A Simulated Prospective Evaluation of a Deep Learning Model for Real-Time Prediction of Clinical Deterioration Among Ward Patients.

OBJECTIVES: The National Early Warning Score, Modified Early Warning Score, and quick Sepsis-related Organ Failure Assessment can predict clinical deterioration. These scores exhibit only moderate performance and are often evaluated using aggregated measures over time. A simulated prospective validation strategy that assesses multiple predictions per patient-day would provide the best pragmatic evaluation. We developed a deep recurrent neural network deterioration model and conducted a simulated prospective evaluation. DESIGN: Retrospective cohort study. SETTING: Four hospitals in Pennsylvania. PATIENTS: Inpatient adults discharged between July 1, 2017, and June 30, 2019. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We trained a deep recurrent neural network and logistic regression model using data from electronic health records to predict hourly the 24-hour composite outcome of transfer to ICU or death. We analyzed 146,446 hospitalizations with 16.75 million patient-hours. The hourly event rate was 1.6% (12,842 transfers or deaths, corresponding to 260,295 patient-hours within the predictive horizon). On a hold-out dataset, the deep recurrent neural network achieved an area under the precision-recall curve of 0.042 (95% CI, 0.04-0.043), comparable with logistic regression model (0.043; 95% CI 0.041 to 0.045), and outperformed National Early Warning Score (0.034; 95% CI, 0.032-0.035), Modified Early Warning Score (0.028; 95% CI, 0.027- 0.03), and quick Sepsis-related Organ Failure Assessment (0.021; 95% CI, 0.021-0.022). For a fixed sensitivity of 50%, the deep recurrent neural network achieved a positive predictive value of 3.4% (95% CI, 3.4-3.5) and outperformed logistic regression model (3.1%; 95% CI 3.1-3.2), National Early Warning Score (2.0%; 95% CI, 2.0-2.0), Modified Early Warning Score (1.5%; 95% CI, 1.5-1.5), and quick Sepsis-related Organ Failure Assessment (1.5%; 95% CI, 1.5-1.5). CONCLUSIONS: Commonly used early warning scores for clinical decompensation, along with a logistic regression model and a deep recurrent neural network model, show very poor performance characteristics when assessed using a simulated prospective validation. None of these models may be suitable for real-time deployment.

A Machine Learning Algorithm to Predict Severe Sepsis and Septic Shock: Development, Implementation, and Impact on Clinical Practice.

OBJECTIVES: Develop and implement a machine learning algorithm to predict severe sepsis and septic shock and evaluate the impact on clinical practice and patient outcomes. DESIGN: Retrospective cohort for algorithm derivation and validation, pre-post impact evaluation. SETTING: Tertiary teaching hospital system in Philadelphia, PA. PATIENTS: All non-ICU admissions; algorithm derivation July 2011 to June 2014 (n = 162,212); algorithm validation October to December 2015 (n = 10,448); silent versus alert comparison January 2016 to February 2017 (silent n = 22,280; alert n = 32,184). INTERVENTIONS: A random-forest classifier, derived and validated using electronic health record data, was deployed both silently and later with an alert to notify clinical teams of sepsis prediction. MEASUREMENT AND MAIN RESULT: Patients identified for training the algorithm were required to have International Classification of Diseases, 9th Edition codes for severe sepsis or septic shock and a positive blood culture during their hospital encounter with either a lactate greater than 2.2 mmol/L or a systolic blood pressure less than 90 mm Hg. The algorithm demonstrated a sensitivity of 26% and specificity of 98%, with a positive predictive value of 29% and positive likelihood ratio of 13. The alert resulted in a small statistically significant increase in lactate testing and IV fluid administration. There was no significant difference in mortality, discharge disposition, or transfer to ICU, although there was a reduction in time-to-ICU transfer. CONCLUSIONS: Our machine learning algorithm can predict, with low sensitivity but high specificity, the impending occurrence of severe sepsis and septic shock. Algorithm-generated predictive alerts modestly impacted clinical measures. Next steps include describing clinical perception of this tool and optimizing algorithm design and delivery.

Essential Resources and Strategies for Antibiotic Stewardship Programs in the Acute Care Setting.

Background: Antibiotic stewardship programs improve clinical outcomes and patient safety and help combat antibiotic resistance. Specific guidance on resources needed to structure stewardship programs is lacking. This manuscript describes results of a survey of US stewardship programs and resultant recommendations regarding potential staffing structures in the acute care setting. Methods: A cross-sectional survey of members of 3 infectious diseases subspecialty societies actively involved in antibiotic stewardship was conducted. Survey responses were analyzed with descriptive statistics. Logistic regression models were used to investigate the relationship between stewardship program staffing levels and self-reported effectiveness and to determine which strategies mediate effectiveness. Results: Two-hundred forty-four respondents from a variety of acute care settings completed the survey. Prior authorization for select antibiotics, antibiotic reviews with prospective audit and feedback, and guideline development were common strategies. Eighty-five percent of surveyed programs demonstrated effectiveness in at least 1 outcome in the prior 2 years. Each 0.50 increase in pharmacist and physician full-time equivalent (FTE) support predicted a 1.48-fold increase in the odds of demonstrating effectiveness. The effect was mediated by the ability to perform prospective audit and feedback. Most programs noted significant barriers to success. Conclusions: Based on our survey's results, we propose an FTE-to-bed ratio that can be used as a starting point to guide discussions regarding necessary resources for antibiotic stewardship programs with executive leadership. Prospective audit and feedback should be the cornerstone of stewardship programs, and both physician leadership and pharmacists with expertise in stewardship are crucial for success.

Implementation of a systematic culturing program to monitor the efficacy of endoscope reprocessing: outcomes and costs.

BACKGROUND AND AIMS: In 2015, the U.S. Food and Drug Administration and Centers for Disease Control and Prevention (CDC) issued guidance for duodenoscope culturing and reprocessing in response to outbreaks of carbapenem-resistant Enterobacteriaceae (CRE) duodenoscope-related infections. Based on this guidance, we implemented best practices for reprocessing and developed a systematic process for culturing endoscopes with elevator levers. The aim of this study is to report the outcomes and direct costs of this program. METHODS: First, clinical microbiology data from 2011 to 2014 were reviewed retrospectively to assess for possible elevator lever-equipped endoscope-related CRE infections. Second, a program to systematically culture elevator lever-equipped endoscopes was implemented. Each week, about 25% of the inventory of elevator lever-equipped endoscopes is cultured based on the CDC guidelines. If any cultures return bacterial growth, the endoscope is quarantined pending repeat culturing. The costs of the program, including staff time and supplies, have been calculated. RESULTS: From 2011 to 2014, none of 17 patients with documented CRE infection had undergone ERCP or endoscopic ultrasound in the previous 36 months. From June 2015 to September 2016, 285 cultures were performed. Three (1.1%) had bacterial growth, 2 with skin contaminants and 1 with an oral contaminant. The associated endoscopes were quarantined and reprocessed, and repeat cultures were negative. The total estimated cost of our program for an inventory of 20 elevator lever-equipped endoscopes was $30,429.60 per year ($1521.48 per endoscope). CONCLUSIONS: This 16-month evaluation of a systematic endoscope culturing program identified a low rate of positive cultures after elevator lever endoscope reprocessing. All positive cultures were with non-enteric microorganisms. The program was of modest cost and identified reprocessing procedures that may have led to a low rate of positive cultures.

Sepsis National Hospital Inpatient Quality Measure (SEP-1): Multistakeholder Work Group Recommendations for Appropriate Antibiotics for the Treatment of Sepsis.

The Center for Medicare and Medicaid Services adopted the Early Management Bundle, Severe Sepsis/Septic Shock (SEP-1) performance measure to the Hospital Inpatient Quality Reporting Program in July 2015 to help address the high mortality and high cost associated with sepsis. The SEP-1 performance measure requires, among other critical interventions, timely administration of antibiotics to patients with sepsis or septic shock. The multistakeholder workgroup recognizes the need for SEP-1 but strongly believes that multiple antibiotics listed in the antibiotic tables for SEP-1 are not appropriate and the use of these antibiotics, as called for in the SEP-1 measure, is not in alignment with prudent antimicrobial stewardship. To promote the appropriate use of antimicrobials and combat antimicrobial resistance, the workgroup provides recommendations for appropriate antibiotics for the treatment of sepsis.

Executive Summary: Implementing an Antibiotic Stewardship Program: Guidelines by the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America.

Evidence-based guidelines for implementation and measurement of antibiotic stewardship interventions in inpatient populations including long-term care were prepared by a multidisciplinary expert panel of the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America. The panel included clinicians and investigators representing internal medicine, emergency medicine, microbiology, critical care, surgery, epidemiology, pharmacy, and adult and pediatric infectious diseases specialties. These recommendations address the best approaches for antibiotic stewardship programs to influence the optimal use of antibiotics.

Implementing an Antibiotic Stewardship Program: Guidelines by the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America.

Evidence-based guidelines for implementation and measurement of antibiotic stewardship interventions in inpatient populations including long-term care were prepared by a multidisciplinary expert panel of the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America. The panel included clinicians and investigators representing internal medicine, emergency medicine, microbiology, critical care, surgery, epidemiology, pharmacy, and adult and pediatric infectious diseases specialties. These recommendations address the best approaches for antibiotic stewardship programs to influence the optimal use of antibiotics.

Duration of Colonization and Determinants of Earlier Clearance of Colonization With Methicillin-Resistant Staphylococcus aureus.

BACKGROUND: The duration of colonization and factors associated with clearance of methicillin-resistant Staphylococcus aureus (MRSA) after community-onset MRSA skin and soft-tissue infection (SSTI) remain unclear. METHODS: We conducted a prospective cohort study of patients with acute MRSA SSTI presenting to 5 adult and pediatric academic hospitals from 1 January 2010 through 31 December 2012. Index patients and household members performed self-sampling for MRSA colonization every 2 weeks for 6 months. Clearance of colonization was defined as negative MRSA surveillance cultures during 2 consecutive sampling periods. A Cox proportional hazards regression model was developed to identify determinants of clearance of colonization. RESULTS: Two hundred forty-three index patients were included. The median duration of MRSA colonization after SSTI diagnosis was 21 days (95% confidence interval [CI], 19-24), and 19.8% never cleared colonization. Treatment of the SSTI with clindamycin was associated with earlier clearance (hazard ratio [HR], 1.72; 95% CI, 1.28-2.30; P < .001). Older age (HR, 0.99; 95% CI, .98-1.00; P = .01) was associated with longer duration of colonization. There was a borderline significant association between increased number of household members colonized with MRSA and later clearance of colonization in the index patient (HR, 0.85; 95% CI, .71-1.01; P = .06). CONCLUSIONS: With a systematic, regular sampling protocol, duration of MRSA colonization was noted to be shorter than previously reported, although 19.8% of patients remained colonized at 6 months. The association between clindamycin and shorter duration of colonization after MRSA SSTI suggests a possible role for the antibiotic selected for treatment of MRSA infection.

Public reporting of health care-associated surveillance data: recommendations from the healthcare infection control practices advisory committee.

Health care-associated infection (HAI) rates are used as measures of a health care facility's quality of patient care. Recently, these outcomes have been used to publicly rank quality efforts and determine facility reimbursement. The value of comparing HAI rates among health care facilities is limited by many factors inherent to HAI surveillance, and incentives that reward low HAI rates can lead to unintended consequences that can compromise medical care surveillance efforts, such as the use of clinical adjudication panels to veto events that meet HAI surveillance definitions.The Healthcare Infection Control Practices Advisory Committee, a federal advisory committee that provides advice and guidance to the Centers for Disease Control and Prevention (CDC) and the Secretary of the Department of Health and Human Services about strategies for surveillance, prevention, and control of HAIs, assessed the challenges associated with using HAI surveillance data for external quality reporting, including the unintended consequences of clinician veto and clinical adjudication panels. Discussions with stakeholder liaisons and committee members were then used to formulate recommended standards for the use of HAI surveillance data for external facility assessment to ensure valid comparisons and to provide as level a playing field as possible.The final recommendations advocate for consistent, objective, and independent application of CDC HAI definitions with concomitant validation of HAIs and surveillance processes. The use of clinician veto and adjudication is discouraged.

Introduction to A Compendium of Strategies to Prevent Healthcare-Associated Infections In Acute-Care Hospitals: 2022 Updates.

Since the initial publication of A Compendium of Strategies to Prevent Healthcare-Associated Infections in Acute Care Hospitals in 2008, the prevention of healthcare-associated infections (HAIs) has continued to be a national priority. Progress in healthcare epidemiology, infection prevention, antimicrobial stewardship, and implementation science research has led to improvements in our understanding of effective strategies for HAI prevention. Despite these advances, HAIs continue to affect ∼1 of every 31 hospitalized patients, leading to substantial morbidity, mortality, and excess healthcare expenditures, and persistent gaps remain between what is recommended and what is practiced.The widespread impact of the coronavirus disease 2019 (COVID-19) pandemic on HAI outcomes in acute-care hospitals has further highlighted the essential role of infection prevention programs and the critical importance of prioritizing efforts that can be sustained even in the face of resource requirements from COVID-19 and future infectious diseases crises.The Compendium: 2022 Updates document provides acute-care hospitals with up-to-date, practical expert guidance to assist in prioritizing and implementing HAI prevention efforts. It is the product of a highly collaborative effort led by the Society for Healthcare Epidemiology of America (SHEA), the Infectious Disease Society of America (IDSA), the Association for Professionals in Infection Control and Epidemiology (APIC), the American Hospital Association (AHA), and The Joint Commission, with major contributions from representatives of organizations and societies with content expertise, including the Centers for Disease Control and Prevention (CDC), the Pediatric Infectious Disease Society (PIDS), the Society for Critical Care Medicine (SCCM), the Society for Hospital Medicine (SHM), the Surgical Infection Society (SIS), and others.

Impact of antimicrobial stewardship programme changes on unnecessary double anaerobic coverage therapy.

BACKGROUND: Concern has been raised over the practice of unnecessary double anaerobic coverage therapy (DACT) in the hospital setting. However, the incidence of and risk factors for unnecessary DACT are not well studied. On 8 September 2008, the antimicrobial stewardship programme (ASP) at our institution was modified such that several antibiotics, including ampicillin/sulbactam and metronidazole, no longer required pre-approval. We anticipated that this change would increase both unnecessary DACT and target antibiotic consumption. METHODS: A nested case-control study was conducted to determine the cumulative incidence of and risk factors for unnecessary DACT. Cases were subjects who received unnecessary DACT while controls were subjects who did not receive DACT or who received necessary DACT. Segmented regression analysis was subsequently performed to evaluate the impact of ASP changes on unnecessary DACT and consumption of target antibiotics. RESULTS: From October 2007 to September 2009, the cumulative incidence of unnecessary DACT was 2.3% [95% confidence interval (CI) 1.7-3.1]. Independent risk factors for unnecessary DACT [adjusted odds ratio (95% CI); P value] included hospitalization on a surgical ward [3.51 (1.03-12.02); P = 0.002], hospitalization on an obstetrics and gynaecology ward [9.07 (2.54-32.40); P = 0.002] and underlying metastatic malignancy [3.18 (1.38-7.09); P = 0.006]. The ASP change was associated with an increase in ampicillin/sulbactam and metronidazole consumption. However, there was no significant impact on unnecessary DACT prescribing. CONCLUSIONS: Although uncommon, unnecessary DACT is more prevalent in specific services. Future qualitative studies focusing on these specific subgroups would be useful in elucidating this problem more clearly. The ASP changes were not associated with increases in unnecessary DACT.

Preparedness Tested: Severe Cerebral Malaria Presenting as a High-Risk Person Under Investigation for Ebola Virus Disease at a US Hospital.

In 2019, a 42-year-old African man who works as an Ebola virus disease (EVD) researcher traveled from the Democratic Republic of Congo (DRC), near an ongoing EVD epidemic, to Philadelphia and presented to the Hospital of the University of Pennsylvania Emergency Department with altered mental status, vomiting, diarrhea, and fever. He was classified as a "wet" person under investigation for EVD, and his arrival activated our hospital emergency management command center and bioresponse teams. He was found to be in septic shock with multisystem organ dysfunction, including circulatory dysfunction, encephalopathy, metabolic lactic acidosis, acute kidney injury, acute liver injury, and diffuse intravascular coagulation. Critical care was delivered within high-risk pathogen isolation in the ED and in our Special Treatment Unit until a diagnosis of severe cerebral malaria was confirmed and EVD was definitively excluded.This report discusses our experience activating a longitudinal preparedness program designed for rare, resource-intensive events at hospitals physically remote from any active epidemic but serving a high-volume international air travel port-of-entry.

The prevalence of fluoroquinolone resistance mechanisms in colonizing Escherichia coli isolates recovered from hospitalized patients.

BACKGROUND: Fluoroquinolones are the most commonly prescribed antimicrobials. The epidemiology of fecal colonization with Escherichia coli demonstrating reduced susceptibility to fluoroquinolones remains unclear. METHODS: During a 3-year period (15 September 2004 through 19 October 2007), all patients hospitalized for >3 days were approached for fecal sampling. All E. coli isolates with reduced susceptibility to fluoroquinolones (minimum inhibitory concentration [MIC] of levofloxacin, 0.125 microg/mL) were identified. We characterized gyrA and parC mutations and organic solvent tolerance. Isolates were compared using pulsed-field gel electrophoresis. RESULTS: Of 353 patients colonized with E. coli demonstrating reduced fluoroquinolone susceptibility, 300 (85.0%) had 1 gyrA mutation, 161 (45.6%) had 1 parC mutation, and 171 (48.6%) demonstrated organic solvent tolerance. The mean numbers of total mutations (ie, gyrA and parC) for E. coli isolates with a levofloxacin MIC of 8 microg/mL versus <8.0 microg/mL were 2.70 and 0.82 (P < .001). Of the 136 E. coli isolates with a levofloxacin MIC of 8 microg/mL, 90 (66.2%) demonstrated a nalidixic acid MIC of 16 microg/mL. Significant differences were found over time in the proportion of E. coli isolates demonstrating gyrA mutation, parC mutation, and organic solvent tolerance. There was little evidence of clonal spread of isolates. Conclusions. Gastrointestinal tract colonization with E. coli demonstrating reduced susceptibility to levofloxacin is common. Although 40% of study isolates exhibited a levofloxacin MIC of <8 microg/mL (and would thus be missed by current Clinical and Laboratory Standards Institute breakpoints), nalidixic acid resistance may be a useful marker for detection of such isolates. Significant temporal changes occurred in the proportion of isolates exhibiting various resistance mechanisms.

Clinical and microbiological outcomes of serious infections with multidrug-resistant gram-negative organisms treated with tigecycline.

Eighteen patients received tigecycline as treatment for infection due to multidrug-resistant gram-negative bacilli, including Acinetobacter baumannii and Klebsiella pneumoniae carbapenemase- and extended-spectrum beta-lactamase-producing Enterobacteriaceae. Pretherapy minimum inhibitory concentration values for tigecycline predicted clinical success. Observed evolution of resistance during therapy raises concern about routine use of tigecycline in treatment of such infections when other therapies are available.

Treatment and outcomes for patients with bacteremic pneumococcal pneumonia.

Delayed time to antibiotic administration has been linked with higher mortality for patients with community-acquired pneumonia, but the impact of antibiotic resistance on clinical outcomes has been controversial. In the current study we assess the combined impact of antibiotic resistance and antibiotic timing on outcomes, including inhospital mortality, complications, length of stay, and time to stability, for patients hospitalized with community-acquired bacteremic pneumococcal pneumonia. We conducted a retrospective cohort study in 43 hospitals in the Southeastern Pennsylvania region from 2001 to 2004. Eligible adult patients had pneumococcal bacteremia and radiographic evidence of pneumonia. Outcomes were assessed based on medical record review. Multivariable regression was used to adjust for severity of illness and sequentially assess the impact of antibiotic resistance and time to active antibiotic therapy. The overall inhospital mortality was 10%. Overall, levels of macrolide, cephalosporin, and fluoroquinolone resistance were low and did not adversely impact the time to administration of active antibiotic therapy. Receipt of at least 1 active antibiotic within 4 hours was associated with reduced mortality (odds ratio [OR], 0.47; 95% confidence interval [CI], 0.2-1.0) and shortened length of stay (OR, 0.77; CI, 0.6-1.0) but did not reduce the risk of other adverse outcomes. We conclude that early antibiotic administration reduces the risks of mortality in patients with bacteremic pneumococcal pneumonia. Current patterns of drug resistance did not lead to delays in administration of active antimicrobial therapy.

Contact isolation for infection control in hospitalized patients: is patient satisfaction affected?

The effects of contact isolation on patient satisfaction are unknown. We performed a cross-sectional survey and found that most patients lack education and knowledge regarding isolation but feel that it improves their care. In multivariable analysis, isolated patients were not less satisfied with inpatient care than were nonisolated patients.

Patient attitudes regarding participation in studies of antimicrobial resistance.

BACKGROUND: A number of recent studies of antimicrobial resistance have focused on the role of antimicrobial-resistant pathogens that colonize the gastrointestinal tract. However, participation rates have been low in studies that involve fecal sampling. Attitudes toward such studies among potential study participants have not been assessed. METHODS: We conducted a cross-sectional survey, enrolling 3 groups of inpatients from a large academic center. Group 1 consisted of patients who had previously participated in a cohort study of fluoroquinolone-resistant Escherichia coli, which involved the collection of perirectal swab samples. Group 2 consisted of patients who had previously refused to participate in the study of fluoroquinolone-resistant E. coli. Group 3 consisted of patients who had never been asked to participate in the study of the fluoroquinolone-resistant E. coli. The survey assessed patients' attitudes and beliefs regarding medical research and their willingness to consent to collection of a perirectal swab sample. Response options were recorded on a 5-point Likert scale. The Fisher exact test was used to compare dichotomized responses across study groups. RESULTS: A total of 90 patients were surveyed: there were 29 in group 1 and in group 2 and 32 in group 3. Of 90 patients, 31 (35%) believed researchers might run additional tests on collected samples without informing the patient, whereas 25 (27%) believed persons other than the research team might gain access to study results. The belief that a person could get sicker as a result of a having a perirectal swab sample collected was significantly more common among patients who had previously refused to participate in the fluoroquinolone-resistant E. coli study. CONCLUSION: This study highlights important beliefs and attitudes that are associated with the likelihood of participating in studies of antimicrobial resistance. Explicitly addressing these concerns with eligible patients is critical to optimize participation in future studies.

Effect of accessibility of influenza vaccination on the rate of childcare staff vaccination.

We assessed the impact of free on-site influenza vaccination on childcare staff vaccination prevalence using 2 before-and-after studies. Vaccination was offered during the 2003-2004 and 2006-2007 influenza seasons. Staff vaccination prevalence was higher in each intervention season compared to the prior, nonintervention season. No baseline characteristics were associated with receipt of vaccination.