Syndromic surveillance for influenza-like illness (ILI) is predominantly performed in the outpatient setting. The objective of this study was to compare patterns of ILI activity in outpatient, emergency department (ED), and inpatient settings using an electronic syndromic surveillance algorithm.

Retrospective cohort study over 7.5 years.

A large community health system comprised of 5 hospitals and >50 clinics.

We applied an electronic syndromic surveillance algorithm for ILI to all primary-care outpatient visits, inpatient encounters, and ED encounters at our health system. Comparisons of ILI activity over time were performed using Spearman’s rank correlation coefficient. Cross correlation was used to compare the timing of ILI activity among treatment settings.

Overall, 4,447,769 patient encounters occurred during the study period; 152,607 of these (3.4%) were consistent with ILI. The correlation coefficient for ILI activity in the outpatient versus ED setting was 0.877, and for the outpatient versus inpatient setting, the correlation coefficient was 0.699. ILI activity among outpatients preceded ILI activity among inpatients by 1 week. ILI activity among children in the outpatient setting preceded ILI activity among adults in all 3 settings by 1 week.

Syndromic surveillance for ILI in the outpatient setting yields similar results to surveillance in the ED setting, but it produces less similar results than ILI surveillance in the inpatient setting. ILI activity in the pediatric outpatient population is a potential predictor of future ILI activity in the general population.

Infect Control Hosp Epidemiol 2017;38:393–398

Influenza A (H1N1) pdm09 became the predominant circulating strain in the United States during the 2013–2014 influenza season. Little is known about the epidemiology of severe influenza during this season.

A retrospective cohort study of severely ill patients with influenza infection in intensive care units in 33 US hospitals from September 1, 2013, through April 1, 2014, was conducted to determine risk factors for mortality present on intensive care unit admission and to describe patient characteristics, spectrum of disease, management, and outcomes.

A total of 444 adults and 63 children were admitted to an intensive care unit in a study hospital; 93 adults (20.9%) and 4 children (6.3%) died. By logistic regression analysis, the following factors were significantly associated with mortality among adult patients: older age (>65 years, odds ratio, 3.1 [95% CI, 1.4–6.9], P=.006 and 50–64 years, 2.5 [1.3–4.9], P=.007; reference age 18–49 years), male sex (1.9 [1.1–3.3], P=.031), history of malignant tumor with chemotherapy administered within the prior 6 months (12.1 [3.9–37.0], P<.001), and a higher Sequential Organ Failure Assessment score (for each increase by 1 in score, 1.3 [1.2–1.4], P<.001).

Risk factors for death among US patients with severe influenza during the 2013–2014 season, when influenza A (H1N1) pdm09 was the predominant circulating strain type, shifted in the first postpandemic season in which it predominated toward those of a more typical epidemic influenza season.

Infect. Control Hosp. Epidemiol. 2015;36(11):1251–1260

To identify predictors of community-onset extended-spectrum β-lactamase (ESBL)-producing Escherichia coli infection.

Prospective case-control study.

Acute care hospitals and ambulatory clinics in the Chicago, Illinois, region.

Adults with E. coli clinical isolates cultured in ambulatory settings or within 48 hours of hospital admission.

Cases were patients with ESBL-producing E. coli clinical isolates cultured in ambulatory settings or within 48 hours of admission, and controls were patients with non-ESBL-producing E. coli isolates, matched to cases by specimen, location, and date. Clinical variables were ascertained through interviews and medical record review. Molecular methods were used to identify ESBL types, sequence type ST131, and aac(6′)-Ib-cr.

We enrolled 94 cases and 158 controls. Multivariate risk factors for ESBL-producing E. coli infection included travel to India in the past year (odds ratio [OR], 14.40 [95% confidence interval (CI), 2.92-70.95]), ciprofloxacin use (OR, 3.92 [95% CI, 1.90-8.1]), and age (OR, 1.04 [95% CI, 1.02-1.06]). Case isolates exhibited high prevalence of CTX-M-15 (78%), ST131 (50%), and aac(6′)-Ib-cr (66% of isolates with CTX-M-15).

Providers should be aware of the increased risk of ESBL-producing E. coli infection among returned travelers, especially those from India.

A major challenge in treating Clostridium difficile infection (CDI) is relapse. Many new therapies are being developed to help prevent this outcome. We sought to establish risk factors for relapse and determine whether fields available in an electronic health record (EHR) could be used to identify high-risk patients for targeted relapse prevention strategies.

Retrospective cohort study.

Large clinical data warehouse at a 4-hospital healthcare organization.

Data were gathered from January 2006 through October 2010. Subjects were all inpatient episodes of a positive C. difficile test where patients were available for 56 days of follow-up.

Relapse was defined as another positive test between 15 and 56 days after the initial test. Multivariable regression was performed to identify factors independently associated with CDI relapse.

Eight hundred twenty-nine episodes met eligibility criteria, and 198 resulted in relapse (23.9%). In the final multivariable analysis, risk of relapse was associated with age (odds ratio [OR], 1.02 per year [95% confidence interval (CI), 1.01–1.03]), fluoroquinolone exposure in the 90 days before diagnosis (OR, 1.58 [95% CI, 1.11–2.26]), intensive care unit stay in the 30 days before diagnosis (OR, 0.47 [95% CI, 0.30–0.75]), cephalosporin (OR, 1.80 [95% CI, 1.19–2.71]), proton pump inhibitor (PPI; OR, 1.55 [95% CI, 1.05–2.29]), and metronidazole exposure after diagnosis (OR, 2.74 [95% CI, 1.64–4.60]). A prediction model tuned to ensure a 50% probability of relapse would flag 14.6% of CDI episodes.

Data from a comprehensive EHR can be used to identify patients at high risk for CDI relapse. Major risk factors include antibiotic and PPI exposure.

Interventions for reducing methicillin-resistant Staphylococcus aureus (MRSA) healthcare-associated disease require outcome assessment; this is typically done by manual chart review to determine infection, which can be labor intensive. The purpose of this study was to validate electronic tools for MRSA healthcare-associated infection (HAI) trending that can replace manual medical record review.

This was an observational study comparing manual medical record review with 3 electronic methods: raw culture data from the laboratory information system (LIS) in use by our healthcare organization, LIS data combined with admission-discharge-transfer (ADT) data to determine which cultures were healthcare associated (LIS + ADT), and the CareFusion MedMined Nosocomial Infection Marker (NIM). Each method was used for the same 7-year period from August 2003 through July 2010.

The data set was from a 3-hospital organization covering 342,492 admissions.

Correlation coefficients for raw LIS, LIS + ADT, and NIM were 0.976, 0.957, and 0.953, respectively, when assessed on an annual basis. Quarterly performance for disease trending was also good, with R2 values exceeding 0.7 for all methods.

The electronic tools accurately identified trends in MRSA HAI incidence density when all infections were combined as quarterly or annual data; the performance is excellent when annual assessment is done. These electronic surveillance systems can significantly reduce (93% [in-house-developed program] to more than 99.9999% [commercially available systems]) the personnel resources needed to monitor the impact of a disease control program.

Healthcare providers need a better empiric antibiotic prescribing aid than the traditional antibiogram, which supplies no information on the relative frequency of organisms recovered in a given infection and which is uninformative in situations where multiple antimicrobials are used or multiple organisms are anticipated. We aimed to develop and demonstrate a novel empiric prescribing decision aid.

This is a demonstration involving more than 9,000 unique encounters for abdominal-biliary infection (ABI) and urinary tract infection (UTI) to a large healthcare system with a fully integrated electronic health record (EHR).

We developed a novel method of displaying microbiology data called the weighted-incidence syndromic combination antibiogram (WISCA) for 2 clinical syndromes, ABI and UTI. The WISCA combines simple diagnosis and microbiology data from the EHR to (1) classify patients by syndrome and (2) determine, for each patient with a given syndrome, whether a given regimen (1 or more agents) would have covered all the organisms recovered for their infection. This allows data to be presented such that clinicians can see the probability that a particular regimen will cover a particular infection rather than the probability that a single drug will cover a single organism.

There were 997 encounters for ABI and 8,232 for UTI. A WISCA was created for each syndrome and compared with a traditional antibiogram for the same period.

Novel approaches to data compilation and display can overcome limitations to the utility of the traditional antibiogram in helping providers choose empiric antibiotics.

To describe the identification, management, and clinical characteristics of hospitalized patients with influenza-like illness (ILI) during the peak period of activity of the 2009 pandemic strain of influenza A virus subtype H1N1 (2009 H1N1).

Retrospective review of electronic medical records.

Hospitalized patients who presented to the emergency department during the period October 18 through November 14, 2009, at 4 hospitals in Cook County, Illinois, with the capacity to perform real-time reverse-transcriptase polymerase chain reaction testing for influenza.

Vital signs and notes recorded within 1 calendar day after emergency department arrival were reviewed for signs and symptoms consistent with ILI. Cases of ILI were classified as recognized by healthcare providers if an influenza test was performed or if influenza was mentioned as a possible diagnosis in the physician notes. Logistic regression was used to determine the patient attributes and symptoms that were associated with ILI recognition and with influenza infection.

We identified 460 ILI case patients, of whom 412 (90%) had ILI recognized by healthcare providers, 389 (85%) were placed under airborne or droplet isolation precautions, and 243 (53%) were treated with antiviral medication. Of 401 ILI case patients tested for influenza, 91 (23%) had a positive result. Fourteen (3%) ILI case patients and none of the case patients who tested positive for influenza had sore throat in the absence of cough.

Healthcare providers identified a high proportion of hospitalized ILI case patients. Further improvements in disease detection can be made through the use of advanced electronic health records and efficient diagnostic tests. Future studies should evaluate the inclusion of sore throat in the ILI case definition.

To evaluate two different methods of measuring catheter-associated urinary tract infection (CAUTI) rates in the setting of a quality improvement initiative aimed at reducing device utilization.

Comparison of CAUTI measurements in the context of a before-after trial of acute care adult admissions to a multicentered healthcare system.

CAUTIs were identified with an automated surveillance system, and device-days were measured through an electronic health record. Traditional surveillance measures of CAUTI rates per 1,000 device-days (R1) were compared with CAUTI rates per 10,000 patient-days (R2) before (T1) and after (T2) an intervention aimed at reducing catheter utilization.

The device-utilization ratio declined from 0.36 to 0.28 between T1 and T2 (P< .001), while infection rates were significantly lower when measured by R2 (28.2 vs 23.2, P = .02). When measured by R1, however, infection rates trended upward by 6% (7.79 vs. 8.28, P = .47), and at the nursing unit level, reduction in device utilization was significantly associated with increases in infection rate.

The widely accepted practice of using device-days as a method of risk adjustment to calculate device-associated infection rates may mask the impact of a successful quality improvement program and reward programs not actively engaged in reducing device usage.

Considerable hospital resources are dedicated to minimizing the number of methicillin-resistant Staphylococcus aureus (MRSA) infections. One tool that is commonly used to achieve this goal is surveillance for MRSA colonization. This process is costly, and false-positive test results lead to isolation of individuals who do not carry MRSA. The performance of this technique would improve if patients who are at high risk of colonization could be readily targeted.

Five MRSA colonization prediction rules of varying complexity were derived in a population of 23,314 patients who were consecutively admitted to a US hospital and tested for colonization. Rules incorporated only prospectively collected, structured electronic data found in a patient's record within 1 day of hospital admission. These rules were tested in a validation cohort of 26,650 patients who were admitted to 2 other hospitals.

The prevalence of MRSA at hospital admission was 2.2% and 4.0% in the derivation and validation cohorts, respectively. Multivariable modeling identified predictors of MRSA colonization among demographic, admission-related, pharmacologic, laboratory, physiologic, and historical variables. Five prediction rules varied in their performance, but each could be used to identify the 30% of patients who accounted for greater than 60% of all cases of MRSA colonization and approximately 70% of all MRSA-associated patient-days. Most rules could also identify the 20% of patients with a greater than 8% chance of colonization and the 40% of patients among whom colonization prevalence was 2% or less.

We report electronic prediction rules that can fully automate triage of patients for MRSA-related hospital admission testing and that offer significant improvements on previously reported rules. The efficiencies introduced may result in savings to infection control programs with little sacrifice in effectiveness.

We evaluated the usefulness of topical decolonization therapy for reducing the risk of methicillin-resistant Staphylococcus aureus (MRSA) infection among MRSA-colonized inpatients.

Retrospective cohort study.

Three hospitals with universal surveillance for MRSA; at their physician's discretion, colonized patients could be treated with a 5-day course of nasal mupirocin calcium 2%, twice daily, plus Chlorhexidine gluconate 4% every second day.

MRSA carriers were later retested for colonization (407 subjects; study 1) or followed up for development of MRSA infection (933 subjects; study 2). Multivariable methods were used to determine the impact of decolonization therapy on the risks of sustained colonization (in study 1) and MRSA infection (in study 2).

Independent risk factors for sustained colonization included residence in a long-term care facility (odds ratio [OR], 1.8 [95% confidence interval {CI}, 1.1–3.2]) and a pressure ulcer (OR, 2.3 195% CI, 1.2–4.4]). Mupirocin at any dose decreased this risk, particularly during the 30-60-day period after therapy; mupirocin resistance increased this risk (OR, 4.1 [95% CI, 1.6–10.7]). Over a median follow-up duration of 269 days, 69 (7.4%) of 933 patients developed infection. Independent risk factors for infection were length of stay (hazard ratio [HR], 1.2 per 5 additional days [95% CI, 1.0–1.4]), chronic lung disease (HR, 1.7 [95% CI, 1.0–2.8]), and receipt of non-MRSA-active systemic antimicrobial agents (HR, 1.8 [95% CI, 1.1–3.1]). Receipt of mupirocin did not affect the risk of infection, although there was a trend toward delayed infection among patients receiving mupirocin (median time to infection, 50 vs 15.5 days; P = .06).

Mupirocin-based decolonization therapy temporarily reduced the risk of continued colonization but did not decrease the risk of subsequent infection.