A 2-year-old, previously healthy boy with an upper respiratory infection presents to the emergency department with acute-onset stridor and respiratory distress. He quickly improves with intramuscular dexamethasone and initiation of heliox via high-flow nasal cannula. He is admitted to the PICU for respiratory insufficiency, requiring heliox, and high risk of respiratory failure. Before calling the PICU team for admission, the emergency department resident sends a respiratory viral polymerase chain reaction (PCR) panel. When her attending questions its necessity, she answers, “They [PICU admitting team] will ask for it. They always want to know which bug it is.” Historically, at our institution, respiratory viral PCR panels were commonly ordered for patients with more severe disease requiring PICU admission to find the culprit virus. Local practice had evolved to sending respiratory viral PCR tests for the majority of PICU patients with respiratory complaints, even before admission.
Since its introduction, respiratory viral PCR panel–testing proponents have advocated its use to reduce unnecessary health care resource use.1–3 However, studies in which authors evaluate the clinical impact of respiratory viral PCR panel testing have varied in their results and lacked consistent evidence to support changes in management or outcomes, including antibiotic administration, chest radiographic imaging, hospital admission, and hospital length of stay.3–7 In addition, commonly held beliefs include the need to test for isolation and cohort practices or to identify treatable viruses. The Choosing Wisely guidelines of the American Society for Clinical Pathology recommend against routine, broad respiratory viral PCR panels on the basis of the evidence and, in place, recommend specific rapid molecular tests that impact management, such as respiratory syncytial virus (RSV) or influenza A and B testing.8,9 In critically ill pediatric patients, current evidence does not support routine testing for common diagnoses, such as status asthmaticus or bronchiolitis.8,10
We aim to “bend the value curve”11 through a pediatric critical care fellow–led quality improvement project addressing overuse of respiratory viral PCR testing. The primary specific aim was to reduce respiratory viral PCR panel testing per PICU patient day in a 35-bed, medical-surgical PICU at Cincinnati Children’s Hospital Medical Center by 30% within 1 year and demonstrate sustainability. We selected a 30% overall reduction because of the necessity for extended respiratory viral PCR panel testing in specific PICU populations for prognostication and treatment. The secondary specific aim was to reduce patient charges for viral respiratory testing within the PICU, defined as patient charges per patient day. The balancing metrics were rate of rapid molecular testing (RSV, influenza A and B, pertussis and/or parapertussis), rate of total overall respiratory viral testing (respiratory viral PCR panel and rapid molecular testing), ventilator days, and hospital length of stay. The project was initiated in November 2018, and data were collected throughout 2 winter seasons. The winter viral respiratory season was defined yearly as November 1 through March 31, mirroring the institutional guideline’s time frame for palivizumab prophylaxis.
Before initiating tests of change, we surveyed PICU fellows, attending physicians, and advanced practice registered nurses to establish consensus (defined as >80% agreement) regarding eliminating routine respiratory viral PCR panel testing for specific patient populations. The majority indicated that routine respiratory viral PCR panel testing could be safely eliminated for patients who are nonintubated with bronchiolitis (96%), status asthmaticus (92%), and viral croup or stridor receiving heliox (88%) because test results do not typically inform clinical management. However, routine respiratory viral PCR panel testing impacted PICU providers’ clinical management decisions for patients intubated for respiratory decompensation due to confirmed or suspected lower respiratory tract infection, patients with acute-on-chronic respiratory failure in the setting of complex chronic medical problems, and patients with risk factors for progression to severe disease. Having achieved consensus, PICU providers agreed to avoid routine testing with respiratory viral PCR panels in nonintubated patients with bronchiolitis, status asthmaticus, and croup.
Our first plan, do, study, act cycle was initiated on November 20, 2018, and concentrated on the education of frontline providers. Although education alone is a low-reliability intervention, a clear knowledge gap existed regarding testing costs and diagnostic implications that necessitated focused education. Knowledge gaps on the utility of viral PCR testing have been described previously. The results of cross-sectional studies, such as those by Gill et al12 and Huang et al,13 highlight reasons often communicated by clinicians for obtaining respiratory viral PCR panel testing. These include reassurance or diagnosis validation to lessen self-doubt, desire to reduce further unnecessary medical interventions, support for medication initiation or discontinuation, test administration directed by another clinician, and parental desire for a viral diagnosis.12,13 Common misconceptions in our unit that were discovered during this education included the need for respiratory viral PCR panel testing to place patients in cohorts or to determine disease severity and prognostic information for all patients. These misconceptions were addressed directly in the educational process, and the use of targeted testing for treatable viruses, such as influenza, was also discussed. Next, sensitivity, specificity, and comparative cost data about the viral respiratory tests (ie, extended or syndromic respiratory viral PCR, rapid RSV, rapid influenza A and B antigen, and Bordetella pertussis and/or parapertussis PCR) were provided to all members of the PICU team, including attending physicians, fellows, residents, advanced practice registered nurses, nurses, and respiratory therapists. We delivered education via multiple formats, including displays of flyers in the PICU, discussions during PICU resident orientation, and presentations at monthly PICU morbidity and mortality conferences. By achieving testing consensus and using focused cost and diagnostic education, respiratory viral PCR testing decreased 27% from a baseline median of 70 tests per 1000 patient days to 51 tests per 1000 patient days by the end of the first winter season, 2018–2019.
Because of concern, education would not reliably sustain practice change, the next plan, do, study, act cycle standardized the PICU respiratory viral PCR panel order process. Using results from the initial consensus survey, we developed a testing algorithm (November 27, 2019), which aided decision-making and prioritized cost-effective, targeted testing for patients with a clinical suspicion of influenza, RSV, or pertussis, that was supported by the Choosing Wisely guidelines of the American Society of Clinical Pathology.9 The algorithm emphasized specific testing that may alter clinical course or management (eg, treatment with oseltamivir). Additionally, the algorithm reserved respiratory viral PCR testing for patients with acute respiratory decompensation (Fig 1). This algorithm, along with repeated education, facilitated a sustained decrease to a median of 28 respiratory viral PCR panel tests per 1000 patient days at the end of the second winter season, 2019–2020 (Fig 2A), with an overall median decrease for the 2 winter seasons to 33 respiratory viral PCR tests per 1000 patient days, a 53% decrease.
For the secondary specific aim, namely, reduction in patient charges for respiratory viral testing, patient charges were tracked over the 2 winter seasons. Respiratory viral PCR panel–testing patient charges decreased from $108 per patient day to $66 per patient day. Rapid molecular (RSV and/or influenza) testing increased from $1.60 per patient day to only $2.78 per patient day. This represents an overall decrease in hospital charges related to respiratory viral testing of $40.82 per patient day, a 38% decrease.
For the first balancing metric, that is, the rate of rapid molecular (RSV and/or influenza) tests sent, testing increased from a median of 10 rapid RSV and/or influenza tests per 1000 patient days preproject initiation to 22 rapid RSV and/or influenza tests per 1000 patient days postproject initiation, a 120% increase. The second balancing metric, median total overall respiratory viral testing, decreased from 83 total tests per 1000 patient days preproject implementation to 59 total tests per 1000 patient days postproject implementation. These results reveal that although other rapid molecular testing increased, we still demonstrated an overall reduction in testing and a clear reduction in charges. No significant changes in other balancing metrics occurred during the course of the project, including ventilator days or hospital length of stay. Furthermore, PICU providers did not feel using the algorithm led to any unsafe patient care scenarios, such as those reviewed at monthly morbidity and mortality conferences.
Through incorporating a value-focused quality improvement project in the pediatric critical care curriculum over a 2-year period, we successfully facilitated and sustained a practice change to reduce routine respiratory viral PCR testing. By achieving buy-in from key stakeholders, establishing a multidisciplinary education program, and implementing a standardized testing algorithm, we surpassed our goal. A trainee-led initiative can reduce unnecessary testing, thereby decreasing patient charges and resource use while maintaining high-quality patient care. Even in the pediatric critical care environment, we believe that for the 9-month-old infant with bronchiolitis on high-flow nasal cannula, the 4-year-old child with status asthmaticus on continuous albuterol, and the 2-year-old toddler with croup on heliox, you do not always need proof of the culprit.
Dr Innis conceptualized and designed the study and drafted the initial manuscript; Drs Hasson, Bodilly, Sveen, and Stalets conceptualized and designed the study; Dr Dewan conceptualized and designed the study and conducted the initial analyses; and all authors reviewed and revised the manuscript and approved the final manuscript as submitted.
FUNDING: Dr Dewan has a grant from the Agency for Healthcare Research and Quality (K08HS026975) payable to her institution.
POTENTIAL CONFLICTS OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
FINANCIAL DICLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.