Assessment of the Appropriateness of Antimicrobial Use in US Hospitals

This cross-sectional study evaluates the appropriateness of antimicrobial use for hospitalized patients treated for community-acquired pneumonia or a urinary tract infection present at admission or for patients who had received fluoroquinolone or intravenous vancomycin treatment.


Urinary tract infection analysis pathway
We included patients treated for urinary tract infection (UTI) present on admission to the survey hospital who were not reported to be pregnant, neutropenic, or to have had a solid organ or hematopoietic stem cell transplant, and who received ≥1 calendar day of inpatient antimicrobial medications for treatment of UTI only. Patients with other infections during their hospitalization were excluded (eFigure 2). Neutropenia was assessed both as an underlying condition and on the first day of UTI treatment.
Qualifying signs or symptoms of UTI in the first 2 hospital days included: fever, urgency, frequency, pain or burning with urination, costovertebral angle pain or tenderness, suprapubic pain or tenderness, or visible blood in urine. Mental status change in the first 2 hospital days also qualified if it was accompanied by: 1) white blood cell count >10,000 cells/mm 3 on the first day of UTI treatment; 2) systolic BP <90 mmHg, mean arterial pressure <65 mmHg, or vasopressor administration on the first day of UTI treatment; 3) lactate >2 mmol/L (>18 mg/dL) on the first day of UTI treatment; or 4) ≥2 of the following on the first day of UTI treatment: temperature <36°C; heart rate >90 beats per minute; respiratory rate >20 breaths per minute (or PaCO2 <32 mmHg); white blood cell count <4,000 cells/mm 3 or >10% bands.
When evaluating available urine culture data, we included each patient's first positive urine culture and first positive blood culture, plus any positive urine or blood cultures collected within 1 day of the first culture(s). We included the following urine culture sources: clean catch urine, catheter urine, other urine, fluid from the kidney, or fluid from a nephrostomy tube. Eligible positive urine cultures were defined as those with ≤2 bacteria isolated. 2 The following organisms were considered colonizers and excluded: yeast, Candida species, and normal or mixed flora. 2 Data on bacterial colony forming units per ml were collected but were not used in the analysis pathway. Positive blood cultures were defined as those with ≥1 National Healthcare Safety Network (NHSN)-defined recognized pathogen isolated. 3 Blood cultures positive for common commensals, as defined in the NHSN bloodstream infection protocol, were excluded. Antimicrobial susceptibility data were frequently not available, so we determined whether identified pathogens were likely to have been susceptible to antimicrobial medications being given to the patient.
In some cases, patients who were reported to have been given post-discharge UTI treatment had missing duration of post-discharge treatment. To estimate total treatment duration for these patients (inpatient plus postdischarge), we determined the median duration of post-discharge treatment among those with known duration and the same duration of inpatient treatment and added this median value to the inpatient treatment duration for patients with missing post-discharge treatment duration.
We evaluated microbiology data from cultures or culture-independent tests collected during the period starting 5 days before the start date of FQ treatment and ending on the last date of FQ treatment. We sought to apply similar rules to determining the eligibility of culture data for inclusion in the pathway as we applied in the CAP and UTI pathways. In addition, we included blood cultures that were positive for NHSN common commensals if there were two or more separate, positive cultures. 3 For some positive cultures, and for all negative culture and cultureindependent diagnostic tests, the date of final microbiological test result was not available to use in determining whether FQ treatment was stopped promptly in response to test results. In these instances, we applied a proxy time of 3 days from specimen collection to final result date using the median time from collection to final result date among patients for whom these data were available. Fluoroquinolone susceptibility data were frequently not available, so we determined whether identified pathogens were likely to have been susceptible to the FQ being given to the patient.
In some cases, patients who were reported to have been given post-discharge FQ treatment had missing duration of post-discharge treatment. To estimate total treatment duration for these patients (inpatient plus postdischarge), we determined the median duration of post-discharge treatment among those with known duration and the same duration of inpatient treatment and added this median value to the inpatient treatment duration for patients with missing post-discharge treatment duration.
We evaluated microbiology data from cultures or culture-independent tests collected during the period starting 5 days before the start date of VANC treatment and ending on the last date of VANC treatment. We sought to apply similar rules to determining the eligibility of culture data for inclusion in the pathway as we applied in the CAP and UTI pathways. In addition, we included blood cultures that were positive for NHSN common commensals if there were two or more separate, positive cultures. 3 For some positive cultures, and for all negative culture and culture-independent diagnostic tests, the date of final microbiological test result was not available to use in determining whether VANC treatment was stopped promptly in response to test results. In these instances, we applied a proxy time of 3 days from specimen collection to final result date using the median time from collection to final result date among patients for whom these data were available. Vancomycin, oxacillin, and penicillin or ampicillin susceptibility data were frequently not available, so we determined whether identified pathogens were likely to have been susceptible.
Patients were considered to have a severe penicillin allergy based on the following reported reactions: hives or urticaria, blisters, wheezing, throat tightness, trouble breathing, angioedema, face, tongue, oral or lip swelling, or anaphylaxis. Free text fields clearly falling into any of these categories, or indicating any of the following, were also considered severe reactions: Stevens-Johnson syndrome, toxic epidermal necrolysis, Drug Reaction with Eosinophilia and Systemic Symptoms, fainting, syncope, joint swelling, liver failure, seizure, convulsions, encephalopathy, pancytopenia, thrombocytopenia, nephritis or acute kidney injury, and neuropathy. Examples of reported allergic reactions that were not considered severe: any gastrointestinal tract-related symptoms, other skin rash (including reports of cracked or peeling skin), fever, yeast infection, itching, and hand numbness.
In some cases, patients who were reported to have been given post-discharge VANC treatment had missing duration of post-discharge treatment. To estimate total treatment duration for these patients (inpatient plus postdischarge), we determined the median duration of post-discharge treatment among those with known duration and the same duration of inpatient treatment and added this median value to the inpatient treatment duration for patients with missing post-discharge treatment duration. Did the patient have: 1) signs or symptoms of UTI in the first 2 hospital days; OR 2) a matching pathogen isolated from an eligible positive urine culture and an eligible blood culture collected within 1 calendar day of each other?

Yes, n=278
Did the patient have a positive urine culture (any CFU/ml) with ≤2 species isolated OR a positive blood culture for a recognized pathogen?

Yes, n=167
Were pathogens isolated from eligible positive culture(s) susceptible (or if unknown, likely to have been susceptible) to antimicrobial medications given to the patient to treat UTI?

No, n=4
Was antimicrobial treatment for UTI stopped within 3 calendar days?
Was the patient treated only with a fluoroquinolone (inpatient and outpatient)?

No, n=95
Yes, n=171 No, n=107 Yes, n=12 Did the patient have a fever in the first 2 hospital days or on the first day of UTI treatment, OR did the patient have a positive blood culture?