Factors associated with daptomycin consumption in French hospitals between 2019 and 2020: A nationwide surveillance study.

OBJECTIVES: This study aimed to investigate factors associated with daptomycin consumption in French healthcare facilities (HCF) between 2019 and 2020. METHODS: Antibiotic consumption expressed as number of defined daily doses (DDD) per 1,000 patient-days (PD) and antimicrobial resistance (AMR) expressed as incidence densities per 1,000PD were extracted each year from the nationwide surveillance network run by the SPARES project (Surveillance and Prevention of Antimicrobial RESistance in hospitals), collecting data at ward level among voluntary HCFs using standardized methodology and webtool. All HCF participating both in 2019 and 2020 were included. A multivariable linear regression was fitted. RESULTS: Among 622 HCFs, we analyzed daptomycin consumption and AMR data in 1,637 clinical wards. Incidence densities of methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-resistant coagulase-negative staphylococci (MRCNS) were the highest in intensive care unit wards (0.54 and 6.83 respectively in 2020). On the most adjusted model, the year 2020 was correlated with a higher daptomycin consumption (1.53; p = 0.01). A greater number of inpatient beds (0.01; p < 0.001), the presence of orthopedic surgery activity in the HCF (1.66; p < 0.02), MRSA (4.38; p < 0.001) and MRCNS (0.61; p < 0.001) incidence densities were associated with a higher daptomycin use. The final model explained 18% of the observed variance. CONCLUSIONS: This study showed that daptomycin consumption was correlated to MRSA and MRCNS incidence densities, to the year 2020 and to non-modifiable HCF-related factors. Prevention of coagulase-negative staphylococci infections should be considered by antimicrobial stewardship teams when daptomycin use is going up in HCF.

Benefits of a 14-year surgical site infections active surveillance programme in a French teaching hospital.

BACKGROUND: Surgical site infections (SSIs) are the second most common healthcare-associated infection. Active SSI surveillance can help inform preventative measures and assess the impact of these measures. AIM: We aimed to describe the evolution in trends over 14 years of prospective active SSI surveillance and implementations of SSI prevention measures in a French Teaching Hospital. METHODS: We monitored and included in the study all surgical procedures performed from 2003 to 2016 in eight surgical units. The semi-automated surveillance method consisted of weekly collection of SSI declaration forms (pre-filled with patient and procedure administrative data and microbiology laboratory data), filled-in by surgeons and then monitored by the infection control practitioners. FINDINGS: A total of 181,746 procedures were included in our analysis and 3270 SSIs recorded (global SSI rate 1.8%). The SSI rate decreased significantly from 3.0% in 2003 to 1.1% in 2016. This decrease was mainly in superficial SSIs and high infectious risk procedures. Higher SSI rates were observed for procedures associated with the usual risk factors. During this 14-year period, several evolutions in surgical practices occurred that might have contributed to this decrease. CONCLUSIONS: With an overall decrease in SSI rate throughout the surveillance, our results revealed the benefits of an active and comprehensive hospital SSI surveillance programme for understanding the SSI rate trends, analysing local risk factors and assessing the effectiveness of prevention strategies. These findings also highlighted the importance of the collaboration between surgeons and infection control practitioners.

Exogenous acquisition of Pseudomonas aeruginosa in intensive care units: a prospective multi-centre study (DYNAPYO study).

BACKGROUND: Pseudomonas aeruginosa remains one of the most common nosocomial pathogens in intensive care units (ICUs). Although exogenous acquisition has been widely documented in outbreaks, its importance is unclear in non-epidemic situations. AIM: To elucidate the role of exogenous origin of P. aeruginosa in ICU patients. METHODS: A chronological analysis of the acquisition of P. aeruginosa was performed using samples collected in 2009 in the DYNAPYO cohort study, during which patients and tap water were screened weekly. Molecular relatedness of P. aeruginosa isolates was investigated by pulsed-field gel electrophoresis. Exogenous acquisition was defined as identification of a P. aeruginosa pulsotype previously isolated from another patient or tap water in the ICU. FINDINGS: The DYNAPYO cohort included 1808 patients (10,402 samples) and 233 water taps (4946 samples). Typing of 1515 isolates from 373 patients and 375 isolates from 81 tap water samples identified 296 pulsotypes. Analysis showed exogenous acquisition in 170 (45.6%) of 373 patients. The pulsotype identified had previously been isolated from another patient and from a tap water sample for 86 and 29 patients, respectively. The results differed according to the ICU. CONCLUSION: Exogenous acquisition of P. aeruginosa could be prevented in half of patients. The overall findings of this survey support the need for studies on routes of transmission and risk assessment approach to better define how to control exogenous acquisition in ICUs.

Risk factors for colonization and infection by Pseudomonas aeruginosa in patients hospitalized in intensive care units in France.

OBJECTIVE: To assess the role of environment, medical care and individual risks factors for P. aeruginosa colonization and infection. STUDY DESIGN AND SETTING: A French multicentric prospective study involved ten ICUs for a five months period. Every adult patient newly hospitalized in ICUs with no P. aeruginosa carriage up to 48 hours after admission was included and weekly screened before discharge or death. Screening swabs were either rectal, sputum or oropharyngeal samples. Hydric environment was also sampled each week. Data on patient clinical features, environmental and device exposures, and antibiotics supports were regularly collected. Multivariate analysis was performed with a multistate model. RESULTS: The overall prevalence of P. aeruginosa carriage was 15.3% (201/1314). Risk factors associated with patient colonization were: use of inactive antibiotics against P. aeruginosa (HR = 1.60 [1.15-2.21] p<0.01), tap water contamination at the entry in the room (HR = 1.66 [1.01-2.27] p<0.05) and mechanical invasive ventilation (HR = 4.70 [2.66-8.31] p<0.0001). Active antibiotics prevented from colonization (HR = 0.67 [0.48-0.93] p = 0.02) and from infection (HR = 0.64 [0.41-1.01] p = 0.05). Interaction between hydric environment antibiotics support was not statistically associated with patient colonization. CONCLUSION: Hydric contamination and antibiotics pressure seem to remain key independent risk factors in P. aeruginosa colonization. These results advocate the need to carry on preventive and targeted interventions toward healthcare associated infections.

Relationship between rates of antimicrobial consumption and the incidence of antimicrobial resistance in Staphylococcus aureus and Pseudomonas aeruginosa isolates from 47 French hospitals.

OBJECTIVE: To investigate relationships between rates of antimicrobial consumption and the incidence of antimicrobial resistance in Staphylococcus aureus and Pseudomonas aeruginosa isolates from hospitals. METHODS: We conducted an observational study that used retrospective data from 2002 and linear regression to model relationships. Hospitals were asked to collect data on consecutive S. aureus and P. aeruginosa isolates, consumption rates for antibiotics (ie, anti-infectives for systemic use as defined by Anatomical Therapeutic Chemical class J01), and hospital characteristics, including infection control policies. Rates of methicillin resistance in S. aureus and rates of ceftazidime and ciprofloxacin resistance in P. aeruginosa were expressed as the percentage of isolates that were nonsusceptible (ie, either resistant or intermediately susceptible) and as the incidence of nonsuceptible isolates (ie, the number of nonsuceptible isolates recovered per 1,000 patient-days). The rate of antimicrobial consumption was expressed as the number of defined daily doses per 1,000 patient-days. SETTING: Data were obtained from 47 French hospitals, and a total of 12,188 S. aureus isolates and 6,370 P. aeruginosa isolates were tested. RESULTS: In the multivariate analysis, fewer antimicrobials showed a significant association between the consumption rate and the percentage of isolates that were resistant than an association between the consumption rate and the incidence of resistance. The overall rate of antibiotic consumption, not including the antibiotics used to treat methicillin-resistant S. aureus infection, explained 13% of the variance between hospitals in the incidence of methicillin resistance among S. aureus isolates. The incidence of methicillin resistance in S. aureus isolates increased with the use of ciprofloxacin and levofloxacin and with the percentage of the hospital's beds located in intensive care units (adjusted multivariate coefficient of determination [aR(2)], 0.30). For P. aeruginosa, the incidence of ceftazidime resistance was greater in hospitals with higher consumption rates for ceftazidime, levofloxacin, and gentamicin (aR(2), 0.37). The incidence of ciprofloxacin resistance increased with the use of fluoroquinolones and with the percentage of a hospital's beds located in intensive care ( aR(2), 0.28). CONCLUSIONS: A statistically significant relationship existed between the rate of fluoroquinolone use and the rate of antimicrobial resistance among S. aureus and P. aeruginosa isolates. The incidence of resistant isolates showed a stronger association with the rate of antimicrobial use than did the percentage of isolates with resistance.

Surveillance of occupational blood and body fluid exposures among French healthcare workers in 2004.

OBJECTIVE: To estimate the incidence rate of reported occupational blood and body fluid exposures among French healthcare workers (HCWs). DESIGN: Prospective national follow-up of HCWs from January 1 to December 31, 2004. SETTING: University hospitals, hospitals, clinics, local medical centers, and specialized psychiatric centers were included in the study on a voluntary basis. PARTICIPANTS: At participating medical centers, every reported blood and body fluid exposure was documented by the occupational practitioner in charge of the exposed HCW by use of an anonymous, standardized questionnaire. RESULTS: A total of 375 medical centers (15% of French medical centers, accounting for 29% of hospital beds) reported 13,041 blood and body fluid exposures; of these, 9,396 (72.0%) were needlestick injuries. Blood and body fluid exposures were avoidable in 39.1% of cases (5,091 of 13,020), and 52.2% of percutaneous injuries (4,986 of 9,552) were avoidable (5.9% due to needle recapping). Of 10,656 percutaneous injuries, 22.6% occurred during an injection, 17.9% during blood sampling, and 16.6% during surgery. Of 2,065 splashes, 22.6% occurred during nursing activities, 19.1% during surgery, 14.1% during placement or removal of an intravenous line, and 12.0% during manipulation of a tracheotomy tube. The incidence rates of exposures were 8.9 per 100 hospital beds (95% confidence interval [CI], 8.7-9.0 exposures), 2.2 per 100 full-time-equivalent physicians (95% CI, 2.4-2.6 exposures), and 7.0 per 100 full-time-equivalent nurses (95% CI, 6.8-7.2 exposures). Human immunodeficiency virus serological status was unknown for 2,789 (21.4%) of 13,041 patients who were the source of the blood and body fluid exposures. CONCLUSION: National surveillance networks for blood and body fluid exposures help to better document their characteristics and risk factors and can enhance prevention at participating medical centers.

Determinants of glycopeptides consumption in hospitals.

The aim of this study was to describe consumption of glycopeptides and to study factors associated with their use in 47 French hospitals. Consumption of glycopeptides for systemic use (defined daily doses per 1,000 patient-days: DDD/1,000 PD and per 100 admissions), number of methicillin-resistant Staphylococcus aureus (MRSA) (percentage and incidence per 1,000 patient-days), and number of venous central lines and hospital characteristics (size, length of stay, number of beds: total and for each hospital inpatient areas and antibiotic policies) were recorded from January, 2002, through December, 2002. Multiple linear regression was performed to check for hospital characteristics. The median rate of total consumption of glycopeptides was 4.11(range 0.21-27.22) DDD per 1,000 PD with higher consumption in large public hospitals and in intensive care areas (median 46.51; range 7.19-134) than in surgery areas (median 4.5; range 0.17-24.76). The consumption of glycopeptides correlated with MRSA incidence, but not with the proportion of MRSA. In the multivariate analysis, the incidence of MRSA and the number of beds in surgery areas were independent predictors of total glycopeptides use in the hospital, expressed in DDD per 1,000 PD (R2 adjusted, 0.39). The incidence of MRSA, the number of venous central lines, and the number of beds in the medicine areas were significant determinants associated with higher consumption of glycopeptides expressed in DDD per 100 admissions (R2 adjusted, 0.73). To reduce glycopeptides use in hospitals, the first effort required is that hospitals focus increased attention on the prevention of cross transmission for MRSA between patients but also on the use of the venous central line. Furthermore, hospitals have to compare their data with others to identify overuse of glycopeptides and to plan control interventions.

Patient and bacterial determinants involved in symptomatic urinary tract infection caused by Escherichia coli with and without bacteraemia.

Risk-factors for bacteraemia were determined in a case-control study of patients with Escherichia coli urinary tract infection. Cases were defined as patients with E. coli urinary source bacteraemia, and controls were chosen from among patients with E. coli urinary tract infection without bacteraemia. Patient characteristics were collected prospectively and the bacterial traits were determined. The phylogenetic background and virulence factors of E. coli isolates did not differ between cases and controls. In multivariate analysis, being female and having a urinary catheter were significantly less prevalent among patients with urinary source bacteraemia than among patients with uncomplicated urinary tract infection.

Control of the duration of urinary catheterization: impact on catheter-associated urinary tract infection.

Catheter-associated urinary tract infection (CAUTI) is one of the most common types of nosocomial infection. Reducing the duration of catheterization may be a key intervention in CAUTI prevention. We carried out a prospective time-sequence, non-randomized intervention study in five departments of our university hospital. We wished to determine the impact of daily reminders from nurses to physicians to remove unnecessary urinary catheters four days after insertion. The intervention significantly decreased the duration of catheterization in two out of five departments. The frequency of late CAUTI (LCAUTI) among catheterized patients in all five departments decreased from 10.6 to 1.1 per 100 patients (P = 0.003) and the incidence of LCAUTI decreased from 12.3 to 1.8 per 1000 catheter-days (P = 0.03). Logistic regression analysis showed that duration of catheterization and iterative catheter changes were associated with LCAUTI. This study demonstrates that a simple measure can reduce the frequency of LCAUTI.

Outbreak of staphylococcal bullous impetigo in a maternity ward linked to an asymptomatic healthcare worker.

An outbreak of staphylococcal bullous impetigo occurred over a period of five months in a maternity ward involving seven infected and two colonised neonates. The skin lesions were due to epidermolytic toxin A-producing Staphylococcus aureus. Infection control measures were implemented and a retrospective case-control study performed. Contact with an auxiliary nurse was the only risk factor for cases of bullous impetigo (P<0.01). The nurse cared for all seven cases and was an asymptomatic nasal carrier of the epidemic strain. Repeated courses of decontamination treatment failed to eradicate carriage. Nine months after the last case, another neonate developed a more severe form of bullous impetigo and the auxiliary nurse was reassigned to an adult ward.

[Relationship between antibiotic policies and antibiotic consumption in hospitals]. / Relation entre politique de bon usage et consommation des antibiotiques dans les établissements de santé.

OBJECTIVE: The authors had for aim to study the relationship between antibiotic policies and antibiotic consumption in hospitals. DESIGN: A component analysis was used to summarize the various measures of the antibiotic policies developed by hospitals. Antibiotics consumption was expressed as a number of Daily Defined Doses per 1,000 patient-days. The relation was studied by a multiple linear regression model with adjustment on hospital activity. RESULTS: The first component illustrated an active program of antibiotic policies associating: a local antibiotic committee, an infectious diseases consultant, written local guidelines for prescriptions and restrictive measures before dispensing. The highest antibiotic consumption was associated with active policies, especially for carbapenems, glycopeptides, and aminoglycosides. However the relationship differed according to the type of hospital, with lower antibiotic consumption, especially for penicillins, in private hospitals that had more active policies. CONCLUSION: The difference between public and private hospitals could be explained by the type of institution but also by the unit used to measure antibiotic consumption.

Root cause analysis to support infection control in healthcare premises.

Infection control teams (ICTs) seek to prevent healthcare-associated infections (HCAIs). They undertake surveillance and prevention, promote safety and quality of care, and evaluate and manage risk. Root cause analysis (RCA) can support this work but is not widely used by ICTs. This paper describes how ICTs can use RCA to enhance their day-to-day work. Many different tools and methods exist for RCA. Its primary aim is to identify the factors that have led to HCAI, but RCA can also be used for near-misses. A team effort and multidisciplinary work are usually required. Published accounts and personal experience in the field indicate that an ICT that correctly uses RCA implements more effective prevention measures, improves practice and collaborative working, enhances teamwork, and reduces the risk of HCAI. RCA should be promoted among ICTs because it adds value to their work and helps to develop a hospital culture that anticipates and pre-empts problems.

Surgical site infection after valvular or coronary artery bypass surgery: 2008-2011 French SSI national ISO-RAISIN surveillance.

BACKGROUND: Multisite information regarding surgical site infection (SSI) rates for cardiac surgery programmes is not widely available. Ward characteristics that may affect outcomes have not been analysed previously. AIM: To determine individual- and ward-level factors associated with SSI occurrence after coronary artery bypass grafting (CABG) and valvular surgery. METHODS: A dataset from the French national SSI database ISO-RAISIN 2008-2011 was used. Only adult patients were included. A standardized questionnaire was completed for each patient who underwent surgery, and patients with and without SSI were characterized. Patients and ward risk factors for SSI were analysed using a multilevel logistic regression model with SSI as binary outcome (two levels: patient and ward). RESULTS: Out of 8569 patients from 39 wards, the SSI rate was 2.2%. Micro-organisms were isolated in 144 patients (74%): 35% coagulase-negative staphylococci (N = 51), 23% Staphylococcus aureus (N = 33), 6% Escherichia coli (N = 8). Higher probability of SSI was associated with the duration of preoperative hospitalization, the duration of follow-up, the duration of surgery >75th percentile and the SSI rate in the surgery ward. The residual heterogeneity between wards (median odds ratio: 1.53) was as relevant as duration of preoperative hospitalization (odds ratio: 1.57). CONCLUSION: Although patient risk factors were more strongly associated with SSI occurrence, this study provided evidence for the existence of a ward-level effect. This should be taken into account when considering possible corrective interventions.

[Effects of imiquimod on latent human papillomavirus anal infection in HIV-infected patients]. / Effets de l'imiquimod sur l'infection périanale latente à papillomavirus humain chez des malades infectés par le virus de l'immunodéficience humaine.

INTRODUCTION: High risk human papillomaviruses (HPV) have emerged as risk factors for anal carcinoma particularly in HIV-infected patients who demonstrate a high rate of anal HPV infection. Considering the relationship between the presence of anal infection and the development of neoplastic lesions, we wished to assess the capacity of imiquimod to eradicate latent HPV infection in HIV-infected patients. PATIENTS AND METHODS: We conducted a prospective, randomized, double-blind and vehicle controlled study. Two consecutive anal swabs were taken at 2 month intervals and only patients with two consecutive tests positive for the detection of HPV-DNA (Hybrid Capture II) were included. Patients with persistent latent HPV infection were divided into 2 groups who applied imiquimod versus vehicle during 6 weeks. HPV-DNA presence was then investigated 2 and 4 months following the onset of treatment. RESULTS: Among the 80 HIV-infected patients, 26 (32.5 p. 100) had 2 positive consecutive assays, and 19 patients were included in the study. After randomization, 9 patients received imiquimod and 10 vehicle. There was no significant difference between treatment groups according to the following criteria: gender, route of HIV transmission, CDC stage, prior medical history of sexually transmitted diseases or anogenital warts. 33.3 p. 100 (3/9) of patients treated with imiquimod were negative at M2, whereas 100 p. 100 (10/10) vehicle treated-patients remained positive (p=0.08). Similar results were observed at the M4 visit. DISCUSSION: Our study confirmed the increased prevalence of latent HPV-DNA infection in HIV-infected patients. In spite of the low number of treated patients, we did not observe a statistically significant decrease in HPV-DNA in anal swabs from imiquimod-treated patients as compared to placebo-treated patients.

Risk factors for Pseudomonas aeruginosa acquisition in intensive care units: a prospective multicentre study.

BACKGROUND: Pseudomonas aeruginosa is a major nosocomial pathogen in intensive care units (ICUs); however, endogenous versus exogenous origin of contamination remains unclear. AIM: To identify individual and environmental ICU risk factors for P. aeruginosa acquisition. METHODS: A five-month prospective multicentric study was performed in ten French ICUs. Adult patients hospitalized in ICU for ≥ 24 h were included and screened for P. aeruginosa colonization on admission, weekly and before discharge. P. aeruginosa acquisition was defined by a subsequent colonization or infection if screening swabs on admission were negative. Water samples were obtained weekly on water taps of the ICUs. Data on patient characteristics, invasive devices exposure, antimicrobial therapy, P. aeruginosa water and patient colonization pressures, and ICU characteristics were collected. Hazard ratios (HRs) were estimated using multivariate Cox model. FINDINGS: Among the 1314 patients without P. aeruginosa on admission, 201 (15%) acquired P. aeruginosa during their ICU stay. Individual characteristics significantly associated with P. aeruginosa acquisition were history of previous P. aeruginosa infection or colonization, cumulative duration of mechanical ventilation and cumulative days of antibiotics not active against P. aeruginosa. Environmental risk factors for P. aeruginosa acquisition were cumulative daily ward 'nine equivalents of nursing manpower use score' (NEMS) [hazard ratio (HR): 1.47 for ≥ 30 points; 95% confidence interval (CI): 1.06-2.03] and contaminated tap water in patient's room (HR: 1.76; CI: 1.09-2.84). CONCLUSION: Individual risk factors and environmental factors for which intervention is possible were identified for P. aeruginosa acquisition.

Nosocomial urinary tract infection in the intensive care unit: when should Pseudomonas aeruginosa be suspected? Experience of the French national surveillance of nosocomial infections in the intensive care unit, Rea-Raisin.

Individual and ward risk factors for P. aeruginosa-induced urinary tract infection in the case of nosocomial urinary tract infection in the intensive care unit were determined with hierarchical (multilevel) logistic regression. The 2004-2006 prospective French national intensive care unit nosocomial infection surveillance dataset was used and 3252 patients with urinary tract infection were included; 16% were infected by P. aeruginosa. Individual risk factors were male sex, duration of stay, antibiotics at admission and transfer from another intensive care unit. Ward risk factors were patient turnover and incidence of P. aeruginosa-infected patients.

A multicentre analysis of catheter-related infection based on a hierarchical model.

PURPOSE: To decrease intensive care unit (ICU)-acquired catheter-related infections (CRI), it is essential that healthcare workers receive training and that quality improvement programmes are in place. The aim of our study was to evaluate risk factors for catheter colonisation and infection, focussing specifically on local care bundles. METHODS: Data were collected prospectively in 51 ICUs [7,188 patients, 8,626 central venous catheters (CVCs)] during two 6-month periods in 2007 and 2008, using a standardized questionnaire on catheter insertion, care and removal. Colonisation and CRI incidence were 6.1 and 2.2/1,000 CVC-days, respectively. A hierarchical mixed logistic model was used to identify risk factors for CRI and colonisation. RESULTS: Written CVC protocols were available in 46 (90 %) ICUs and were strictly followed in 38 ICUs. Factors significantly associated with CRI fell into three overall categories: (1) patient-related factors-immunosuppression [odds ratio (OR) 1.42, p = 0.02], medical diagnosis at admission (OR 1.64, p = 0.03) and trauma patient (OR 2.54, p < 0.001); (2) catheter-related factors-catheter rank (OR 1.7, p < 0.0001, non-subclavian catheter (OR 2.1, p < 0.001) and longer time with the catheter (p < 10(-4)); (3) centre-related factors-quantitative tip culture method (OR 2.55, p = 0.005) and alcohol-based povidone-iodine [OR 0.68, 95 % confidence interval (CI) 0.49-0.96] or alcohol-based chlorhexidine preparations (OR 0.69, 95 % CI 0.34-1.39) as compared to an aqueous povidone-iodine preparation (p < 0.001). CONCLUSIONS: We identified several risk factors for CRI that are amenable to improvement (preference for the subclavian route and use of an antiseptic solution containing alcohol). However, several patient-related factors were also found, and the use of quantitative catheter culture methods increased culture sensitivity, thereby increasing the CRI rate. Case-mix issues and the culture method should be taken into account when assessing the risk of CRI across centres.

Factors associated with the prevalence of antibiotic use for the treatment of hospital-acquired infections at 393 French hospitals: a regional variation analysis.

OBJECTIVE: The present study investigated regional variations in antibiotic use for the treatment of hospital-acquired infections (HAIs) in France by means of a multilevel analysis, to identify targets for quality improvement. METHODS: Data were obtained from the 2001 and 2006 French national point-prevalence surveys of HAIs and antibiotic use. The present study was conducted using data from 393 nonteaching public hospitals. Data included patient characteristics calculated at the hospital level (mean age and proportion of patients with the following: HAI, presence of a vascular catheter, presence of a urinary catheter, surgical procedure, and immunodeficiency) and hospital characteristics (size and presence of an intensive care unit). The regional effect was measured using a random intercept on a regional variable. RESULTS: Overall, the prevalence of antibiotic use was 5.35% for both study years. The most commonly used antibiotics for HAIs were fluoroquinolones (2001, 1.33%; 2006, 1.35%) and combinations of penicillins with a ß-lactamase inhibitor (2001, 0.92%; 2006, 1.02%). Mapping of antibiotic use showed wide variation between regions. The regional effect explained 3% of antibiotic variation in the unadjusted analysis. In the multivariable analysis, hospital size, high prevalence of patients with immunodeficiency, and infection characteristics explained 45% of the variability in antibiotic use. The regional effect was not retained in the final model. CONCLUSION: The pattern of antibiotic use for HAIs differed over time, and regional variations were mostly explained by patient characteristics; there was no regional effect. Models that take data hierarchy into account are essential to better approach antibiotic use and develop relevant strategies for improvement.