Diagnostic Stewardship for Ventilator Associated Pneumonia

Diagnostic Stewardship for Ventilator Associated Pneumonia: A Pragmatic Cluster-randomized Crossover Trial of a Hybrid Order-review and Laboratory Reporting Intervention

Centers for Disease Control and Prevention, George Washington University, Baylor College of Medicine, Virginia Commonwealth University

The goal of this pragmatic cluster-randomized crossover trial is to test if less unnecessary antibiotics are prescribed when the lab reports respiratory culture test results in a specific way for patients who have respiratory cultures obtained, but do not meet clinical criteria for ventilator associated pneumonia (VAP). The main question it aims to answer is: Does a modified culture reporting intervention reduce unnecessary antibiotics for ventilated patients in the intensive care unit (ICU)? Researchers will compare antibiotic use outcomes between eligible patients whose test results are communicated using the modified reporting and those with standard reporting of results.

The specific aims of this project are: Specific Aim 1: In a cluster-randomized crossover trial among 6 ICUs across 3 medical centers, evaluate the impact of a VAP diagnostic stewardship intervention on antibiotic use, VAP diagnoses, and adverse events. Hypothesis: A change in unnecessary antibiotics for VAP and in VAP clinical diagnoses in the intervention vs. control periods across all sites, without a change in adverse events, is expected. Specific Aim 2: Evaluate overall impact of intervention including clinical and antibiotic outcomes using the "Desirability of Outcome Ranking (DOOR)/ Response Adjusted for Duration of Antibiotic Risk (RADAR)" methodology. Hypothesis: A change in overall patient outcomes (better DOOR ranking, accounting for duration of antibiotic use) in the intervention vs. control period is expected.

This is a pragmatic cluster-randomized crossover trial in 6 adult ICUs (1 medical and 1 surgical ICU each at 3 hospitals). ICUs (clusters) will be randomized at the hospital-level (1:1) to a sequence of intervention (6 months) followed by control (6 months), or vice-versa. The study population is ventilated patients with respiratory culture orders but not meeting clinical pneumonia criteria, and organism(s) growing in respiratory culture. A research physician will evaluate likelihood of pneumonia using a standard clinical algorithm, within 24 hours of the order Mon-Fri. If pneumonia criteria are not met AND there is growth of organism(s) (except normal respiratory flora), during the intervention period, the culture report will be modified to reflect the likelihood of asymptomatic colonization instead of reporting bacterial identification. Clinicians may call the lab for identification and susceptibilities if necessary. During the control period, no modification to reporting will occur.

Ordering care providers will not be aware if the test results are eligible for modified reporting or not.

Laboratory reporting of respiratory cultures will continue per regular or routine laboratory protocols (standard reporting).

The lab will publish a modified report for respiratory cultures which do not meet clinical criteria for pneumonia and have growth of organisms (other than normal respiratory flora). The modified report will include the likelihood of colonization instead of reporting bacterial identification.

If appropriateness of culturing i.e., clinical criteria for pneumonia testing does not meet the algorithm AND there is growth of one or more organism(s) that are not considered normal upper respiratory flora, during the intervention period, the result will be modified to reflect the likelihood of asymptomatic colonization.

The primary efficacy outcome is change in proportion of unnecessary antibiotic therapy for presumed VAP in the intervention (modified reporting) period compared to control (standard reporting) period. Antibiotic use for VAP relative to the index respiratory culture will be classified as follows: Not on empiric antibiotic(s), new antibiotic started based on index respiratory culture result Not on empiric antibiotic(s), no new antibiotics initiated based on culture result On empiric antibiotic(s), changed based on index respiratory culture, completed course for VAP On empiric antibiotic(s), continued without change following index respiratory culture, completed course for VAP On empiric antibiotic(s), antibiotics discontinued based on culture results On empiric antibiotics(s), but this was not directed towards VAP Antibiotic use in categories a, c, and d will together be considered "unnecessary antibiotic therapy for VAP treatment".

To understand the impact on antibiotic use in ICU patients, we will measure the duration of antibiotic therapy for study patients, unit-level total antibiotic days of therapy (DOTs) standardized to patient census, and unit-level respiratory antibiotic days of therapy (DOTs) standardized to patient census.

Measurement of this outcome will help understand how modification of reporting would influence the likelihood of a patient receiving a diagnosis of VAP. The clinical diagnosis of VAP is an intermediate outcome that precedes the endpoint of antibiotic use. Because there are no consensus definitions for VAP, we will use the treating clinician's documentation of pneumonia or VAP and treatment for those conditions to capture this outcome. Similarly, clinical diagnosis of tracheobronchitis within 7 days after the index culture will also be recorded.

We will perform a sensitivity analysis of the impact on antibiotic use outcomes after excluding patients diagnosed with tracheobronchitis within 7 days of index culture. The hypothesis is that the intervention will have limited impact on those diagnosed and treated for tracheobronchitis, and thus, we may see a larger difference between intervention and control when excluding patients with diagnoses of tracheobronchitis.

Ventilator-free days is defined as the number of calendar days within 28 days after the index respiratory culture on which the patient was not mechanically ventilated. Any patient dying within 28 days of the index respiratory culture collection will be assigned zero ventilator-free days. Ventilator-free days will be compared between the intervention and control periods for all study ICUs combined, and by hospital.

Frequency of requests for complete reports in the intervention period - these data are applicable in the intervention period only. This will be a primarily descriptive analysis to help us understand if the intervention worked as intended i.e., how frequently did clinicians still want the full report with organism identification despite the nudge to consider asymptomatic colonization.

Death, bacteremia, and septic shock (from any cause) within 7 days of the index culture order will be recorded as potential significant adverse events from a delayed or missed true VAP diagnosis or delay in initiation of appropriate antimicrobial therapy. Two-person adjudication will be used to determine whether the above events were attributed to the intervention, during the intervention period. Rates of these events will be compared between intervention and control periods individually for each event, and as a combined adverse event outcome.

The clinical outcome of a participant will be ranked from most to least desirable outcome, to obtain a 5-level ordinal outcome. We will conduct a series of pairwise comparisons of each intervention patient to each control patient using two rules: When comparing patients with different overall clinical outcomes, the patient with a better overall clinical outcome based on pre-specified ordered ranks is determined to have the better outcome. When comparing two patients who have the same rank based on overall clinical outcome, the patient with a shorter course of antibiotics has the better outcome. The DOOR/RADAR probability represents the probability that a randomly selected participant who received the intervention (modified reporting) has a better overall outcome than a randomly selected participant who received standard reporting (control), accounting for antibiotic duration of therapy.

Inclusion Criteria: Patient located on ICU unit included in the study Patient is mechanically ventilated Patient had respiratory culture sent >48 hours after admission Patient age ≥ 18 years Exclusion Criteria: Extra corporeal membrane oxygenation (ECMO) at time of respiratory culture Heart or lung transplant Culture rejected by lab per standard lab protocol Prisoners Severe immunosuppression as defined by: <6 months from solid organ transplant (SOT) OR <6 months from treatment for acute rejection following SOT Active treatment for lymphoreticular malignancies Neutropenic < 1000 Receiving lymphodepleting chemotherapy Allogeneic stem cell transplants <6 months Autologous stem cell transplants or chimeric antigen receptor T-cell (CAR-T) therapy <6 months out Allogeneic stem cell transplant with graft vs host disease (GVHD) or receiving 2 or more immunosuppressants Advanced or untreated human immunodeficiency virus (HIV) infection with CD4 < 200 Receiving biologics within 6 months