – Conventional method
Excess LOS post-infection, days	2.2	1.2	3.3	0.027	< 0.001
Excess hospital charges, $USD	889	378	1400	0.046	0.001
Healthcare-associated infection – “Post-infection HAI” method
Excess LOS post-infection, days	2.1	0.9	3.2	0.140	< 0.001
Excess hospital charges, $USD	1297	627	1966	0.274	< 0.001
Healthcare-associated infection – Matched method
Excess LOS post-infection, days	1.6	0.3	2.9	0.135	0.016
Excess hospital charges, $USD	1139	471	1807	0.229	0.001

Linear regression models were used assuming linear distribution. Unstandardized beta coefficient, excess LOS post-infection and excess hospital charges displayed in the table are those for the key independent variable: infection due at least one resistant bacteria. The other independent variables included in the community-associated infection regression model were patient age, gender, Charlson co-morbidity index and previous antibiotic exposure. In the healthcare-associated infection analyses, “post HAIs” analysis the other independent variables included in the regression models were patient age, gender, Charlson co-morbidity index, previous exposure to antibiotic, transfer to the ICU, insertion of central venous catheter, insertion of urinary catheter, mechanical ventilation, days of mechanical ventilation, time to infection, appropriateness of antibiotic therapy. p value < 0.05 is considered significant
Abbreviations: LOS, Length of stay; Coefficient, unstandardized Beta; CI, confidence interval; HAIs, healthcare –associated infections; ICU, intensive care unit

ISSN: 1471-2334