Value-Based Care for Healthy Children With First Episode of Febrile Neutropenia Free

This QI initiative took place in a large urban quaternary care pediatric hospital in Canada. The ED has a census of ∼80 000 visits per year. At our institution, patients with nononcologic febrile neutropenia who require hospitalization are managed by general pediatricians and may have a consultation with hematology during hospitalization or in outpatient follow-up.

Using the Model for Improvement methodology,¹⁹  we first assembled a team of key stakeholders including pediatric emergency medicine physicians, hematologists, hematology nurses, a pediatric hospitalist, and an infectious disease physician who leads the institutional antimicrobial stewardship program. We conducted a baseline audit of the management of febrile neutropenia in previously healthy patients seen in the ED. We performed a literature review to identify larger studies to determine risk of complications, including serious bacterial infections and any consensus recommendations for this patient population.

The paucity of existing best practice guidelines in the literature supported the theory that a knowledge gap was the main contributor to the current practice variation. Therefore, the main intervention planned was an ED guideline for the management of healthy children presenting with a first episode of febrile neutropenia. The literature review and experts’ consensus helped create the guideline and identify inclusion and exclusion criteria. The guideline was presented to the pediatrics, hematology, and emergency medicine divisions for educational purposes. Feedback from these sessions was integrated for further guideline iterations.

The guideline (Fig 1) is composed of 3 sections: inclusion criteria (age 6 months and older, first episode of febrile neutropenia, and ANC below 0.5 × 10⁹/L), exclusion criteria (risk factors, laboratory results such as ANC 0.2 × 10⁹/L or less, and challenges with follow-up), and recommended management of low-risk patients (blood culture collection, no routine hospitalization or antibiotics, rapid follow-up, and proper discharge instructions). Very severe neutropenia is defined as ANC 0.2 × 10⁹/L or lower. Taking a conservative approach, hematology expert consensus suggested this cutoff as an exclusion for this study given evidence of increased risk of infection extrapolated from the oncology, severe aplastic anemia, and severe congenital anemia populations.^20,21 

If follow-up with a primary care physician within 24 to 48 hours was not feasible, emergency physicians were instructed to prebook a scheduled follow-up appointment with a community pediatrician partner, some of whom had weekend availability. The discharge instructions included clear follow-up explanations for the patient and the primary care physician and were incorporated into our electronic medical system to ensure appropriate outpatient follow-up for this population. The guideline was officially launched for use in the ED at the end of January 2020.

The guideline was further refined 1 year after its launch because data analysis revealed that several patients were being excluded because of an associated mild thrombocytopenia (platelets 100–150 × 10⁹/L), which is also known to sometimes occur with viral illnesses in addition to neutropenia. Of these patients with associated mild thrombocytopenia hospitalized with antibiotics, none had positive blood cultures and their final discharge diagnoses were viral illnesses. Through consensus from project stakeholders, including full hematology divisional support, patients with associated mild thrombocytopenia, in addition to their neutropenia, were included in the low-risk category. These changes were shared with all stakeholders by e-mail, and the revised guideline was relaunched in the ED in April 2021 (Fig 1).

The guideline was sent via e-mail to all pediatric emergency physicians and fellows and was displayed in different areas in the ED as a reminder. The guideline was also distributed to the general pediatric residents and fellows and was added to their shared-drive resource folder. Multiple reminders were sent throughout this QI study to account for normal trainee turnover in an academic institution. Targeted audit and feedback were conducted when a low-risk patient did not follow the guideline (Supplemental Fig 5). An e-mail with the guideline attached was sent to the physicians, notifying them of the deviation and reminding them of the new guideline. In Fall 2020, this initiative was accepted as 1 of 5 new recommendations for our hospital’s Choosing Wisely campaign, further promoting and prioritizing this improvement work across the organization.

Retrospective monthly chart reviews were done to establish our baseline data from June 2018 to December 2019, and prospectively from January 2020 to January 2022, to assess the effectiveness of our interventions. A database was created by our laboratory’s information system’s technology team and updated monthly. To avoid missing cases, all patients presenting to the ED who had a neutrophil count below 0.5 × 10⁹/L were included in the database. Data were automatically extracted using the electronic medical record (EPIC-Epic Systems Corporation, Verona, WI). Every chart was reviewed manually by 2 authors (C.G.B., C.K.L.) to apply inclusion and exclusion criteria from the guideline.

The primary outcome measure was the percentage of low-risk patients with a first episode of febrile neutropenia receiving empirical antibiotics and/or hospitalization. The secondary outcome measure was the percentage of inaccurate diagnoses of severe neutropenia (neutrophil count below 0.5 × 10⁹/L but an ANC at or above 0.5 × 10⁹/L). The process measure was the percentage of cases managed incorrectly that received audit and feedback. Balancing measures included missed SBI, ED return visit within 72 hours, and a new, unanticipated hematologic diagnosis within 6 months after discharge (eg, leukemia) for patients with first episode of febrile neutropenia.

Microsoft Excel and QI macros (QI Macros for Excel, version 2022; KnowWare International, Inc, Denver, CO) were used to create statistical process charts to monitor changes in the outcome measures over time. Special cause variation was assessed using standard statistical process chart methods.²²  This project was approved by our institutional quality risk management team and did not require research ethics board approval.

Baseline data identified 15 low-risk patients over 18 months with a first episode of severe febrile neutropenia; 73% of these patients were either hospitalized and/or received antibiotics. Of these, 36% received antibiotics and were discharged from the hospital and 64% received antibiotics and were hospitalized. Additionally, of 18 patients who were misdiagnosed with severe febrile neutropenia by excluding bands from the absolute neutrophils count, 55% received antibiotics and/or were hospitalized. After implementation, there were 88 patients with a diagnosis of febrile neutropenia, 57 of whom met exclusion criteria, leaving 31 patients (35%) with low-risk criteria over a 24-month period (Fig 2).

For the primary outcome measure, the percentage of low-risk patients with a first episode of severe febrile neutropenia (ANC between 0.2 and 0.5 × 10⁹/L) that were hospitalized and/or received antibiotics decreased from 73.3% to 12.9% (n = 4) (Fig 3). This represents 27 patients who were managed appropriately and avoided unnecessary hospitalization or antibiotics. Of the 4 patients managed inappropriately, 25% received antibiotics and were discharged, 25% were hospitalized without antibiotics, and 50% received antibiotics and were hospitalized. No patient received antibiotics for another diagnosis. Special cause variation was detected in April/May 2020 and was sustained for 9 months.

For the secondary outcome measure, the misdiagnosis of severe febrile neutropenia receiving antibiotics and/or hospitalization decreased from 55.6% to 7.7% (n = 6) (Fig 4). This represents 72 patients who were managed appropriately and avoided unnecessary hospitalization or antibiotics. Special cause was detected in June/July 2020 and sustained for 8 months. Four cases (100%) of low-risk patients with febrile neutropenia managed incorrectly according to the guideline underwent audit and feedback.

Eight patients with low-risk febrile neutropenia and associated mild thrombocytopenia received antibiotics and were hospitalized before our guideline iteration. After our guideline iteration to include mild thrombocytopenia, 2 patients with febrile neutropenia associated with mild thrombocytopenia were managed as per the new guideline recommendations.

One low-risk patient had a positive blood culture growing coagulase-negative staphylococci, which was identified as a contaminant. No cases of true bacteremia were identified. Two (6%) patients had a return ED visit within 72 hours after the initial discharge. Neither patient was admitted on the return visit. One case was for the aforementioned contaminated positive blood culture. The second case was for ongoing fever with a subsequent diagnosis of viral illness, who was discharged again from the ED. No hematologic diagnosis was made for any of the study patients in the 6 months after discharge.

This QI initiative reduced the use of unnecessary antibiotics and hospitalization in healthy, low-risk children presenting with a first episode of febrile neutropenia. With the development of a standardized clinical guideline reinforced by education, reminders, and targeted feedback, we were able to standardize the medical management and improve the value of care provided to these children. Additionally, the recognition of true severe neutropenia with inclusion of bands was also improved throughout our study, leading to fewer inaccurate diagnoses of severe neutropenia. Equally important, these changes were achieved without any unintended consequences because there were no missed cases of SBIs or other serious hematologic diagnoses.

This novel QI initiative is the first published study to our knowledge attempting to standardize the approach to healthy children presenting with a first episode of febrile neutropenia while reducing unnecessary and potentially harmful care. Hospitalizations and antibiotics are not without risk and can lead to antibiotic side effects, antimicrobial resistance, nosocomial infections, caregiver stress, and other avoidable consequences, especially when not indicated. Many recent publications on fever and neutropenia in healthy children support the efficacy and safety of a standardized, less aggressive approach to the management of these patients and support risk stratification for the management of previously healthy children with febrile neutropenia, which our QI initiative has accomplished.^1,15,16 

Given the absence of a published clinical practice guideline by larger pediatric societies, we recognized the significant practice variation in the management of this population within our institution, with the majority of patients receiving unnecessary antibiotics and/or hospitalization. This quality gap demonstrated the need for the development of a local guideline to reduce variation and ultimately improve patient care. Clinical pathways have been well described in the literature demonstrating that standardizing care improves quality of care and patient safety.^23,24 

Various components likely contributed to the success of our QI initiative, mainly the development of our local guideline that efficiently summarized best practice recommendations at the bedside. Furthermore, the engagement of our key stakeholders to streamline management recommendations across departments, along with education and reminders, audit, and feedback, and, importantly, the eventual alignment with the local Choosing Wisely campaign all helped support the implementation and sustainability of this initiative. Making this initiative a hospital Choosing Wisely recommendation positively affected our outcomes because it helped disseminate this work even more broadly and supported the sustainability of these improvements. The support by organizational leadership for our hospital’s Choosing Wisely campaign along with the evolving resource stewardship culture in our hospital engages and motivates clinicians to critically reflect on their practice.²⁵  The hospital’s Choosing Wisely campaign will help support ongoing sustainability of this work. Choosing Wisely initiatives are presented at the beginning and middle of every academic year as a reminder and educational opportunity.

This QI initiative was completed in a single quaternary care pediatric and may not be generalizable to other contexts. Parts of the guideline may need to be modified to comply with local practices, patient populations, and available resources. For example, outpatient follow-up with a physician within 24 to 48 hours was feasible in our center either with the patient’s primary care physician or through a scheduled follow-up, but may not be in other contexts, where next-day primary care access may be more difficult. Furthermore, our local Choosing Wisely campaign helped disseminate our work and supported sustainability, which may be more challenging in other contexts. Because charts were reviewed retrospectively, we were unable to ensure that the guideline was used for each patient. However, with a significant change in practice variation temporally after our guideline implementation, we would expect that the guideline was the cause for change. Furthermore, our balancing measures were only monitored at our center. Patients could have returned to another ED within 72 hours. Although the patients also could have had a subsequent hematologic diagnosis, our center is the only quaternary care pediatric hospital in the area that would make this less likely. Furthermore, we were unable to confirm that the discharge instructions were followed and that a follow-up occurred with the recommended bloodwork. Although the number of low-risk patients with severe febrile neutropenia only amounted to 31 over a 2-year period, this is not out of keeping with what has been published in the literature. Our region has several other EDs where children with viral-associated neutropenia are seen. Hindie et al published a retrospective chart review in 2018 on the risk of bacteremia in previously healthy children presenting with severe neutropenia and fever. Over a 6-year period, they identified 47 cases of severe de novo neutropenia with fever, which is lower than our case rate.¹⁷  Furthermore, this guideline was launched shortly before the COVID-19 pandemic, which led to a decrease in other circulating viruses, increase in multisystem inflammatory syndrome in children screening, and other practice changes. All of these pandemic-related factors likely affected the incidence of viral-induced neutropenia.

Through a multifaceted, multidisciplinary QI study, we improved resource stewardship and value-based care by reducing unnecessary hospitalizations and antibiotics in low-risk patients with a first episode of febrile neutropenia. This work can be adopted by other pediatric and community sites caring for children to expand its impact. The development of society-based practice guidelines with a risk-stratified approach to febrile neutropenia would support further dissemination of this work.

We thank Sheila Butchart, BScN, and Michelle Fantauzzi, BScN, The Hospital for Sick Children.