To describe the epidemiology of ventilator-associated pneumonia (VAP) in community hospitals.

Prospective study in 31 community hospitals from 2007 to 2011.

VAP surveillance was performed by infection preventionists using the National Healthcare Safety Network protocol. VAP incidence was reported as number of events per 1,000 ventilator-days. We categorized hospitals into small (<30,000 patient-days/year), medium (30,000–60,000 patient-days/year), and large (>60,000 patient-days/year) groups and compared VAP incidence by hospital size.

The median VAP incidence was 1.4 (interquartile range, 0.4–2.4), and ventilator utilization ratio (VUR) was 0.33 (0.25–0.47). VAP incidence was higher in small hospitals (2.1) than medium (0.85) or large (0.69) hospitals (P = .03) despite a lower VUR in small hospitals (0.29 vs 0.31 vs 0.44, respectively; P = .01). The median age of 247 VAP cases was 64 (53-73); 136 (55.1%) were female; 142 (57.5%) were Caucasian; 170 (68.8%) were admitted from home. The length of stay and duration of ventilation were 26 (14–42) and 12 (4–21) days, respectively. The pre- and postinfection hospital stays were 8 (3–13) days and 14 (8–30) days, respectively. Data on outcomes were available in 214 cases (86.6%), and 75 (35.0%) cases died during hospitalization. The top 3 pathogens were methicillin-resistant Staphylococcus aureus (MRSA; n = 70, 27.9%), Pseudomonas species (n = 40, 16.3%), and Klebsiella species (n = 34, 13.3%).

VAP incidence was inversely associated with size of hospital. VAP in community hospitals was frequently caused by MRSA. Importantly, predictors of VAP incidence in tertiary care hospitals such as VUR may not be predictive in community hospitals with few ventilated patients.

To evaluate the application of the National Healthcare Safety Network (NHSN) central line-associated bloodstream infection (CLABSI) definition in pediatric intensive care units (PICUs) and pediatric hematology/oncology units (PHOUs) participating in a multicenter quality improvement collaborative to reduce CLABSIs; to identify sources of variability in the application of the definition.

Online survey using 18 standardized case scenarios. Each described a positive blood culture in a patient and required a yes-or-no answer to the question “Is this a CLABSI?” NHSN staff responses were the reference standard.

Sixty-five US PICUs and PHOUs.

Staff who routinely adjudicate CLABSIs using NHSN definitions.

Sixty responses were received from 58 (89%) of 65 institutions; 78% of respondents were infection preventionists, infection control officers, or infectious disease physicians. Responses matched those of NHSN staff for 78% of questions. The mean (SE) percentage of concurring answers did not differ for scenarios evaluating application of 1 of the 3 criteria (“known pathogen,” 78% [1.7%]; “skin contaminant, >1 year of age,” 76% [SE, 2.5%]; “skin contaminant, ≤1 year of age,” 81% [3.8%]; P = .3 ). The mean percentage of concurring answers was lower for scenarios requiring respondents to determine whether a CLABSI was present or incubating on admission (64% [4.6%]; P = .017) or to distinguish between primary and secondary bacteremia (65% [2.5%]; P = .021).

The accuracy of application of the CLABSI definition was suboptimal. Efforts to reduce variability in identifying CLABSIs that are present or incubating on admission and in distinguishing primary from secondary bloodstream infection are needed.

To describe the epidemiology of infections related to the use of implantable central venous access devices (CVADs) in cancer patients and to evaluate measures aimed at reducing the rates of such infections.

Prospective cohort study.

Referral hospital for cancer in São Paulo, Brazil.

We prospectively evaluated all implantable CVADs employed between January 2009 and December 2011. Inpatients and outpatients were followed until catheter removal, transfer to another facility, or death.

Outcome measures were bloodstream infection and pocket infection. We also evaluated the effects that the creation of a multidisciplinary team for CVAD care, avoiding in-hospital implantation of CVADs, and limiting CVAD insertion in neutropenic patients have on the rates of such infections.

During the study period, 966 CVADs (mostly venous ports) were implanted in 933 patients, for a combined total of 243,792 catheter-days. We identified 184 episodes of infection: 154 (84%) were bloodstream infections, 21 (11%) were pocket infections, and 9 (5%) were surgical site infections. During the study period, the rate of CVAD-related infection dropped from 2.2 to 0.24 per 1,000 catheter-days (P < .001). Multivariate analysis revealed that relevant risk factors for such infection include surgical reintervention, implantation in a neutropenic patient, in-hospital implantation, use of a cuffed catheter, and nonchemotherapy indication for catheter use.

Establishing a multidisciplinary team specifically focused on CVAD care, together with systematic reporting of infections, appears to reduce the rates of infection related to the use of these devices.

Little is known about whether those performing healthcare-associated infection (HAI) surveillance vary in their interpretations of HAI definitions developed by the Centers for Disease Control and Prevention's National Healthcare Safety Network (NHSN). Our primary objective was to characterize variations in these interpretations using clinical vignettes. We also describe predictors of variation in responses.

Cross-sectional study.

United States.

A sample of US-based members of the Society for Healthcare Epidemiology of America (SHEA) Research Network.

Respondents assessed whether each of 6 clinical vignettes met criteria for an NHSN-defined HAI. Individual- and institutional-level data were also gathered.

Surveys were distributed to 143 SHEA Research Network members from 126 hospitals. In total, 113 responses were obtained, representing at least 61 unique hospitals (30 respondents did not identify a hospital); 79.2% (84 of 106 nonmissing responses) were infection preventionists, and 79.4% (81 of 102 nonmissing responses) worked at academic hospitals. Among the 6 vignettes, the proportion of respondents correctly characterizing the vignettes was as low as 27.3%. Combining all 6 vignettes, the mean percentage of correct responses was 61.1% (95% confidence interval, 57.7%–63.8%). Percentage of correct responses was associated with presence of a clinical background (ie, nursing or physician degrees) but not with hospital size or infection prevention and control department characteristics.

Substantial heterogeneity exists in the application of HAI definitions in this survey of infection preventionists and hospital epidemiologists. Our data suggest a need to better clarify these definitions, especially when comparing HAI rates across institutions.

To determine the prevalence of methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), and Clostridium difficile infection (CDI) in Canadian hospitals.

National point prevalence survey in November 2010.

Canadian acute care hospitals with at least 50 beds.

Adult inpatients colonized or infected with MRSA or VRE or with CDI.

The prevalence (per 100 inpatients) of MRSA, VRE, and CDI was determined. Associations between prevalence and institutional characteristics and infection control policies were evaluated.

One hundred seventy-six hospitals (65% of those eligible) participated. The median (range) prevalence rates for MRSA and VRE colonization or infection and CDI were 4.2% (0%–22.1%), 0.5% (0%–13.1%), and 0.9% (0%–8.6%), respectively. Median MRSA and VRE infection rates were low (0.3% and 0%, respectively). MRSA, VRE, and CDI were thought to have been healthcare associated in 79%, 96%, and 84% of cases, respectively. In multivariable analysis, routine use of a private room for colonized/infected patients was associated with lower median MRSA infection rate (prevalence ratio [PR], 0.44 [95% confidence interval (CI), 0.22–0.88]) and VRE prevalence (PR, 0.26 [95% CI, 0.12–0.57]). Lower VRE rates were also associated with enhanced environmental cleaning (PR, 0.52 [95% CI, 0.36–0.75]). Higher bed occupancy rates were associated with higher rates of CDI (PR, 1.02 [95% CI, 1.01–1.03]).

These data provide the first national prevalence estimates for MRSA, VRE, and CDI in Canadian hospitals. Certain infection prevention and control policies were found to be associated with prevalence and deserve further investigation.

To assess the impact of an electronic surveillance system on isolation practices and rates of methicillin-resistant Staphylococcus aureus (MRSA).

A pre-post test intervention.

Inpatient units (except psychiatry and labor and delivery) in 4 New York City hospitals.

All patients for whom isolation precautions were indicated, May 2009–December 2011.

Trained observers assessed isolation sign postings, availability of isolation carts, and staff use of personal protective equipment (PPE). Infection rates were obtained from the infection control department. Regression analyses were used to examine the association between the surveillance system, infection prevention practices, and MRSA infection rates.

A total of 54,159 isolation days and 7,628 staff opportunities for donning PPE were observed over a 31-month period. Odds of having an appropriate sign posted were significantly higher after intervention than before intervention (odds ratio [OR], 1.10 [95% confidence interval {CI}, 1.01–1.20]). Relative to baseline, postintervention sign posting improved significantly for airborne and droplet precautions but not for contact precautions. Sign posting improved for vancomycin-resistant enterococci (OR, 1.51 [95% CI, 1.23–1.86]; P = .0001), Clostridium difficile (OR, 1.59 [95% CI, 1.27–2.02]; P = .00005), and Acinetobacter baumannii (OR, 1.41 [95% CI, 1.21–1.64]; P = .00001) precautions but not for MRSA precautions (OR, 1.11 [95% CI, 0.89–1.39]; P = .36). Staff and visitor adherence to PPE remained low throughout the study but improved from 29.1% to 37.0% after the intervention (OR, 1.14 [95% CI, 1.01–1.29]). MRSA infection rates were not significantly different after the intervention.

An electronic surveillance system resulted in small but statistically significant improvements in isolation practices but no reductions in infection rates over the short term. Such innovations likely require considerable uptake time.

The year 2010 is the first time that the Rhode Island hospital discharge database included present on admission (POA) indicators, which give us the opportunity to distinguish cases of hospital-onset Clostridium difficile infection (CDI) from cases of community-onset CDI and to assess the burden of hospital-onset CDI in patients discharged from Rhode Island hospitals during 2010 and 2011.

Observational study.

Patients 18 years of age or older discharged from one of Rhode Island's 11 acute-care hospitals between January 1, 2010, and December 31, 2011.

Using the newly available POA indicators in the Rhode Island 2010 and 2011 hospital discharge database, we identified patients with hospital-onset CDI and without CDI. Adjusting for patient demographic and clinical characteristics using propensity score matching, we measured between-group differences in mortality, length of stay, and cost for patients with hospital-onset CDI and without CDI.

In 2010 and 2011, the 11 acute-care hospitals in Rhode Island had 225,999 discharges. Of 4,531 discharged patients with CDI (2.0% of all discharges), 1,211 (26.7%) had hospital-onset CDI. After adjusting for patient demographic and clinical characteristics, discharged patients with hospital-onset CDI were found to have higher mortality rates, longer lengths of stay, and higher costs than those without CDI.

Our results highlight the burden of hospital-onset CDI in Rhode Island. These findings emphasize the need to track longitudinal trends to tailor and target population-health and quality-improvement initiatives.

Acute hepatitis B virus (HBV) infections have been reported in long-term care facilities (LTCFs), primarily associated with infection control breaks during assisted blood glucose monitoring. We investigated HBV outbreaks that occurred in separate skilled nursing facilities (SNFs) to determine factors associated with transmission.

Outbreak investigation with case-control studies.

Two SNFs (facilities A and B) in Durham, North Carolina, during 2009–2010.

Residents with acute HBV infection and controls randomly selected from HBV-susceptible residents during the outbreak period.

After initial cases were identified, screening was offered to all residents, with repeat testing 3 months later for HBV-susceptible residents. Molecular testing was performed to assess viral relatedness. Infection control practices were observed. Case-control studies were conducted to evaluate associations between exposures and acute HBV infection in each facility.

Six acute HBV cases were identified in each SNF. Viral phylogenetic analysis revealed a high degree of HBV relatedness within, but not between, facilities. No evaluated exposures were significantly associated with acute HBV infection in facility A; those associated with infection in facility B (all odds ratios >20) included injections, hospital or emergency room visits, and daily blood glucose monitoring. Observations revealed absence of trained infection control staff at facility A and suboptimal hand hygiene practices during blood glucose monitoring and insulin injections at facility B.

These outbreaks underscore the vulnerability of LTCF residents to acute HBV infection, the importance of surveillance and prompt investigation of incident cases, and the need for improved infection control education to prevent transmission.

To assess the knowledge, attitudes, and practices of infection control among staff in a residential care facility for children and young adults with neurologic and neurodevelopmental conditions.

Self-administered survey.

Residential care facility (facility A).

Facility A staff (N = 200).

We distributed a survey to staff at facility A. We classified staff with direct care responsibilities as clinical (ie, physicians, nurses, and therapists) or nonclinical (ie, habilitation assistants, volunteers, and teachers) and used X2 tests to measure differences between staff agreement to questions.

Of 248 surveys distributed, 200 (81%) were completed; median respondent age was 36 years; 85% were female; and 151 were direct care staff (50 clinical, 101 nonclinical). Among direct care staff respondents, 86% agreed they could identify residents with respiratory symptoms, 70% stayed home from work when ill with respiratory infection, 64% agreed that facility administration encouraged them to stay home when ill with respiratory infection, and 72% reported that ill residents with respiratory infections were separated from well residents. Clinical and nonclinical staff differed in agreement about using waterless hand gel as a substitute for handwashing (96% vs 78%; P = .005) and whether handwashing was done after touching residents (92% vs 75%; P = .04).

Respondents' knowledge, attitudes, and practices regarding infection control could be improved, especially among nonclinical staff. Facilities caring for children and young adults with neurologic and neurodevelopmental conditions should encourage adherence to infection control best practices among all staff having direct contact with residents.

To describe practices for influenza vaccination of healthcare personnel (HCP) with emphasis on correlates of increased vaccination rates.

Survey.

Volunteer sample of hospitals in Louisiana.

All hospitals in Louisiana were invited to participate. A 17-item questionnaire inquired about the hospital type, patients served, characteristics of the vaccination campaign, and the resulting vaccination rate.

Of 254 hospitals, 153 (60%) participated and were included in the 124 responses that were received. Most programs (64%) required that HCP either receive the vaccine or sign a declination form, and the rest were exclusively voluntary (36%); no program made vaccination a condition of employment. The median vaccination rate was 67%, and the vaccination rate was higher among hospitals that were accredited by the Joint Commission; provided acute care; served children, pregnant women, oncology patients, or intensive care unit patients; required a signed declination form; or imposed consequences for unvaccinated HCP (the most common of which was to require that a mask be worn on patient contact). Hospitals that provided free vaccine, made vaccine widely available, advertised the program extensively, required a declination form, and imposed consequences had the highest vaccination rates (median, 86%; range, 81%–91%).

The rate of influenza vaccination of HCP remains low among the hospitals surveyed. Recommended practices may not be enough to reach 90% vaccination rates unless a signed declination requirement and consequences are implemented. Wearing a mask is a strong consequence. Demanding influenza vaccination as a condition of employment was not reported as a practice by the participating hospitals.