Michael D. Howell

Strategies to Prevent Ventilator-Associated Pneumonia in Acute Care Hospitals: 2014 Update

Previously published guidelines are available that provide comprehensive recommendations for detecting and preventing healthcare-associated infections (HAIs). The intent of this document is to highlight practical recommendations in a concise format to assist acute care hospitals in implementing and prioritizing strategies to prevent ventilator-associated pneumonia (VAP) and other ventilator-associated events (VAEs) and to improve outcomes for mechanically ventilated adults, children, and neonates. This document updates “Strategies to Prevent Ventilator-Associated Pneumonia in Acute Care Hospitals,” published in 2008. This expert guidance document is sponsored by the Society for Healthcare Epidemiology of America (SHEA) and is the product of a collaborative effort led by SHEA, the Infectious Diseases Society of America (IDSA), the American Hospital Association (AHA), the Association for Professionals in Infection Control and Epidemiology (APIC), and The Joint Commission, with major contributions from representatives of a number of organizations and societies with content expertise. The list of endorsing and supporting organizations is presented in the introduction to the 2014 updates.

Reappraisal of Routine Oral Care With Chlorhexidine Gluconate for Patients Receiving Mechanical Ventilation: Systematic Review and Meta-Analysis

IMPORTANCEnRegular oral care with chlorhexidine gluconate is standard of care for patients receiving mechanical ventilation in most hospitals. This policy is predicated on meta-analyses suggesting decreased risk of ventilator-associated pneumonia, but these meta-analyses may be misleading because of lack of distinction between cardiac surgery and non-cardiac surgery studies, conflation of open-label vs double-blind investigations, and insufficient emphasis on patient-centered outcomes such as duration of mechanical ventilation, length of stay, and mortality.nnnOBJECTIVEnTo evaluate the impact of routine oral care with chlorhexidine on patient-centered outcomes in patients receiving mechanical ventilation.nnnDATA SOURCESnPubMed, Embase, CINAHL, and Web of Science from inception until July 2013 without limits on date or language.nnnSTUDY SELECTIONnRandomized clinical trials comparing chlorhexidine vs placebo in adults receiving mechanical ventilation. Of 171 unique citations, 16 studies including 3630 patients met inclusion criteria.nnnDATA EXTRACTION AND SYNTHESISnEligible trials were independently identified, evaluated for risk of bias, and extracted by 2 investigators. Differences were resolved by consensus. We stratified studies into cardiac surgery vs non-cardiac surgery and open-label vs double-blind investigations. Eligible studies were pooled using random-effects meta-analysis.nnnMAIN OUTCOMES AND MEASURESnVentilator-associated pneumonia, mortality, duration of mechanical ventilation, intensive care unit and hospital length of stay, antibiotic prescribing.nnnRESULTSnThere were fewer lower respiratory tract infections in cardiac surgery patients randomized to chlorhexidine (relative risk [RR], 0.56 [95% CI, 0.41-0.77]) but no significant difference in ventilator-associated pneumonia risk in double-blind studies of non-cardiac surgery patients (RR, 0.88 [95% CI, 0.66-1.16]). There was no significant mortality difference between chlorhexidine and placebo in cardiac surgery studies (RR, 0.88 [95% CI, 0.25-2.14]) and nonsignificantly increased mortality in non-cardiac surgery studies (RR, 1.13 [95% CI, 0.99-1.29]). There were no significant differences in mean duration of mechanical ventilation or intensive care length of stay. Data on hospital length of stay and antibiotic prescribing were limited.nnnCONCLUSIONS AND RELEVANCEnRoutine oral care with chlorhexidine prevents nosocomial pneumonia in cardiac surgery patients but may not decrease ventilator-associated pneumonia risk in non-cardiac surgery patients. Chlorhexidine use does not affect patient-centered outcomes in either population. Policies encouraging routine oral care with chlorhexidine for non-cardiac surgery patients merit reevaluation.

Automated Surveillance for Ventilator-Associated Events

BACKGROUNDnThe US Centers for Disease Control and Prevention has implemented a new, multitiered definition for ventilator-associated events (VAEs) to replace their former definition of ventilator-associated pneumonia (VAP). We hypothesized that the new definition could be implemented in an automated, efficient, and reliable manner using the electronic health record and that the new definition would identify different patients than those identified under the previous definition.nnnMETHODSnWe conducted a retrospective cohort analysis using an automated algorithm to analyze all patients admitted to the ICU at a single urban, tertiary-care hospital from 2008 to 2013.nnnRESULTSnWe identified 26,466 consecutive admissions to the ICU, 10,998 (42%) of whom were mechanically ventilated and 675 (3%) of whom were identified as having any VAE. Any VAE was associated with an adjusted increased risk of death (OR, 1.91; 95% CI, 1.53-2.37; P < .0001). The automated algorithm was reliable (sensitivity of 93.5%, 95% CI, 77.2%-98.8%; specificity of 100%, 95% CI, 98.8%-100% vs a human abstractor). Comparison of patients with a VAE and with the former VAP definition yielded little agreement (κ = 0.06).nnnCONCLUSIONSnA fully automated method of identifying VAEs is efficient and reliable within a single institution. Although VAEs are strongly associated with worse patient outcomes, additional research is required to evaluate whether and which interventions can successfully prevent VAEs.

Management of Sepsis and Septic Shock

of the Clinical Problem Sepsis results when the body’s response to infection causes lifethreatening organ dysfunction. Septic shock is sepsis that results in tissue hypoperfusion, with vasopressor-requiring hypotension and elevated lactate levels.1 Sepsis is a leading cause of death, morbidity, and expense, contributing to one-third to half of deaths of hospitalized patients,2 depending on definitions.3 Management of sepsis is a complicated clinical challenge requiring early recognition and management of infection, hemodynamic issues, and other organ dysfunctions.

Acid suppression therapy does not predispose to Clostridium difficile infection: the case of the potential bias.

Objective An adverse effect of acid-suppression medications on the occurrence of Clostridium difficile infection (CDI) has been a common finding of many, but not all studies. We hypothesized that association between acid-suppression medications and CDI is due to the residual confounding in comparison between patients with infection to those without, predominantly from non-tested and less sick subjects. We aimed to evaluate the effect of acid suppression therapy on incidence of CDI by comparing patients with CDI to two control groups: not tested patients and patients suspected of having CDI, but with a negative test. Methods We conducted a case-control study of adult patients hospitalized in internal medicine department of tertiary teaching hospital between 2005–2010 for at least three days. Controls from each of two groups (negative for CDI and non-tested) were individually matched (1∶1) to cases by primary diagnosis, Charlson comorbidity index, year of hospitalization and gender. Primary outcomes were diagnoses of International Classification of Diseases (ICD-9)–coded CDI occurring 72 hours or more after admission. Results Patients with CDI were similar to controls with a negative test, while controls without CDI testing had lower clinical severity. In multivariable analysis, treatment by acid suppression medications was associated with CDI compared to those who were not tested (ORu200a=u200a1.88, p-valueu200a=u200a0.032). Conversely, use of acid suppression medications in those who tested negative for the infection was not associated with CDI risk as compared to the cases (ORu200a=u200a0.66; pu200a=u200a0.059). Conclusions These findings suggest that the reported epidemiologic associations between use of acid suppression medications and CDI risk may be spurious. The control group choice has an important impact on the results. Clinical differences between the patients with CDI and those not tested and not suspected of having the infection may explain the different conclusions regarding the acid suppression effect on CDI risk.

Scalable and accurate deep learning with electronic health records

Predictive modeling with electronic health record (EHR) data is anticipated to drive personalized medicine and improve healthcare quality. Constructing predictive statistical models typically requires extraction of curated predictor variables from normalized EHR data, a labor-intensive process that discards the vast majority of information in each patient’s record. We propose a representation of patients’ entire raw EHR records based on the Fast Healthcare Interoperability Resources (FHIR) format. We demonstrate that deep learning methods using this representation are capable of accurately predicting multiple medical events from multiple centers without site-specific data harmonization. We validated our approach using de-identified EHR data from two US academic medical centers with 216,221 adult patients hospitalized for at least 24u2009h. In the sequential format we propose, this volume of EHR data unrolled into a total of 46,864,534,945 data points, including clinical notes. Deep learning models achieved high accuracy for tasks such as predicting: in-hospital mortality (area under the receiver operator curve [AUROC] across sites 0.93–0.94), 30-day unplanned readmission (AUROC 0.75–0.76), prolonged length of stay (AUROC 0.85–0.86), and all of a patient’s final discharge diagnoses (frequency-weighted AUROC 0.90). These models outperformed traditional, clinically-used predictive models in all cases. We believe that this approach can be used to create accurate and scalable predictions for a variety of clinical scenarios. In a case study of a particular prediction, we demonstrate that neural networks can be used to identify relevant information from the patient’s chart.Artificial intelligence: Algorithm predicts clinical outcomes for hospital inpatientsArtificial intelligence outperforms traditional statistical models at predicting a range of clinical outcomes from a patient’s entire raw electronic health record (EHR). A team led by Alvin Rajkomar and Eyal Oren from Google in Mountain View, California, USA, developed a data processing pipeline for transforming EHR files into a standardized format. They then applied deep learning models to data from 216,221 adult patients hospitalized for at least 24u2009h each at two academic medical centers, and showed that their algorithm could accurately predict risk of mortality, hospital readmission, prolonged hospital stay and discharge diagnosis. In all cases, the method proved more accurate than previously published models. The authors provide a case study to serve as a proof-of-concept of how such an algorithm could be used in routine clinical practice in the future.

An Emergency Department Validation of the SEP-3 Sepsis and Septic Shock Definitions and Comparison With 1992 Consensus Definitions

Study objective: The Third International Consensus Definitions Task Force (SEP‐3) proposed revised criteria defining sepsis and septic shock. We seek to evaluate the performance of the SEP‐3 definitions for prediction of inhospital mortality in an emergency department (ED) population and compare the performance of the SEP‐3 definitions to that of the previous definitions. Methods: This was a secondary analysis of 3 prospectively collected, observational cohorts of infected ED subjects aged 18 years or older. The primary outcome was all‐cause inhospital mortality. In accordance with the SEP‐3 definitions, we calculated test characteristics of sepsis (quick Sequential Organ Failure Assessment [qSOFA] score ≥2) and septic shock (vasopressor dependence plus lactate level >2.0 mmol/L) for mortality and compared them to the original 1992 consensus definitions. Results: We identified 7,754 ED patients with suspected infection overall; 117 had no documented mental status evaluation, leaving 7,637 patients included in the analysis. The mortality rate for the overall population was 4.4% (95% confidence interval [CI] 3.9% to 4.9%). The mortality rate for patients with qSOFA score greater than or equal to 2 was 14.2% (95% CI 12.2% to 16.2%), with a sensitivity of 52% (95% CI 46% to 57%) and specificity of 86% (95% CI 85% to 87%) to predict mortality. The original systemic inflammatory response syndrome–based 1992 consensus sepsis definition had a 6.8% (95% CI 6.0% to 7.7%) mortality rate, sensitivity of 83% (95% CI 79% to 87%), and specificity of 50% (95% CI 49% to 51%). The SEP‐3 septic shock mortality was 23% (95% CI 16% to 30%), with a sensitivity of 12% (95% CI 11% to 13%) and specificity of 98.4% (95% CI 98.1% to 98.7%). The original 1992 septic shock definition had a 22% (95% CI 17% to 27%) mortality rate, sensitivity of 23% (95% CI 18% to 28%), and specificity of 96.6% (95% CI 96.2% to 97.0%). Conclusion: Both the new SEP‐3 and original sepsis definitions stratify ED patients at risk for mortality, albeit with differing performances. In terms of mortality prediction, the SEP‐3 definitions had improved specificity, but at the cost of sensitivity. Use of either approach requires a clearly intended target: more sensitivity versus specificity.

Antibiotic and Duration of Perioperative Prophylaxis Predicts Surgical Site Infection in Head and Neck Surgery

Objective To examine the effect of giving antibiotics on the day of surgery (DOS) vs DOS and first postoperative day (DOS+1) for prophylaxis against surgical site infection (SSI) in clean-contaminated head and neck surgery (CCHNS). Study Design Retrospective multi-institution analysis using University HealthSystem Consortium data. Methods A multivariate logistic regression model of 8836 discharge records from patients undergoing CCHNS was used to determine the odds of SSI for antibiotic agent/duration combinations. Setting Ninety-two academic and affiliated medical centers from 2008 to 2011. Results Ampicillin/sulbactam, clindamycin, cefazolin + metronidazole, and cefazolin alone were the most common antibiotics. For patients receiving antibiotics only on DOS, there was no significant difference in odds of SSI based on antibiotic choice. When given on the DOS and DOS+1, patients receiving ampicillin/sulbactam had a reduction in odds of SSI by over two-thirds (odds ratio [OR], 0.28 [95% confidence interval, 0.13-0.61], P = .001, compared with ampicillin/sulbactam on DOS only), whereas this effect was not seen with clindamycin (1.82 [0.93-3.56], P = .078, compared with clindamycin on DOS only). Prolonging clindamycin beyond the DOS was associated with a higher odds of SSI compared with DOS-only ampicillin/sulbactam (OR, 2.66; 95% CI, 1.33-5.30; P = .006). These relationships held in a subset of physicians and hospitals that used multiple different regimens. DOS+1 regimens were not associated with an increased odds of antibiotic-induced complications. Conclusion Prolonging ampicillin/sulbactam beyond the day of surgery may have a protective effect against SSI, and 1 or more days of ampicillin/sulbactam may be preferable to multiple days of clindamycin. New randomized trials are needed to define the ideal regimen for CCHNS.

Antipsychotic utilization in the intensive care unit and in transitions of care.

PURPOSEnThe objective of this study was to quantify the rate at which newly initiated antipsychotic therapy is continued on discharge from the intensive care unit (ICU) and describe risk factors for continuation post-ICU discharge.nnnMATERIALS AND METHODSnThis is a retrospective cohort study of all patients receiving an antipsychotic in the ICUs of a large academic medical center from January 1, 2005, to October 31, 2011. Medical record review was conducted to ascertain whether a patient was newly started on antipsychotic therapy and whether therapy was continued post-ICU discharge.nnnRESULTSnA total of 39,248 ICU admissions over the 7-year period were evaluated. Of these, 4468 (11%) were exposed to antipsychotic therapy, of which 3119 (8%) were newly initiated. In the newly initiated cohort, 642 (21%) were continued on therapy on discharge from the hospital. Type of drug (use of quetiapine vs no use of quetiapine: odds ratio, 3.2; 95% confidence interval, 2.5-4.0; P < .0001 and use of olanzapine: odds ratio, 2.4, 95% confidence interval, 2.0-3.1; P ≤ .0001) was a significant risk factor for continuing antipsychotics on discharge despite adjustment for clinical factors.nnnCONCLUSIONSnAntipsychotic use is common in the ICU setting, and a significant number of newly initiated patients have therapy continued upon discharge from the hospital.

Association Between In-Hospital Critical Illness Events and Outcomes in Patients on the Same Ward

Association Between In-Hospital Critical Illness Events and Outcomes in Patients on the Same Ward Major critical illness events, such as cardiopulmonary arrest and intensive care unit (ICU) transfer, disrupt workflow in a hospital ward. Other patients on the same ward may receive inadequate attention, especially if their care team is distracted by the emergency. Most studies have concentrated on patient-level variables associated with outcomes.1-3 To our knowledge, no study has quantified the risk to ward occupants associated with patients on the same ward experiencing critical illness.