This is a prospective pilot/feasibility trial of a bundled diagnostic stewardship intervention at the level of the microbiologic testing pathway in ventilator-associated pneumonia (VAP). The study utilized a pre/post design and was initially registered as a single-arm trial because the study intervention could only be applied prospectively to a single group (all patients hospitalized in study ICUs requiring ventilation during the trial intervention period). The study objectives are to safely and effectively reduce antibiotic overuse and its attendant hazards (adverse drug events, Clostridioides difficile diarrhea and generation of multidrug-resistant organisms) among mechanically-ventilated patients. Participating ICUs will have the following three modifications made in their respiratory culture workflows for mechanically-ventilated patients: 1) providers will be required to select a valid indication for respiratory culture performance (worsening ventilator requirements, purulent sputum production, and/or new radiographic infiltrate on chest imaging); 2) respiratory cultures will be preferentially obtained via bronchoscopic or nonbronchoscopic BAL (by respiratory therapists) rather than via endotracheal aspiration; and 3) BAL samples will be sent for cell count and differentials, and respiratory culture results will not be released for samples with <50% neutrophils. The study will carefully monitor adherence to study interventions, ICU-specific antibiotic utilization rates, and important safety metrics including rates of mortality, ventilator-dependence and ventilator-associated events.

The trial hypotheses are:

• Implementation of a VAP diagnostic stewardship bundle will be successfully implemented without significant increases in mortality or ventilator-associated events.
• Implementation of a VAP diagnostic stewardship bundle will be associated with a reduction in ICU-specific antibiotic utilization rates

Overdiagnosis of ventilator-associated pneumonia (VAP) is common among mechanically-ventilated patients and contributes to antibiotic overuse and the generation of multidrug resistant organisms within intensive care units (ICUs). Identification of interventions that safely and effectively lower VAP overdiagnosis and antibiotic overuse are important for antimicrobial stewardship programs. Antibiotic stewardship interventions targeting VAP have heretofore focused on therapeutic interventions-antibiotic de-escalation or discontinuation-in established VAP cases, but have not leveraged interventions at the level of the VAP diagnostic testing pathway to minimize overdiagnosis in the first place. Current microbiologic testing practices-specifically, indiscriminate ordering, collection and interpretation of respiratory cultures-incents VAP overdiagnosis and antibiotic overuse.

This is a prospective pilot/feasibility trial of a bundled diagnostic stewardship intervention targeting the microbiologic diagnostic testing pathway among mechanically-ventilated patients. The study utilized a pre/post design and was initially registered as a single-arm trial because the study intervention could only be applied prospectively to a single group (all patients hospitalized in study ICUs requiring ventilation during the trial intervention period). This pilot/feasibility trial will implement system-based changes within ICUs in microbiologic diagnostic testing workflows with the goal of safely lowering antibiotic utilization rates in mechanically-ventilated patients. Specifically, the study will sequentially implement the following 3 changes in the diagnostic testing workflow for clinical providers:

1. Ordering respiratory cultures:

   1. Current workflow: respiratory cultures can be ordered by frontline clinical providers without preconditions.
   2. Study intervention: frontline clinical providers will be required to select a valid indication for respiratory culture collection through a custom order set in the institutional electronic medical record (EMR).

2. Collection of respiratory cultures:

   1. Current workflow: providers may collect endotracheal aspirates (proximal lung secretions) or bronchoalveolar lavage (distal lung secretions) for analysis.
   2. Study intervention: providers will be required to preferentially use only bronchoalveolar lavage (BAL) to minimize false positive test results (assuming no contraindications to BAL performance). This will be accomplished through a custom order set in the institutional EMR.

3. Reporting of respiratory cultures:

   1. Current workflow: positive respiratory culture results are automatically released in the EMR, regardless of likelihood of infection.
   2. Study intervention: respiratory culture results will be automatically released in the EMR only for BAL samples with a polymorphonuclear (PMN) percentage of >50% (to minimize false positive test results). Respiratory culture results for BAL samples with PMN% <50% will be released if the primary team calls the microbiology lab directly and requests result release.

All proposed interventions are within the standard of care for routine clinical practice, but have not been operationalized in parallel to each other and studied explicitly with the goal of reducing antibiotic overuse. The first two study interventions will be implemented at the beginning of the trial period and the final study intervention will be implemented 6 months after. The interventions will be operationalized through use of a custom order set in the institutional EMR and modifications to Clinical Microbiology Laboratory Workflows, as well as through bimonthly provider educational sessions performed by study personnel during morning ICU rounds. Use of the clinical decision support tool within the EMR by frontline providers will be entirely volitional.

The study population in this trial includes all patients admitted to participating study ICUs for routine clinical care. Providers will engage in routine clinical care utilizing the aforementioned modified diagnostic testing workflows and the study will compare safety outcomes and ICU antibiotic utilization rates pre- and post-intervention implementation.

Healthcare providers in participating ICUs will provide routine clinical care utilizing these novel diagnostic testing workflows. Adherence to study interventions by these healthcare providers will be assessed during this study to inform project feasibility on a larger scale.

Inclusion Criteria Patients included:

• All patients hospitalized within the Michigan Medicine Cardiac Care Unit (CCU, 7D) and the Michigan Medicine Critical Care Medicine Unit (CCMU, 6D)

Exclusion Criteria Patients included:

• For the study intervention requiring performance of bronchoalveolar lavage rather than endotracheal aspirate for respiratory culture collection, the following are exclusion criteria for performance of bronchoalveolar lavage:

  • international normalized ratio (INR)>2,
  • platelet count <50,
  • gross blood in endotracheal secretions,
  • ratio of arterial oxygen partial pressure (PaO2 in mmHg) to fractional inspired oxygen (FiO2), ratio (P/F ratio)<80,
  • major lung surgery within prior 30 days.

Inclusion Criteria healthcare providers included:

• healthcare providers working in units (CCU and CCMU) as part of routine clinical care.

Exclusion Criteria healthcare providers included:

• healthcare providers that are not as part of routine care in CCU and CCMU

Up to 6 years (including 5 years pre-intervention and 1-year post intervention).

Apart from in-hospital mortality, the only adverse event data collected was ventilator-associated events (VAEs). VAE numbers are complete and accurate, but number of unique persons affected by VAEs was not collected. Given the constraints of the clinicaltrials.gov reporting system, the number of events is replicated in number of persons affected. In fact, the number of persons affected could be equal to or less than that number. No non-serious AE data was collected, thus the denominator is 0.

Ventilator-Associated Event numbers are complete and accurate, but number of unique persons affected by ventilator-associated events was not collected. Given the constraints of the clinicaltrials.gov reporting system, the number of events is replicated in number of persons affected. In fact, the number of persons affected could be equal to or less than that number.

Due to the nature of this pilot trial the study will obtain an exemption from informed consent.