Study design and setting
A retrospective study was performed at Hospital del Mar, a 420-bed, urban, tertiary-care teaching hospital that covers an area of 300,000 inhabitants in the city of Barcelona, Catalonia (Spain). The study was carried out in compliance with the Helsinki Declaration. The study was approved by the ethics committee of IMIM-Hospital del Mar.
The study population consisted of all hospital admissions between January 1, 2005 and December 31, 2006 in which a nosocomial incident P. aeruginosa was isolated and identified. We used the microbiology laboratory records of the hospital to identify all inpatients with positive clinical cultures for P. aeruginosa. Clinical specimens were isolated and identified by the microbiology laboratory by means of routine techniques. Nosocomial incident P. aeruginosa acquisitions were defined as those in which the first P. aeruginosa positive culture for a particular patient occurred more than 48 h after patient admission during the study period. Overall, 410 incident nosocomial P. aeruginosa positive cultures were identified. The present report is based on 402 admissions with complete information on antibiotic susceptibility pattern and economical cost estimations (98% of the 410 positive cultures).
The susceptibility of isolates was determined by two methods: the MicroScan Walk away (Siemens Healthcare) (using NC36 and NC38 panels) or the Kirby Bauer method in Muller Hinton plates (Biomerieux Marcy l’etoile). Antibiotic susceptibility tests were performed using a standardized custom microtiter minimum inhibitory concentration (MIC). Testing procedures were validated by determining the MICs for reference strains. P. aeruginosa isolates were classified in three categories according to the antibiotic susceptibility pattern to all studied agents as: 1) non-resistant P. aeruginosa when the organism was susceptible to all the agents studied; 2) multi-drug resistant (MDRPA) for strains resistant to carbapenems, b-lactams, quinolones, tobramycin, and gentamicin and sensitive to colistin and amikacin [5, 8, 14]; and 3) resistant P. aeruginosa all the possible remainder combinations.
This information was linked to a hospital computerized patient records database. The following information was retrieved for each admission during the study period: patient age and sex, cause of hospitalization, medical treatments and comorbidities, previous hospitalization in the same hospital, previous intensive care unit (ICU) stay during the same hospital admission, exposure to invasive devices or procedures (mechanical ventilation, hemodialysis, bronchoscopy, digestive endoscopy and surgery); length of hospital stay (LOS) (including the LOS before detection defined as the period of time from admission to microorganism detection; and total LOS defined as the period of time from admission to the discharge or death of the patient), previous antibiotic therapy during the current hospitalization episode, and in-hospital mortality. Moreover, underlying illness severity was measured by the All Patient Refined Diagnosis Related Group Severity of Illness scale (version 20.0) [15].
Costs estimation
The Municipal Institute of Health uses a hospital cost accounting system based on full-costing allocation that allows for assessing direct cost derived from clinical activity [16]. In the present study, cost estimation was based on a full-costing cost accounting system and on the criteria of clinical Activity-Based Costing (ABC) methods to obtain the highest sensitivity in the assessment of variability in clinical activity. Moreover, this system ensures that the hospital’s total costs are distributed among the patients. Allocation was based on directly assigning the cost of the following services to the patient: laboratory, pharmacy, radiology, nuclear medicine, pathology, and prosthetics [17, 18]. The information systems contain exhaustive data on human resources and their activity: storage, admissions planning, ambulatory and emergency care, operating rooms, diagnostic and complementary tests, and inter-hospital consultations. This information creates and automatically updates the cost drivers for overheads [19, 20].
The principal economic outcome assessed was total hospital cost of nosocomial MDRPA acquisition. To avoid the overestimation effect of costs due to extra LOS, all hospital costs were estimated once a positive culture was detected. This includes fixed cost, variable cost and pharmacy cost. Fixed cost derived from surgical procedures, hospitalization, and ICU stay, was distributed based on routine criteria: operation or intervention time, or the number of days of stay among the various hospitalization units. Variable cost included the costs derived from laboratory, radiology, pathology, prostheses, tests and pharmacy. In addition, we evaluated the effect of nosocomial MDRPA acquisition on pharmacy cost.
Statistical analysis
Patient admission represented the unit of analysis. Univariate statistics were computed as customary. In contingency tables, Fisher's exact test for homogeneity or independence was applied to assess the relationship between two categorical variables; the Kruskal-Wallis and Mann–Whitney U-tests were used to compare continuous variables. The main effects of all predictors were independently explored in base models. Hospital costs were log transformed to achieve a normal distribution as far as possible. Multivariate analyses of hospital cost were performed using generalized linear models with the Gamma distribution and the log link function. The coefficients were converted to the measures of effect using an exponential transformation. Final multivariate models were adjusted for patient age and sex, comorbidities, average total cost per Diagnosis-Related-Groups (DRG) of patients with the lowest severity score, source of infection, stay (or not) in UCI after the infection and in-hospital mortality. The level of statistical significance was set at 0.05, and all tests were two tailed. Analyses were conducted using PASW Statistics 18 version 18.0.0 (IBM corp., 2009).
