To assess hand hygiene improvement and sustainability associated with a Breakthrough Collaborative.

Multicenter analysis of hand hygiene compliance through direct observation by trained observers.

A total of 5 publicly funded hospitals in 14 locations, with a total of 1,152 beds, in the County of Vaud, Switzerland.

Clinical staff.

In total, 59,272 opportunities for hand hygiene were monitored for the duration of the study, for an average of 5,921 per audit (range, 5,449–6,852). An 18-month Hand Hygiene Breakthrough Collaborative was conducted to implement the WHO multimodal promotional strategy including improved access to alcohol-based hand rub, education, performance measurement and feedback, reminders and communication, leadership engagement, and safety culture.

Overall hand hygiene compliance improved from 61.9% to 88.3% (P<.001) over 18 months and was sustained at 88.9% (P=.248) 12 months after the intervention. Hand hygiene compliance among physicians increased from 62% to 85% (P<.001) and finally 86% at follow-up (P=.492); for nursing staff, compliance improved from 64% to 90% (P<.001) and finally 90% at follow-up (P=.464); for physiotherapists compliance improved from 50% to 90% (P<.001) and finally 91% at follow-up (P=.619); for X-ray technicians compliance improved from 45% to 80% (P<.001) and finally 81% at follow-up (P=.686). Hand hygiene compliance also significantly increased with sustained improvement across all hand hygiene indications and all hospitals.

A rigorously conducted multicenter project combining the Breakthrough Collaborative method for its structure and the WHO multimodal strategy for content and measurement was associated with significant and substantial improvement in compliance across all professions, all hand hygiene indications, and all participating hospitals.

Infect Control Hosp Epidemiol 2017;38:1420–1427

For alcohol-based hand rubs, the currently recommended application time of 30 seconds is longer than the actual time spent in clinical practice. We investigated whether a shorter application time of 15 seconds is microbiologically safe in neonatal intensive care and may positively influence compliance with the frequency of hand antisepsis actions.

We conducted in vitro experiments to determine the antimicrobial efficacy of hand rubs within 15 seconds, followed by clinical observations to assess the effect of a shortened hand antisepsis procedure under clinical conditions in a neonatal intensive care unit (NICU). An independent observer monitored the frequency of hand antisepsis actions during shifts.

All tested hand rubs fulfilled the requirement of equal or even significantly higher efficacy within 15 seconds when compared to a reference alcohol propan-2-ol 60% (v/v) within 30 seconds. Microbiologically, reducing the application time to 15 seconds had a similar effect when compared to 30-second hand rubbing, but it resulted in significantly increased frequency of hand antisepsis actions (7.9±4.3 per hour vs 5.8±2.9 per hour; P=.05).

Time pressure and workload are recognized barriers to compliance. Therefore, reducing the recommended time for hand antisepsis actions, using tested and well-evaluated hand rub formulations, may improve hand hygiene compliance in clinical practice.

Infect Control Hosp Epidemiol 2017;38:1430–1434

The VU University Medical Center, a tertiary-care hospital in the Netherlands, has adopted a dress code based on national guidelines. It includes uniforms provided by the hospital and a ‘bare-below-the-elbow’ policy for all healthcare workers (HCWs) in direct patient care. Because compliance was poor, we sought to improve adherence by interventions targeted at the main causes of noncompliance.

To measure compliance with the dress code, to assess causes of noncompliance and to assess whether a behavioral approach (combing a nominal group technique with participatory action) is effective in improving compliance

Between March 2014 and June 2016, a total of 1,920 HCWs were observed in hospital hallways for adherence to the policy, at baseline, and at follow-up measurements. Based on the outcome of the baseline measurement, a nominal group technique was applied to assess causes of noncompliance. The causes revealed served as input for interventions that were developed, prioritized, and tailored to specific groups of HCWs and specific departments through participatory action.

We identified lack of knowledge, lack of facilities, and negative attitudes as the main causes of noncompliance. The importance of each cause varied for different groups of HCWs. Tailored interventions targeted at these causes increased overall compliance by 39.6% (95% CI, 31.7–47.5).

The combination of a nominal group technique and participatory action approach is an effective method to increase and sustain compliance with hospital dress code. This combined approach may also be useful to improve adherence to other guidelines.

Infect Control Hosp Epidemiol 2017;38:1435–1440

Burn patients are particularly vulnerable to infection, and an estimated half of all burn deaths are due to infections. This study explored risk factors for healthcare-associated infections (HAIs) in adult burn patients.

Retrospective cohort study.

Tertiary-care burn center.

Adults (≥18 years old) admitted with burn injury for at least 2 days between 2004 and 2013.

HAIs were determined in real-time by infection preventionists using Centers for Disease Control and Prevention criteria. Multivariable Cox proportional hazards regression was used to estimate the direct effect of each risk factor on time to HAI, with inverse probability of censor weights to address potentially informative censoring. Effect measure modification by burn size was also assessed.

Overall, 4,426 patients met inclusion criteria, and 349 (7.9%) patients had at least 1 HAI within 60 days of admission. Compared to <5% total body surface area (TBSA), patients with 5%–10% TBSA were almost 3 times as likely to acquire an HAI (hazard ratio [HR], 2.92; 95% CI, 1.63–5.23); patients with 10%–20% TBSA were >6 times as likely to acquire an HAI (HR, 6.38; 95% CI, 3.64–11.17); and patients with >20% TBSA were >10 times as likely to acquire an HAI (HR, 10.33; 95% CI, 5.74–18.60). Patients with inhalational injury were 1.5 times as likely to acquire an HAI (HR, 1.61; 95% CI, 1.17–2.22). The effect of inhalational injury (P=.09) appeared to be larger among patients with ≤20% TBSA.

Larger burns and inhalational injury were associated with increased incidence of HAIs. Future research should use these risk factors to identify potential interventions.

Infect Control Hosp Epidemiol 2017;38:1441–1448

To examine knowledge and attitudes about influenza vaccination and infection prevention practices among healthcare personnel (HCP) in a long-term-care (LTC) setting.

Knowledge, attitudes, and practices (KAP) survey.

An LTC facility in St Louis, Missouri.

All HCP working at the LTC facility were eligible to participate, regardless of department or position. Of 170 full- and part-time HCP working at the facility, 73 completed the survey, a 42.9% response rate.

Most HCP agreed that respiratory viral infections were serious and that hand hygiene and face mask use were protective. However, only 46% could describe the correct transmission-based precautions for an influenza patient. Correctly answering infection prevention knowledge questions did not vary by years of experience but did vary for HCP with more direct patient contact versus less patient contact. Furthermore, 42% of respondents reported working while sick, and 56% reported that their coworkers did. In addition, 54% reported that facility policies made staying home while ill difficult. Some respondents expressed concerns about the safety (22%) and effectiveness (27%) of the influenza vaccine, and 28% of respondents stated that they would not get the influenza vaccine if it was not required.

This survey of staff in an LTC facility identified several areas for policy improvement, particularly sick leave, as well as potential targets for interventions to improve infection prevention knowledge and to address HCP concerns about influenza vaccination to improve HCP vaccination rates in LTCs.

Infect Control Hosp Epidemiol 2017;38:1449–1456

The relationship between hospital antibiotic use and antibiotic resistance is poorly understood. We evaluated the association between antibiotic utilization and resistance in academic and community hospitals in Ontario, Canada.

We conducted a multicenter observational ecological study of 37 hospitals in 2014. Hospital antibiotic purchasing data were used as an indicator of antibiotic use, whereas antibiotic resistance data were extracted from hospital indexes of resistance. Multivariate regression was performed, with antibiotic susceptibility as the primary outcome, antibiotic consumption as the main predictor, and additional covariates of interest (ie, hospital type, laboratory standards, and patient days).

With resistance data representing more than 90,000 isolates, we found the increased antibiotic consumption in defined daily doses per 1,000 patient days (DDDs/1,000 PD) was associated with decreased antibiotic susceptibility for Pseudomonas aeruginosa (−0.162% per DDD/1,000 PD; P=.119). However, increased antibiotic consumption predicted increased antibiotic susceptibility significantly for Escherichia coli (0.173% per DDD/1,000 PD; P=.005), Klebsiella spp (0.124% per DDD/1,000 PD; P=.004), Enterobacter spp (0.194% per DDD/1,000 PD; P=.003), and Enterococcus spp (0.309% per DDD/1,000 PD; P=.001), and nonsignificantly for Staphylococcus aureus (0.012% per DDD/1,000 PD; P=.878). Hospital type (P=.797) and laboratory standard (P=.394) did not significantly predict antibiotic susceptibility, while increased hospital patient days generally predicted increased organism susceptibility (0.728% per 10,000 PD; P<.001).

We found that hospital-specific antibiotic usage was generally associated with increased, rather than decreased hospital antibiotic susceptibility. These findings may be explained by community origins for many hospital-diagnosed infections and practitioners choosing agents based on local antibiotic resistance patterns.

Infect Control Hosp Epidemiol 2017;38:1457–1463

Infections caused by multidrug-resistant gram-negative organisms (MDRGNOs) have been increasing every year. The objective of this study was to describe the prevalence of MDRGNOs and factors associated with MDRGNOs in patients with spinal cord injury or disorder (SCI/D).

Retrospective cohort study.

Department of Veterans Affairs (VA) electronic health record data from 142 VA facilities were evaluated for 19,642 patients with SCI/D. Multivariable cluster-adjusted models were fit to identify factors associated with MDRGNO.

Gram-negative (GN) cultures occurred in 44% of patients with SCI/D receiving care at VA facilities, and 11,527 (41.3%) GN cultures had an MDRGNO. The most frequent GN organisms (GNOs) were Escherichia coli (28.5%), Klebsiella pneumoniae (17.0%), and Pseudomonas aeruginosa (16.0%). Two-thirds of GN cultures were from the outpatient setting, where MDRGNO prevalence was 37.6%. Significant geographic variation in the prevalence of MDRGNOs was identified (South, 44.7%; Northeast, 44.3%; West, 36.8%; Midwest, 34.4%). Other factors associated with an MDRGNO were older age, injury characteristics, comorbidities, specimen type, healthcare setting, and healthcare exposure. Black (odds ratio [OR], 1.58; 95% confidence interval [CI], 1.39–1.78) and Hispanic race (OR, 1.58; 95% CI, 1.28–1.95), polymicrobial culture (OR, 2.67; 95% CI, 2.46–2.90), and antibiotic use in the previous 90 days (OR, 1.62; 95% CI, 1.50–1.76) were also associated with having an MDRGNO.

MDRGNOs were common in community and healthcare settings among veterans with SCI/D, with significant geographic variation. Health care and antibiotic exposures were significant factors associated with MDRGNOs. Priority should be given to controlling the spread of MDRGNOs in this special population, including a focus on judicious use of antibiotics.

Infect Control Hosp Epidemiol 2017;38:1464–1471

The attributable cost of Clostridium difficile infection (CDI) in children is unknown. We sought to determine a national estimate of attributable cost and length of stay (LOS) of CDI occurring during hospitalization in children.

We analyzed discharge records of patients between 2 and 18 years of age from the Agency for Healthcare Research and Quality (AHRQ) Kids’ Inpatient Database. We created a logistic regression model to predict CDI during hospitalization based on demographic and clinical characteristics. Predicted probabilities from the logistic regression model were then used as propensity scores to match 1:2 CDI to non-CDI cases. Charges were converted to costs and compared between patients with CDI and propensity-score–matched controls. In a sensitivity analysis, we adjusted for LOS as a confounder by including it in both the propensity score and a generalized linear model predicting cost.

We identified 8,527 pediatric hospitalizations (0.53%) with a diagnosis of CDI and 1,597,513 discharges without CDI. In our matched cohorts, the attributable cost of CDI occurring during a hospitalization ranged from $1,917 to $8,317, depending on whether model was adjusted for LOS. When not adjusting for LOS, CDI-associated hospitalizations cost 1.6 times more than non-CDI associated hospitalizations. Attributable LOS of CDI was approximately 4 days.

Clostridium difficile infection in hospitalized children is associated with an economic burden similar to adult estimates. This finding supports a continued focus on preventing CDI in children as a priority. Pediatric CDI cost analyses should account for LOS as an important confounder of cost.

Infect Control Hosp Epidemiol 2017;38:1472–1477

Reported per-patient costs of Clostridium difficile infection (CDI) vary by 2 orders of magnitude among different hospitals, implying that infection control officers need precise, local analyses to guide rational decision making between interventions.

We sought to comprehensively estimate changes in length of stay (LOS) attributable to CDI at a single urban tertiary-care facility using only data automatically extractable from the electronic medical record (EMR).

We performed a retrospective cohort study of 171,938 visits spanning a 7-year period. In total, 23,968 variables were extracted from EMR data recorded within 24 hours of admission to train elastic-net regularized logistic regression models for propensity score matching. To address time-dependent bias (reverse causation), we separately stratified comparisons by time of infection, and we fit multistate models.

The estimated difference in median LOS for propensity-matched cohorts varied from 3.1 days (95% CI, 2.2–3.9) to 10.1 days (95% CI, 7.3–12.2) depending on the case definition; however, dependency of the estimate on time to infection was observed. Stratification by time to first positive toxin assay, excluding probable community-acquired infections, showed a minimum excess LOS of 3.1 days (95% CI, 1.7–4.4). Under the same case definition, the multistate model averaged an excess LOS of 3.3 days (95% CI, 2.6–4.0).

In this study, 2 independent time-to-infection adjusted methods converged on similar excess LOS estimates. Changes in LOS can be extrapolated to marginal dollar costs by multiplying by average costs of an inpatient day. Infection control officers can leverage automatically extractable EMR data to estimate costs of CDI at their own institutions.

Infect Control Hosp Epidemiol. 2017;38:1478–1486

To identify, using a novel enhanced method of recovery, environmental sites where spores of Clostridium difficile persist despite cleaning and hydrogen peroxide aerial decontamination.

Cohort study.

Tertiary referral center teaching hospital.

In total, 16 sites representing high-frequency contact or difficult-to-clean surfaces in a single-isolation room or bed area in patient bed bays were sampled before and after terminal or hydrogen peroxide disinfection using a sponge swab. In some rooms, individual sites were not present (eg, there were no en-suite rooms in the ICU). Swab contents were homogenized, concentrated by membrane-filtration, and plated onto selective media. Results of C. difficile sampling were used to focus cleaning.

Over 1 year, 2,529 sites from 146 rooms and 44 bays were sampled. Clostridium difficile was found on 131 of 572 surfaces (22.9%) before terminal cleaning, on 105 of 959 surfaces (10.6%) after terminal cleaning, and on 43 of 967 surfaces (4.4%) after hydrogen peroxide disinfection. Clostridium difficile persisted most frequently on floor corners (97 of 334; 29.0%) after disinfection. Between the first and third quarters, we observed a significant decrease in the number of positive sites (25 of 390 vs 6 of 256). However, no similar change in the number of isolates before terminal cleaning was observed.

Persistence of C. difficile in the clinical environment was widespread. Although feedback of results did not improve the efficacy of manual disinfection, numbers of C. difficile following hydrogen peroxide gradually declined.

Infect Control Hosp Epidemiol 2017;38:1487–1492

To identify clinical variables that influence blood culture volume recovery

Retrospective chart review and linear model analysis

A 621-bed Academic Medical Center with a Clinical Laboratory that processes 20,000+ blood cultures annually and dedicated phlebotomy staff for venipuncture

Consecutive patients requiring blood culture

Over a 6-day period, blood volume was determined in 568 culture bottles from 128 unique adult patients, and clinical data from the time of phlebotomy were extracted from hospital electronic medical records. Conditional hierarchical linear models with random effects for patient and phlebotomy occasion were utilized to analyze correlations between values collected from the same patient and during the same phlebotomy occasion.

Blood samples obtained from a central venous catheter yielded, on average, 2.53 mL more blood (95% CI, 1.63–3.44 mL; P<.001) than those from peripheral venipuncture, and aerobic bottles contained 0.38 mL more blood (95% CI, 0.1–0.67 mL; P=.009) than the anaerobic bottles. The remaining clinical variables (eg, hospital department, patient age, body mass index, gender, mean arterial pressure, concomitant systemic antibiotic use, and Charlson comorbidity index score) failed to reach statistical significance (P<.05) in relation to volume.

Blood cultures obtained from central venous catheters contain significantly greater volume than those obtained via peripheral venipuncture. These data highlight the clinically significant issue of low culture volume recovery, indicate that diagnostic and prognostic tools that rely on volume-dependent phenomena (ie, time to positivity) may require further validation under usual clinical practice circumstances, and suggest goals for future institutional performance improvement.

Infect Control Hosp Epidemiol 2017;38:1493–1497