Validate a public health model identifying patients at high risk for carbapenem-resistant Enterobacterales (CRE) on admission and evaluate performance across a healthcare network.

Retrospective case-control studies

Adults hospitalized with a clinical CRE culture within 3 days of admission (cases) and those hospitalized without a CRE culture (controls).

Using public health data from Atlanta, GA (1/1/2016–9/1/2019), we validated a CRE prediction model created in Chicago. We then closely replicated this model using clinical data from a healthcare network in Atlanta (1/1/2015–12/31/2021) (“Public Health Model”) and optimized performance by adding variables from the healthcare system (“Healthcare System Model”). We frequency-matched cases and controls based on year and facility. We evaluated model performance in validation datasets using area under the curve (AUC).

Using public health data, we matched 181 cases to 764,408 controls, and the Chicago model performed well (AUC 0.85). Using clinical data, we matched 91 cases to 384,013 controls. The Public Health Model included age, prior infection diagnosis, number of and mean length of stays in acute care hospitalizations (ACH) in the prior year. The final Healthcare System Model added Elixhauser score, antibiotic days of therapy in prior year, diabetes, admission to the intensive care unit in prior year and removed prior number of ACH. The AUC increased from 0.68 to 0.73.

A CRE risk prediction model using prior healthcare exposures performed well in a geographically distinct area and in an academic healthcare network. Adding variables from healthcare networks improved model performance.

Examine the relationship between patients’ race and prescriber antibiotic choice while accounting for differences in underlying illness and infection severity.

Retrospective cohort analysis.

Acute care facilities within an academic healthcare system.

Adult inpatients from January 2019 through June 2022 discharged from the Hospital Medicine Service with an ICD-10 Code for Pneumonia.

We describe variability in days of therapy of antimicrobials with activity against Pseudomonas aeruginosa (anti-Pseudomonas agents) or against MRSA (anti-MRSA agents), by patient’s race and ethnicity. We estimated the likelihood of receipt of any anti-Pseudomonas agents by race and modeled the effect of race on rate of use, adjusting for age, severity, and indication.

5,820 patients with 6,700 encounters were included. After adjusting for broad indication, severity, underlying illness, and age, use of anti-Pseudomonas agents were less likely among non-Hispanic Black patients than other race groups, although this effect was limited to younger patients (adjusted odds ratio [aOR] 0.45, 95% confidence interval [CI] 0.29, 0.70), and not older ones (aOR 0.98; 95% CI 0.85, 1.13); use of anti-MRSA agents were similar between groups. Among patients receiving any anti-Pseudomonas agents, Black patients received them for relatively lower proportion of their inpatient stay (incidence rate ratio 0.91; 95% CI 0.87, 0.96).

We found difference in use of anti-Pseudomonas agents between non-Hispanic Black patients and other patients that could not be easily explained by indications or underlying illness, suggesting unmeasured factors may be playing a role in treatment decisions.

To examine the relationship between race and ethnicity and central line-associated bloodstream infections (CLABSI) while accounting for inherent differences in CLABSI risk related to central venous catheter (CVC) type.

Retrospective cohort analysis.

Acute care facilities within an academic healthcare system.

Adult inpatients from January 2012 through December 2017 with CVC present for ≥2 contiguous days.

We describe variability in demographics, comorbidities, CVC type/configuration, and CLABSI rate by patient’s race and ethnicity. We estimated the unadjusted risk of CLABSI for each demographic and clinical characteristic and then modelled the effect of race on time to CLABSI, adjusting for total parenteral nutrition use and CVC type. We also performed exploratory analysis replacing race and ethnicity with social vulnerability index (SVI) metrics.

32,925 patients with 57,642 CVC episodes met inclusion criteria, most of which (51,348, 89%) were among non-Hispanic White or non-Hispanic Black patients. CVC types differed between race/ethnicity groups. However, after adjusting for CVC type, configuration, and indication in an adjusted cox regression, the risk of CLABSI among non-Hispanic Black patients did not significantly differ from non-Hispanic White patients (adjusted hazard ratio [aHR] 1.19; 95% confidence interval [CI]: 0.94, 1.51). The odds of having a CLABSI among the most vulnerable SVI subset compared to the less vulnerable was no different (odds ratio [OR] 0.95; 95% CI: 0.75–1.2).

We did not find a difference in CLABSI risk between non-Hispanic White and non-Hispanic Black patients when adjusting for CLABSI risk inherent in type and configuration of CVC.

Among inpatients, peer-comparison of prescribing metrics is challenging due to variation in patient-mix and prescribing by multiple providers daily. We established risk-adjusted provider-specific antibiotic prescribing metrics to allow peer-comparisons among hospitalists.

Using clinical and billing data from inpatient encounters discharged from the Hospital Medicine Service between January 2020 through June 2021 at four acute care hospitals, we calculated bimonthly (every two months) days of therapy (DOT) for antibiotics attributed to specific providers based on patient billing dates. Ten patient-mix characteristics, including demographics, infectious disease diagnoses, and noninfectious comorbidities were considered as potential predictors of antibiotic prescribing. Using linear mixed models, we identified risk-adjusted models predicting the prescribing of three antibiotic groups: broad spectrum hospital-onset (BSHO), broad-spectrum community-acquired (BSCA), and anti-methicillin-resistant Staphylococcus aureus (Anti-MRSA) antibiotics. Provider-specific observed-to-expected ratios (OERs) were calculated to describe provider-level antibiotic prescribing trends over time.

Predictors of antibiotic prescribing varied for the three antibiotic groups across the four hospitals, commonly selected predictors included sepsis, COVID-19, pneumonia, urinary tract infection, malignancy, and age >65 years. OERs varied within each hospital, with medians of approximately 1 and a 75th percentile of approximately 1.25. The median OER demonstrated a downward trend for the Anti-MRSA group at two hospitals but remained relatively stable elsewhere. Instances of heightened antibiotic prescribing (OER >1.25) were identified in approximately 25% of the observed time-points across all four hospitals.

Our findings indicate provider-specific benchmarking among inpatient providers is achievable and has potential utility as a valuable tool for inpatient stewardship efforts.

Understanding characteristics of healthcare personnel (HCP) with SARS-CoV-2 infection supports the development and prioritization of interventions to protect this important workforce. We report detailed characteristics of HCP who tested positive for SARS-CoV-2 from April 20, 2020 through December 31, 2021.

CDC collaborated with Emerging Infections Program sites in 10 states to interview HCP with SARS-CoV-2 infection (case-HCP) about their demographics, underlying medical conditions, healthcare roles, exposures, personal protective equipment (PPE) use, and COVID-19 vaccination status. We grouped case-HCP by healthcare role. To describe residential social vulnerability, we merged geocoded HCP residential addresses with CDC/ATSDR Social Vulnerability Index (SVI) values at the census tract level. We defined highest and lowest SVI quartiles as high and low social vulnerability, respectively.

Our analysis included 7,531 case-HCP. Most case-HCP with roles as certified nursing assistant (CNA) (444, 61.3%), medical assistant (252, 65.3%), or home healthcare worker (HHW) (225, 59.5%) reported their race and ethnicity as either non-Hispanic Black or Hispanic. More than one third of HHWs (166, 45.2%), CNAs (283, 41.7%), and medical assistants (138, 37.9%) reported a residential address in the high social vulnerability category. The proportion of case-HCP who reported using recommended PPE at all times when caring for patients with COVID-19 was lowest among HHWs compared with other roles.

To mitigate SARS-CoV-2 infection risk in healthcare settings, infection prevention, and control interventions should be specific to HCP roles and educational backgrounds. Additional interventions are needed to address high social vulnerability among HHWs, CNAs, and medical assistants.

We evaluated the impact of test-order frequency per diarrheal episodes on Clostridioides difficile infection (CDI) incidence estimates in a sample of hospitals at 2 CDC Emerging Infections Program (EIP) sites.

Observational survey.

Inpatients at 5 acute-care hospitals in Rochester, New York, and Atlanta, Georgia, during two 10-workday periods in 2020 and 2021.

We calculated diarrhea incidence, testing frequency, and CDI positivity (defined as any positive NAAT test) across strata. Predictors of CDI testing and positivity were assessed using modified Poisson regression. Population estimates of incidence using modified Emerging Infections Program methodology were compared between sites using the Mantel-Hanzel summary rate ratio.

Surveillance of 38,365 patient days identified 860 diarrhea cases from 107 patient-care units mapped to 26 unique NHSN defined location types. Incidence of diarrhea was 22.4 of 1,000 patient days (medians, 25.8 for Rochester and 16.2 for Atlanta; P < .01). Similar proportions of diarrhea cases were hospital onset (66%) at both sites. Overall, 35% of patients with diarrhea were tested for CDI, but this differed by site: 21% in Rochester and 49% in Atlanta (P < .01). Regression models identified location type (ie, oncology or critical care) and laxative use predictive of CDI test ordering. Adjusting for these factors, CDI testing was 49% less likely in Rochester than Atlanta (adjusted rate ratio, 0.51; 95% confidence interval [CI], 0.40–0.63). Population estimates in Rochester had a 38% lower incidence of CDI than Atlanta (summary rate ratio, 0.62; 95% CI, 0.54–0.71).

Accounting for patient-specific factors that influence CDI test ordering, differences in testing practices between sites remain and likely contribute to regional differences in surveillance estimates.

To determine the incidence of severe acute respiratory coronavirus virus 2 (SARS-CoV-2) infection among healthcare personnel (HCP) and to assess occupational risks for SARS-CoV-2 infection.

Prospective cohort of healthcare personnel (HCP) followed for 6 months from May through December 2020.

Large academic healthcare system including 4 hospitals and affiliated clinics in Atlanta, Georgia.

HCP, including those with and without direct patient-care activities, working during the coronavirus disease 2019 (COVID-19) pandemic.

Incident SARS-CoV-2 infections were determined through serologic testing for SARS-CoV-2 IgG at enrollment, at 3 months, and at 6 months. HCP completed monthly surveys regarding occupational activities. Multivariable logistic regression was used to identify occupational factors that increased the risk of SARS-CoV-2 infection.

Of the 304 evaluable HCP that were seronegative at enrollment, 26 (9%) seroconverted for SARS-CoV-2 IgG by 6 months. Overall, 219 participants (73%) self-identified as White race, 119 (40%) were nurses, and 121 (40%) worked on inpatient medical-surgical floors. In a multivariable analysis, HCP who identified as Black race were more likely to seroconvert than HCP who identified as White (odds ratio, 4.5; 95% confidence interval, 1.3–14.2). Increased risk for SARS-CoV-2 infection was not identified for any occupational activity, including spending >50% of a typical shift at a patient’s bedside, working in a COVID-19 unit, or performing or being present for aerosol-generating procedures (AGPs).

In our study cohort of HCP working in an academic healthcare system, <10% had evidence of SARS-CoV-2 infection over 6 months. No specific occupational activities were identified as increasing risk for SARS-CoV-2 infection.

To estimate prior severe acute respiratory coronavirus virus 2 (SARS-CoV-2) infection among skilled nursing facility (SNF) staff in the state of Georgia and to identify risk factors for seropositivity as of fall 2020.

Baseline survey and seroprevalence of the ongoing longitudinal Coronavirus 2019 (COVID-19) Prevention in Nursing Homes study.

The study included 14 SNFs in the state of Georgia.

In total, 792 SNF staff employed or contracted with participating SNFs were included in this study. The analysis included 749 participants with SARS-CoV-2 serostatus results who provided age, sex, and complete survey information.

We estimated unadjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) for potential risk factors and SARS-CoV-2 serostatus. We estimated adjusted ORs using a logistic regression model including age, sex, community case rate, SNF resident infection rate, working at other facilities, and job role.

Staff working in high-infection SNFs were twice as likely (unadjusted OR, 2.08; 95% CI, 1.45–3.00) to be seropositive as those in low-infection SNFs. Certified nursing assistants and nurses were 3 times more likely to be seropositive than administrative, pharmacy, or nonresident care staff: unadjusted OR, 2.93 (95% CI, 1.58–5.78) and unadjusted OR, 3.08 (95% CI, 1.66–6.07). Logistic regression yielded similar adjusted ORs.

Working at high-infection SNFs was a risk factor for SARS-CoV-2 seropositivity. Even after accounting for resident infections, certified nursing assistants and nurses had a 3-fold higher risk of SARS-CoV-2 seropositivity than nonclinical staff. This knowledge can guide prioritized implementation of safer ways for caregivers to provide necessary care to SNF residents.

To determine the impact of an inpatient stewardship intervention targeting fluoroquinolone use on inpatient and postdischarge Clostridioides difficile infection (CDI).

We used an interrupted time series study design to evaluate the rate of hospital-onset CDI (HO-CDI), postdischarge CDI (PD-CDI) within 12 weeks, and inpatient fluoroquinolone use from 2 years prior to 1 year after a stewardship intervention.

An academic healthcare system with 4 hospitals.

All inpatients hospitalized between January 2017 and September 2020, excluding those discharged from locations caring for oncology, bone marrow transplant, or solid-organ transplant patients.

Introduction of electronic order sets designed to reduce inpatient fluoroquinolone prescribing.

Among 163,117 admissions, there were 683 cases of HO-CDI and 1,104 cases of PD-CDI. In the context of a 2% month-to-month decline starting in the preintervention period (P < .01), we observed a reduction in fluoroquinolone days of therapy per 1,000 patient days of 21% after the intervention (level change, P < .05). HO-CDI rates were stable throughout the study period. In contrast, we also detected a change in the trend of PD-CDI rates from a stable monthly rate in the preintervention period to a monthly decrease of 2.5% in the postintervention period (P < .01).

Our systemwide intervention reduced inpatient fluoroquinolone use immediately, but not HO-CDI. However, a downward trend in PD-CDI occurred. Relying on outcome measures limited to the inpatient setting may not reflect the full impact of inpatient stewardship efforts.

Fluoroquinolones (FQs) and extended-spectrum cephalosporins (ESCs) are associated with higher risk of Clostridioides difficile infection (CDI). Decreasing the unnecessary use of FQs and ESCs is a goal of antimicrobial stewardship. Understanding how prescribers perceive the risks and benefits of FQs and ESCs is needed.

We conducted interviews with clinicians from 4 hospitals. Interviews elicited respondent perceptions about the risk of ESCs, FQs, and CDI. Interviews were audio recorded, transcribed, and analyzed using a flexible coding approach.

Interviews were conducted with 64 respondents (38 physicians, 7 nurses, 6 advance practice providers, and 13 pharmacists). ESCs and FQs were perceived to have many benefits, including infrequent dosing, breadth of coverage, and greater patient adherence after hospital discharge. Prescribers stated that it was easy to make decisions about these drugs, so they were especially appealing to use in the context of time pressures. They described having difficulty discontinuing these drugs when prescribed by others due to inertia and fear. Prescribers were skeptical about targeting specific drugs as a stewardship approach and felt that the risk of a negative outcome from under treatment of a suspected bacterial infection was a higher priority than the prevention of CDI.

Prescribers in this study perceived many advantages to using ESCs and FQs, especially under conditions of time pressure and uncertainty. In making decisions about these drugs, prescribers balance risk and benefit, and they believed that the risk of CDI was acceptable in compared with the risk of undertreatment.

To determine the effect of an electronic medical record (EMR) nudge at reducing total and inappropriate orders testing for hospital-onset Clostridioides difficile infection (HO-CDI).

An interrupted time series analysis of HO-CDI orders 2 years before and 2 years after the implementation of an EMR intervention designed to reduce inappropriate HO-CDI testing. Orders for C. difficile testing were considered inappropriate if the patient had received a laxative or stool softener in the previous 24 hours.

Four hospitals in an academic healthcare network.

All patients with a C. difficile order after hospital day 3.

Orders for C. difficile testing in patients administered a laxative or stool softener in <24 hours triggered an EMR alert defaulting to cancellation of the order (“nudge”).

Of the 17,694 HO-CDI orders, 7% were inappropriate (8% prentervention vs 6% postintervention; P < .001). Monthly HO-CDI orders decreased by 21% postintervention (level-change rate ratio [RR], 0.79; 95% confidence interval [CI], 0.73–0.86), and the rate continued to decrease (postintervention trend change RR, 0.99; 95% CI, 0.98–1.00). The intervention was not associated with a level change in inappropriate HO-CDI orders (RR, 0.80; 95% CI, 0.61–1.05), but the postintervention inappropriate order rate decreased over time (RR, 0.95; 95% CI, 0.93–0.97).

An EMR nudge to minimize inappropriate ordering for C. difficile was effective at reducing HO-CDI orders, and likely contributed to decreasing the inappropriate HO-CDI order rate after the intervention.

To determine the best nursing home facility characteristics for aggregating antibiotic susceptibility testing results across nursing homes to produce a useful annual antibiogram that nursing homes can use in their antimicrobial stewardship programs.

Derivation cohort study.

Center for Medicare and Medicaid Services (CMS) certified skilled nursing facilities in Georgia (N = 231).

All residents of eligible facilities submitting urine culture specimens for microbiologic testing at a regional referral laboratory.

Crude and adjusted metrics of antibiotic resistance prevalence (percent of isolates testing susceptible) for 5 bacterial species commonly recovered from urine specimens were calculated using mixed linear models to determine which facility characteristics were predictive of testing antibiotic susceptibility.

In a single year, most facilities had an insufficient number of isolates tested to create facility-specific antibiograms: 49% of facilities had sufficient Escherichia coli isolates tested, but only about 1 in 10 had sufficient isolates of Klebsiella pneumoniae, Proteus mirabilis, Enterococcus faecalis, or Pseudomonas aeruginosa. After accounting for antibiotic tested and age of the patient, facility characteristics predictive of susceptibility were: E. coli, region, year, average length of stay; K. pneumoniae, region, bed size; P. mirabilis, region; and for E. faecalis or P. aerginosa no facility parameter remained in the model.

Nursing homes often have insufficient data to create facility-specific antibiograms; aggregating data across nursing homes in a region is a statistically sound approach to overcoming data shortages in nursing home stewardship programs.