Context

This manuscript was prepared in accordance with the SQUIRE 2.0 (Standards for QUality Improvement Reporting Excellence) guidelines. ^{Reference Ogrinc, Davies, Goodman, Batalden, Davidoff and Stevens34}

Key stakeholders for this initiative included critical care, pharmacy, infectious diseases, pulmonology, emergency medicine, as well as the antimicrobial stewardship team. Our PICU is a 37-bed unit that manages medical/surgical, postsurgical cardiothoracic, and neurosurgical patients; patients who have undergone bone marrow and kidney transplantation; and those that require continuous renal replacement therapy and extracorporeal membrane oxygenation. The PICU team includes pediatric intensivists, fellows in pediatric critical care and emergency medicine, pediatric and emergency medicine residents, nurse practitioners, clinical pharmacists, registered nurses, respiratory therapists, dieticians, and students. At the time of study implementation, there were over 2,600 annual admissions to the PICU.

The patient population targeted by this initiative consisted of non-neonatal patients with an artificial airway, ETT or tracheostomy, and a concern for VAT. While patients with tracheostomies are at times excluded when studying VAT in patients with ETTs, ^{Reference Tamma, Turnbull, Milstone, Lehmann, Sydnor and Cosgrove26,Reference Ingolfsland, Gonzalez-Villamizar and Moore29} we chose to have an applicable intervention for all patients with an artificial airway. ^{Reference Ormsby, Conrad and Blumenthal28,Reference Sick-Samuels, Linz and Bergmann30}

Prior to this project, a VAT clinical decision-support tool was in place featuring standardized indications for obtaining a culture, a recommendation of a 5-day duration for antibiotic treatment, and guidance on how to interpret Gram stain results and choose the appropriate antibiotic. There was no clear data that practices were being followed.

Interventions

We identified three specific areas for improvement based on a multidisciplinary review of local practice and emerging literature: (1) standardization of recommendations for EAC collection; (2) diagnostic criteria for VAT; and (3) recommendations for antibiotic treatment. There was also an impetus to gain systemwide consensus on care. We therefore created a multidisciplinary improvement team in July 2022 involving one pediatric intensivist, one pediatric critical care medicine fellow, two pediatric infectious disease specialists, one pediatric pulmonologist, and two clinical pharmacists. We conducted a literature search and collected clinical decision-support tools from institutions across the country to inform our understanding of existing practices and to identify strategies that could be adapted and integrated into our own institutional framework (Supplemental Table 1).

We performed a series of Plan-Do-Study-Act (PDSA) cycles to achieve our objectives. A summary of key dates is listed in Table 1. Our first PDSA cycle aimed to achieve consensus through committee member meetings with key stakeholders between February and September 2023; proposed changes were presented to the inpatient clinical team. Our second PDSA cycle in June 2023 identified a PICU nurse “champion” and communicated availability to other nurses for education and questions regarding the improvement effort. Both cycles were meant to solicit feedback, build consensus, and inform iterative refinement of the intervention. Our third PDSA cycle involved the revision of the existing clinical decision-support tool, followed by approval and health systemwide implementation in November 2023 (Figure 1). The tool was shared by email to key stakeholders, posted in the PICU, and shared with the health system internal website.

Table 1.

Key dates

Table 1. Long description

A table titled ‘Key dates’ with two columns and ten rows. The columns are labeled ‘Date’ and ‘Event’. The rows list specific dates and corresponding events as follows: Row 1: Date, July 27th, 2022; Event, Multidisciplinary committee created. Row 2: Date, February 9th, 2023; Event, Presentation to pediatric critical care division to gain consensus. Row 3: Date, May 5th, 2023; Event, Presentation to pediatric critical care division to ensure consensus. Row 4: Date, May 17th, 2023; Event, Presentation to emergency medicine division performance improvement coordinating group to ensure consensus. Row 5: Date, June 20th, 2023; Event, Nurse champion identified and engaged to disseminate revisions to nursing staff. Row 6: Date, January 30th, 2023; Event, Updated changes approved by the Antimicrobial Stewardship Program. Row 7: Date, July 20th, 2023; Event, Pharmacy and Therapeutics committee approval. Row 8: Date, September 14th, 2023; Event, Presentation to pediatric critical care performance improvement coordinating group to disseminate revised recommendations. Row 9: Date, November 29th, 2023; Event, Adoption of revised clinical decision-support tool across the health system pediatric service line.

Figure 1.

Revised clinical decision-support tool. WBC, white blood cell; PIP, peak inspiratory pressure; PEEP, positive end expiratory pressure; FiO₂, fraction of inspired oxygen; PMN, polymorphonuclear neutrophil; ESBL, extended spectrum beta-lactamase; TMP-SMX, trimethoprim-sulfamethoxazole.

Revisions were informed by recently published evidence, review of existing institutional frameworks (Supplemental Table 1), and expert consensus among committee members. Key revisions included recommendations against repeating an EAC within 72 hours of a prior collection, implementation of standardized clinical criteria for VAT diagnosis, incorporation of a simplified Gram stain and culture interpretation support table, ^{Reference Craven, Hudcova and Lei35} and recommendation of a 3-day antibiotic treatment duration for uncomplicated cases. We created a novel definition of VAT that emphasized the necessary presence of both respiratory and systemic signs of infection.

Study Population

Patients admitted to the PICU who had an inpatient EAC performed between January 2022 and February 2025 were included in the analysis. Only cultures sent for patients admitted to the PICU were included, including those acquired elsewhere in the hospital (ie, the emergency department prior to physical transfer). Cultures obtained from other hospitals or as an outpatient were not included. Cultures obtained by bronchoalveolar lavage or intended for cystic fibrosis-specific evaluation were excluded. Data were collected from the electronic medical record. There were two team members who reviewed charts (E.F., J.H.) for a diagnosis of VAT. At time of culture retrieval, the presence of pneumonia on clinician documentation and/or radiographically diagnosed pneumonia led to exclusion of these cases for VAT related measures.

Measures

From a diagnostic stewardship perspective, we aimed to decrease EAC use. To assess this, we tracked the number of EACs sent per 100 ETT/tracheostomy days. We also aimed to decrease the rate of monthly serially repeated cultures per 100 ETT/tracheostomy days, which was defined as the number of cultures repeated within 72 hours for the same patient divided by the total number of cultures sent.

From an antimicrobial stewardship perspective, we aimed to decrease antibiotic usage for VAT. To assess this, we tracked DOT for VAT per 100 ETT/tracheostomy days. We defined this as the administration of antibiotics for patients with an ETT or tracheostomy with clinician documentation of VAT diagnosis. Patients who met diagnostic criteria, but did not receive antibiotics, were not included for the purposes of analyzing antibiotic utilization. Antimicrobials used for other purposes (ie, for pneumonia or non-respiratory infections) were not included in the DOT calculation. Antibiotics were counted starting on the day of initiation and included all consecutive antibiotic days until treatment completion. We tracked the monthly proportion of VAT episodes treated with 3 days of antibiotics, which was defined by the number of VAT episodes treated with a 3-day antibiotic duration divided by the total number of treated VAT episodes.

We tracked guideline-defined VAT diagnosis to assess adherence to the new diagnostic framework. Guideline-defined diagnosis met the appropriate clinical indications to obtain a culture, as outlined in Figure 1, and had a corresponding Gram stain with microscopic purulence and significant bacterial growth. The rate of guideline-defined VAT diagnosis was defined as number of episodes meeting this new framework divided by total number of episodes diagnosed, normalized by 100 ETT/tracheostomy days.

Balancing measures were based solely on data from the PICU. Ventilator-associated events were normalized per 1,000 ventilator days and ventilator days per 1,000 patient days. Mortality rate was defined as number of mortalities per 100 admissions. Average length of stay (LOS) was also included.

Cost savings were calculated based on the decrease in EACs sent to the microbiology laboratory. A charge of $20.00 per EAC included both the material and labor costs for Gram stain, culture, and antibiotic susceptibility testing. This, however, did not include savings related to nursing time to collect samples and a potential related reduction in antibiotic use.

Analysis

We used Shewhart control charts to provide a visual representation of process performance over time. U-charts were used to monitor the number of non-conformities per unit with a variable sample size; I-charts to monitor individual continuous data; and P-charts to monitor proportion of observations over time. Charts were created using Microsoft Excel. Centerlines were re-computed when 8 consecutive data points were above or below the mean. Control limits were set at 3 standard deviations above and below the mean. Cumulative data were analyzed using Microsoft Excel (Microsoft Corporation; Redmond, WA).

Ethical considerations

This initiative did not involve human subjects per the Northwell Institutional Review Board and therefore did not require further review (Human Subjects Research Determination Request 23-0193).