To characterize the top 1% of inpatients who contributed to the 6-month antimicrobial budget in a tertiary, academic medical center and identify cost-effective intervention opportunities targeting high-cost antimicrobial utilization.

Retrospective cohort study.

Top 1% of the antimicrobial budget from July 1 through December 31, 2014.

Patients were identified through a pharmacy billing database. Baseline characteristics were collected through a retrospective medical chart review. Patients were presented to the antimicrobial stewardship team to determine appropriate utilization of high-cost antimicrobials and potential intervention opportunities. Appropriate use was defined as antimicrobial therapy that was effective, safe, and most cost-effective compared with alternative agents.

A total of 10,460 patients received antimicrobials in 6 months; 106 patients accounted for $889,543 (47.2%) of the antimicrobial budget with an antimicrobial cost per day of $219±$192 and antimicrobial cost per admission of $4,733±$7,614. Most patients were immunocompromised (75%) and were followed by the infectious disease consult service (80%). The most commonly prescribed antimicrobials for treatment were daptomycin, micafungin, liposomal amphotericin B, and meropenem. Posaconazole and valganciclovir accounted for most of the prophylactic therapy. Cost-effective opportunities (n=71) were present in 57 (54%) of 106 patients, which included dose optimization, de-escalation, dosage form conversion, and improvement in transitions of care.

Antimicrobial stewardship oversight is important in implementing cost-effective strategies, especially in complex and immunocompromised patients who require the use of high-cost antimicrobials.

Infect Control Hosp Epidemiol 2017;38:259–265

To develop a risk score to predict probability of bloodstream infections (BSIs) due to extended-spectrum β-lactamase–producing Enterobacteriaceae (ESBLE).

Retrospective case-control study.

Two large community hospitals.

Hospitalized adults with Enterobacteriaceae BSI between January 1, 2010, and June 30, 2015.

Multivariate logistic regression was used to identify independent risk factors for ESBLE BSI. Point allocation in extended-spectrum β-lactamase prediction score (ESBL-PS) was based on regression coefficients.

Among 910 patients with Enterobacteriaceae BSI, 42 (4.6%) had ESBLE bloodstream isolates. Most ESBLE BSIs were community onset (33 of 42; 79%), and 25 (60%) were due to Escherichia coli. Independent risk factors for ESBLE BSI and point allocation in ESBL-PS included outpatient procedures within 1 month (adjusted odds ratio [aOR], 8.7; 95% confidence interval [CI], 3.1–22.9; 1 point), prior infections or colonization with ESBLE within 12 months (aOR, 26.8; 95% CI, 7.0–108.2; 4 points), and number of prior courses of β-lactams and/or fluoroquinolones used within 3 months of BSI: 1 course (aOR, 6.3; 95% CI, 2.7–14.7; 1 point), ≥2 courses (aOR, 22.0; 95% CI, 8.6–57.1; 3 points). The area under the receiver operating characteristic curve for the ESBL-PS model was 0.86. Patients with ESBL-PSs of 0, 1, 3, and 4 had estimated probabilities of ESBLE BSI of 0.7%, 5%, 24%, and 44%, respectively. Using ESBL-PS ≥3 to indicate high risk provided a negative predictive value of 97%.

ESBL-PS estimated patient-specific risk of ESBLE BSI with high discrimination. Incorporation of ESBL-PS with acute severity of illness may improve adequacy of empirical antimicrobial therapy and reduce carbapenem utilization.

Infect Control Hosp Epidemiol 2017;38:266–272

To investigate the scale of antimicrobial prescribing without a corresponding visit, and to compare the attributes of patients who received antimicrobials with a corresponding visit with those who did not have a visit.

Retrospective cohort.

We followed up 185,010 Medicare patients for 1 year after an acute myocardial infarction. For each antimicrobial prescribed, we determined whether the patient had an inpatient, outpatient, or provider claim in the 7 days prior to the antimicrobial prescription being filled. We compared the proportions of patient characteristics for those prescriptions associated with a visit and without a visit (ie, phantom prescriptions). We also compared the rates at which different antimicrobials were prescribed without a visit.

We found that of 356,545 antimicrobial prescriptions, 14.75% had no evidence of a visit in the week prior to the prescription being filled. A higher percentage of patients without a visit were identified as white (P<.001) and female (P<.001). Patients without a visit had a higher likelihood of survival and fewer additional cardiac events (acute myocardial infarction, cardiac arrest, stroke, all P<.001). Among the antimicrobials considered, amoxicillin, penicillin, and agents containing trimethoprim and methenamine were much more likely to be prescribed without a visit. In contrast, levofloxacin, metronidazole, moxifloxacin, vancomycin, and cefdinir were much less likely to be prescribed without a visit.

Among this cohort of patients with chronic conditions, phantom prescriptions of antimicrobials are relatively common and occurred more frequently among those patients who were relatively healthy.

Infect Control Hosp Epidemiol 2017;38:273–280

To identify predictors of treatment for urinary tract infections (UTI) among patients undergoing total hip (THA) or knee (TKA) arthroplasties and to assess an intervention based on these predictors.

We conducted a retrospective cohort study of 200 consecutive patients undergoing THA/TKA between February 21, 2011, and June 30, 2011, to identify predictors of treatment for UTI and a prospective cohort study of 50 patients undergoing these procedures between May 21, 2012, and July 17, 2012, to assess the association of signs or symptoms and UTI treatment. We then conducted a before-and-after study to assess whether implementing an intervention affected the frequency of treatment for UTI before or after THA/TKA.

The orthopedics department of a university health center.

Patients undergoing THA or TKA.

Surgeons revised their UTI screening and treatment practices.

Positive leukocyte esterase (P<.0001; P<.0001) and urine white blood cell count>5 (P=.01; P=.01) were associated with preoperative or postoperative UTI treatment. In the prospective study, 12 patients (24%) had signs and symptoms consistent with UTI. The number of patients treated for presumed UTI decreased 80.2% after the surgeons changed their practices, and surgical site infection (SSI) rates, including prosthetic joint infections (PJIs), did not increase.

Urine leukocyte esterase and white blood cell count were the strongest predictors of treatment for UTI before or after THA/TKA. The intervention was associated with a significant decrease in treatment for UTI, and SSI/PJI rates did not increase.

Infect Control Hosp Epidemiol 2017;38:281–286

The impact of healthcare system integration on infection prevention programs is unknown. Using catheter-associated urinary tract infection (CAUTI) prevention as an example, we hypothesize that US Department of Veterans Affairs (VA) nursing homes have a more robust infection prevention infrastructure due to integration and centralization compared with non–VA nursing homes.

VA and non-VA nursing homes participating in the AHRQ Safety Program for Long-Term Care collaborative.

Nursing homes provided baseline information about their infection prevention programs to assess strengths and gaps related to CAUTI prevention via a needs assessment questionnaire.

A total of 353 of 494 nursing homes from 41 states (71%; 47 VA and 306 non-VA facilities) responded. VA nursing homes reported more hours per week devoted to infection prevention-related activities (31 vs 12 hours; P<.001) and were more likely to have committees that reviewed healthcare-associated infections. Compared with non-VA facilities, a higher percentage of VA nursing homes reported tracking CAUTI rates (94% vs 66%; P<.001), sharing CAUTI data with leadership (94% vs 70%; P=.014) and with nursing personnel (85% vs 56%, P=.003). However, fewer VA nursing homes reported having policies for appropriate catheter use (64% vs 81%; P=.004) and catheter insertion (83% vs 94%; P=.004).

Among nursing homes participating in an AHRQ-funded collaborative, VA and non-VA nursing homes differed in their approach to CAUTI prevention. Best practices from both settings should be applied universally to create an optimal infection prevention program within emerging integrated healthcare systems.

Infect Control Hosp Epidemiol 2017;38:287–293

To describe the characteristics and impact of Clostridium difficile infection (CDI) in a long-term acute-care hospital (LTACH).

Retrospective matched cohort study.

A 38-bed, urban, university-affiliated LTACH.

The characteristics of LTACH-onset CDI were assessed among patients hospitalized between July 2008 and October 2015. Patients with CDI were matched to concurrently hospitalized patients without a diagnosis of CDI. Severe CDI was defined as CDI with 2 or more of the following criteria: age ≥65 years, serum creatinine ≥2 mg/dL, or peripheral leukocyte count ≥20,000 cells/μL. A conditional Poisson regression model was developed to determine characteristics associated with a composite primary outcome of 30-day readmission to an acute-care hospital, or mortality.

The overall incidence of CDI was 21.4 cases per 10,000 patient days, with 27% of infections classified as severe. Patients with CDI had a mean age of 70 years (SD, 14 years), a mean Charlson comorbidity index of 3.6 (SD, 2.0), a median length of stay of 33 days (interquartile range [IQR], 24–45 days), and a median time between admission and CDI diagnosis of 16 days (IQR, 9–23 days). The most commonly prescribed antibiotic preceding a CDI diagnosis was a cephalosporin, with median duration of 8 days (IQR, 4–14 days). In multivariate analysis, CDI was not significantly associated with the primary outcome (relative risk, 0.97; 95% CI, 0.59–1.58).

Incidence of CDI in an urban, university-affiliated LTACH was high. Future research should focus on infection prevention measures to decrease the burden of CDI in this complex patient population.

Infect Control Hosp Epidemiol 2017;38:294–299

We investigated an increase in the incidence of healthcare-associated Clostridium difficile infection (CDI) that occurred following a change from a bleach disinfectant to a peracetic acid–based disinfectant.

To evaluate the efficacy of the peracetic acid–based disinfectant.

Laboratory-based product evaluation.

The commercial peracetic acid–based product is activated on site by mixing a small volume of concentrated hydrogen peroxide and peracetic acid present in a “SmartCap” reservoir with the remaining contents of the container. We measured concentrations of peracetic acid in newly activated and in-use product and determined the stability of nonactivated and activated product. We tested the efficacy of the product against C. difficile spores using the American Society for Testing and Materials standard quantitative carrier disk test method.

Measured concentrations of peracetic acid (50–800 parts per million [ppm]) were significantly lower than the level stated on the product label (1,500 ppm), and similar results were obtained for containers from multiple lot numbers and from another hospital. Product with peracetic acid levels below 600 ppm had significantly reduced activity against C. difficile spores. Peracetic acid concentrations were reduced markedly after storage of either activated or nonactivated product for several weeks. The Environmental Protection Agency confirmed the finding of low disinfectant levels and ordered discontinuation of sale of the product.

Use of a defective peracetic acid–based surface disinfectant may have contributed to an increase in healthcare-associated CDI. Our findings highlight the importance of evaluating the efficacy of liquid disinfectants in healthcare settings.

Infect Control Hosp Epidemiol 2017;38:300–305

To define the scope of an outbreak of Legionnaires’ disease (LD), to identify the source, and to stop transmission.

Epidemiologic investigation of an LD outbreak among patients and a visitor exposed to a newly constructed hematology-oncology unit.

An LD case was defined as radiographically confirmed pneumonia in a person with positive urinary antigen testing and/or respiratory culture for Legionella and exposure to the hematology-oncology unit after February 20, 2014. Cases were classified as definitely or probably healthcare-associated based on whether they were exposed to the unit for all or part of the incubation period (2–10 days). We conducted an environmental assessment and collected water samples for culture. Clinical and environmental isolates were compared by monoclonal antibody (MAb) and sequence-based typing.

Over a 12-week period, 10 cases were identified, including 6 definite and 4 probable cases. Environmental sampling revealed Legionella pneumophila serogroup 1 (Lp1) in the potable water at 9 of 10 unit sites (90%), including all patient rooms tested. The 3 clinical isolates were identical to environmental isolates from the unit (MAb2-positive, sequence type ST36). No cases occurred with exposure after the implementation of water restrictions followed by point-of-use filters.

Contamination of the unit’s potable water system with Lp1 strain ST36 was the likely source of this outbreak. Healthcare providers should routinely test patients who develop pneumonia at least 2 days after hospital admission for LD. A single case of LD that is definitely healthcare associated should prompt a full investigation.

Infect Control Hosp Epidemiol 2017;38:306–313

To determine the source of a healthcare-associated outbreak of Pantoea agglomerans bloodstream infections.

Epidemiologic investigation of the outbreak.

Oncology clinic (clinic A).

Cases were defined as Pantoea isolation from blood or catheter tip cultures of clinic A patients during July 2012–May 2013. Clinic A medical charts and laboratory records were reviewed; infection prevention practices and the facility’s water system were evaluated. Environmental samples were collected for culture. Clinical and environmental P. agglomerans isolates were compared using pulsed-field gel electrophoresis.

Twelve cases were identified; median (range) age was 65 (41–78) years. All patients had malignant tumors and had received infusions at clinic A. Deficiencies in parenteral medication preparation and handling were identified (eg, placing infusates near sinks with potential for splash-back contamination). Facility inspection revealed substantial dead-end water piping and inadequate chlorine residual in tap water from multiple sinks, including the pharmacy clean room sink. P. agglomerans was isolated from composite surface swabs of 7 sinks and an ice machine; the pharmacy clean room sink isolate was indistinguishable by pulsed-field gel electrophoresis from 7 of 9 available patient isolates.

Exposure of locally prepared infusates to a contaminated pharmacy sink caused the outbreak. Improvements in parenteral medication preparation, including moving chemotherapy preparation offsite, along with terminal sink cleaning and water system remediation ended the outbreak. Greater awareness of recommended medication preparation and handling practices as well as further efforts to better define the contribution of contaminated sinks and plumbing deficiencies to healthcare-associated infections are needed.

Infect Control Hosp Epidemiol 2017;38:314–319

To describe the investigation and control of a rare cluster of Klebsiella pneumoniae carbapenemase–producing Citrobacter freundii in a hospital in southern Florida.

An epidemiologic investigation, review of infection prevention procedures, and molecular studies including whole genome sequencing were conducted.

An outbreak of K. pneumoniae carbapenemase–3-producing C. freundii was identified at a tertiary hospital in Florida in 2014. Of the 6 cases identified, 3 occurred in the same intensive care unit and were caused by the same clone. For 2 of the 3 remaining cases, the isolates had low carbapenem minimum inhibitory concentrations and were unrelated by whole genome sequencing. As a response to the outbreak, supplementary environmental cleaning was implemented, including closure and terminal cleaning of the unit where the 3 cases clustered, in addition to the infection control bundle already in place at the time. No further cases were identified after these additional interventions.

Although C. freundii is not a species that commonly demonstrates carbapenem resistance, our findings suggest that carbapenemase-producing C. freundii may be underdetected even when active surveillance is in place and has a potential to cause hospital outbreak.

Infect Control Hosp Epidemiol 2017;38:320–326

Adult ventilator-associated event (VAE) definitions include ventilator-associated conditions (VAC) and subcategories for infection-related ventilator-associated complications (IVAC) and possible ventilator-associated pneumonia (PVAP). We explored these definitions for children.

Retrospective cohort

Pediatric, cardiac, or neonatal intensive care units (ICUs) in 6 US hospitals

Patients ≤18 years old ventilated for ≥1 day

We identified patients with pediatric VAC based on previously proposed criteria. We applied adult temperature, white blood cell count, antibiotic, and culture criteria for IVAC and PVAP to these patients. We matched pediatric VAC patients with controls and evaluated associations with adverse outcomes using Cox proportional hazards models.

In total, 233 pediatric VACs (12,167 ventilation episodes) were identified. In the cardiac ICU (CICU), 62.5% of VACs met adult IVAC criteria; in the pediatric ICU (PICU), 54.2% of VACs met adult IVAC criteria; and in the neonatal ICU (NICU), 20.2% of VACs met adult IVAC criteria. Most patients had abnormal white blood cell counts and temperatures; we therefore recommend simplifying surveillance by focusing on “pediatric VAC with antimicrobial use” (pediatric AVAC). Pediatric AVAC with a positive respiratory diagnostic test (“pediatric PVAP”) occurred in 8.9% of VACs in the CICU, 13.3% of VACs in the PICU, and 4.3% of VACs in the NICU. Hospital mortality was increased, and hospital and ICU length of stay and duration of ventilation were prolonged among all pediatric VAE subsets compared with controls.

We propose pediatric AVAC for surveillance related to antimicrobial use, with pediatric PVAP as a subset of AVAC. Studies on generalizability and responsiveness of these metrics to quality improvement initiatives are needed, as are studies to determine whether lower pediatric VAE rates are associated with improvements in other outcomes.

Infect Control Hosp Epidemiol 2017;38:327–333

Few studies have validated ICD-9-CM diagnosis codes for surgical site infection (SSI), and none have validated coding for noninfectious wound complications after mastectomy.

To determine the accuracy of International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes in health insurer claims data to identify SSI and noninfectious wound complications, including hematoma, seroma, fat and tissue necrosis, and dehiscence, after mastectomy.

We reviewed medical records for 275 randomly selected women who were coded in the claims data for mastectomy with or without immediate breast reconstruction and had an ICD-9-CM diagnosis code for a wound complication within 180 days after surgery. We calculated the positive predictive value (PPV) to evaluate the accuracy of diagnosis codes in identifying specific wound complications and the PPV to determine the accuracy of coding for the breast surgical procedure.

The PPV for SSI was 57.5%, or 68.9% if cellulitis-alone was considered an SSI, while the PPV for cellulitis was 82.2%. The PPVs of individual noninfectious wound complications ranged from 47.8% for fat necrosis to 94.9% for seroma and 96.6% for hematoma. The PPVs for mastectomy, implant, and autologous flap reconstruction were uniformly high (97.5%–99.2%).

Our results suggest that claims data can be used to compare rates of infectious and noninfectious wound complications after mastectomy across facilities, even though PPVs vary by specific type of postoperative complication. The accuracy of coding was highest for cellulitis, hematoma, and seroma, and a composite group of noninfectious complications (fat necrosis, tissue necrosis, or dehiscence).

Infect Control Hosp Epidemiol 2017;38:334–339

To determine risk factors for the development of surgical site infections (SSIs) in neurosurgery patients undergoing spinal fusion.

Retrospective case-control study.

Large, academic, quaternary care center.

The study population included all neurosurgery patients who underwent spinal fusion between August 1, 2009, and August 31, 2013. Cases were defined as patients in the study cohort who developed an SSI. Controls were patients in the study cohort who did not develop an SSI.

To achieve 80% power with an ability to detect an odds ratio (OR) of 2, we performed an unmatched case-control study with equal numbers of cases and controls.

During the study period, 5,473 spinal fusion procedures were performed by neurosurgeons in our hospital. With 161 SSIs recorded during the study period, the incidence of SSIs associated with these procedures was 2.94%. While anterior surgical approach was found to be a protective factor (OR, 0.20; 95% confidence interval [CI], 0.08–0.52), duration of procedure (OR, 1.58; 95% CI, 1.29–1.93), American Society of Anesthesiologists score of 3 or 4 (OR, 1.79; 95% CI, 1.00–3.18), and hospitalization within the prior 30 days (OR, 5.8; 95% CI, 1.37–24.57) were found in multivariate analysis to be independent predictors of SSI following spinal fusion. Prior methicillin-resistant Staphylococcus aureus (MRSA) nares colonization was highly associated with odds 20 times higher of SSI following spinal fusion (OR, 20.30; 95% CI, 4.64–8.78).

In additional to nonmodifiable risk factors, prior colonization with MRSA is a modifiable risk factor very strongly associated with development of SSI following spinal fusion.

Infect Control Hosp Epidemiol 2017;38:348–352

Despite significant advances in technological methods for hand hygiene surveillance, a lack of evidence prohibits comparison of systems to one another or against the current gold standard of direct observation.

To validate a hand hygiene monitoring technology (HHMT) designed to capture hand hygiene behaviors aggregated at the hospital-unit level (GOJO Industries, Akron, OH).

Our team followed a rigorous validation approach to assess the sensitivity and positive predictive value (PPV) of an HHMT. A planned path was first used to measure the accuracy of the system when purposefully activated by investigators. Next, behavioral validation was used to quantify accuracy of the system in capturing real-world behaviors.

During the planned path phase, investigators performed 4,872 unique events across 3 distinct hospital buildings varying in size and age since construction. Overall sensitivity across the medical center was 88.7% with a PPV of 99.2%. During the behavioral validation phase, trained direct observers recorded 5,539 unique events across 3 distinct hospital buildings. Overall sensitivity across the medical center was 92.7% and PPV was 84.4%.

Objective measures of sensitivity and PPV indicate the promise of the benefit of this and other HHMTs to capture basic behaviors associated with hand hygiene.

Infect Control Hosp Epidemiol 2017;38:348–352