Medical Outcomes, Quality of Life, and Family Perceptions for Outpatient vs Inpatient Neutropenia Management After Chemotherapy for Pediatric Acute Myeloid Leukemia

Key Points Question What are the clinical outcomes and patient and family experiences associated with outpatient neutropenia management after intensive chemotherapy for pediatric acute myeloid leukemia (AML) compared with inpatient management? Findings In this cohort study including as many as 554 pediatric patients, patients discharged to outpatient management during neutropenia did not experience higher bacteremia incidence, delays to subsequent courses, or worse health-related quality of life compared with those who received inpatient management. Most patients and families were satisfied with the discharge practice offered by the treating institution, but experiences varied, suggesting that outpatient management may not be appropriate for all families. Meaning These findings suggest that outpatient management during neutropenia was a viable approach without excess risk for many patients and was preferred by some families.


Introduction
Pediatric patients with newly diagnosed acute myeloid leukemia (AML) receive multiple intensive chemotherapy courses, causing prolonged neutropenia and a high risk for life-threatening infections. 1,2 In 2000, Children's Cancer Group mandated hospitalization during neutropenia for patients in the CCG2961 clinical trial because of an initial 19% treatment-related mortality rate. 3 This recommendation carried forward in all phase III Children's Oncology Group (COG) AML clinical trials. 4,5 Most centers in the US, Canada, and Europe follow this guideline, but some report outpatient management some or all of the time. [6][7][8] Research comparing outpatient vs inpatient management has not included integrated assessments of medical outcomes, family preferences, and patient-and family-centered outcomes (PCOs). While adult AML studies suggest that outpatient management may be clinically safe, [9][10][11][12][13][14][15] pediatric data are limited by conflicting results. A single-center study by Inoue et al 16 reported similar rates of relapse and mortality for outpatient vs inpatient management of neutropenia. However, a retrospective multicenter observational study of administrative data found that outpatient management was associated with higher rates of antibiotic, vasopressor, and supplemental oxygen use. 6 Additionally, data regarding patient and family preferences for inpatient vs outpatient management during neutropenia is limited to children with low-risk neutropenia, 17,18 and preferences for management in children with AML have not been assessed, to our knowledge.
This study used a comprehensive mixed-methods design to compare medical outcomes, identify PCOs, assess health-related quality of life (HRQOL) and other identified PCOs, and describe the lived experience of outpatient and inpatient management after intensive AML chemotherapy.
We hypothesized that outpatient management would be associated with an increased risk of bacteremia but higher quality of life.

Methods
The study protocols were approved by institutional review boards at all participating sites. Consent requirements were waived for medical record extractions because the research met the criteria of 45 CFR 46.116(d). All patients who participated in interviews or completed questionnaires provided written informed consent. The analyses and reporting of the quantitative objectives followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for observation cohort studies.

Study Design
This cohort study used a multiphase mixed methods design (Figure). First, standardized medical record abstraction captured primary medical outcomes for pediatric patients with AML who received outpatient or inpatient neutropenia management. Next, semistructured interviews identified outcomes important to patients and their families. 19 Patients receiving frontline chemotherapy were then prospectively invited to complete questionnaires measuring patient HRQOL and PCOs identified from the semistructured interviews. Lastly, the qualitative interviews provided depth of understanding and context for the quantitative findings.
Patients were eligible for inclusion in the overall cohort if they had a diagnosis of AML and were aged younger than 19 years at initial diagnosis. Patients with acute promyelocytic leukemia and patients who received reduced intensity chemotherapy or only received a hematopoietic stem cell transplantation at the participating site were ineligible.

Medical Record Abstraction
Trained abstractors collected detailed data on demographics, diagnosis, treatment, antimicrobial prophylaxis, all inpatient and outpatient encounters, and all blood culture results from initial AML admission until recovery after the last frontline chemotherapy course for retrospectively and prospectively identified patients diagnosed from January 2011 to July 2019 at 17 pediatric institutions across the US. Institutions were surveyed for standard supportive care practices, including approach to systemic anti-infective prophylaxis, central line care, and antiseptic bathing protocols.

Qualitative Interviews
From November 2015 to February 2017, semistructured interviews were conducted at 9 pediatric institutions with patients and families experiencing either management strategy. Details about data collection and analysis have been previously reported in accordance with Consolidated Criteria for Reporting Qualitative Research (COREQ) reporting guideline. 19 In addition to identifying PCOs for prospective assessment, interviews captured patient and family perceptions about the lived experience of neutropenia, including their satisfaction with their management strategy, impact on family life, and beliefs about risks and benefits of each approach.

Prospective Questionnaires
From June 2016 to May 2019, patients receiving frontline chemotherapy were prospectively invited to questionnaires administered at 2 time points during a single post-induction I chemotherapy course: within the period from first day to the last day of chemotherapy within the given course (baseline), and within the period from recovery of absolute neutrophil count to greater than 500 cells/uL (to convert to cells ×10 9 /L, multiply by 0.001) until start of next chemotherapy course Based on expert recommendations on the use of established assessments, PedsQL and COST were administered at baseline and follow-up, whereas SDSC-DIMS and PIP-D were administered at follow-up only. SDSC-DIMS, PIP-D, and COST assessments were incorporated based on themes identified from analysis of the qualitative interviews 19 and were implemented as the amended protocol was approved by the institutional review boards of participating sites. Therefore, not all assessments were administered to every participant who consented to participate.

Exposure
Neutropenia management strategy, categorized as inpatient or outpatient, was the primary exposure. Patients discharged home within 3 days after course-specific chemotherapy completion were categorized as receiving outpatient management, even if they were subsequently readmitted.
Patients meeting discharge eligibility criteria but remaining in inpatient care more than 3 days after chemotherapy completion were categorized as receiving inpatient management. The threshold defining outpatient management was chosen based on prior data from Getz et al 6 on timing of initial discharge relative to course-specific chemotherapy completion and represents a period within which neutropenia recovery is improbable.

Outcomes
Primary medical outcomes were course-specific bacteremia incidence and time to initiation of the next chemotherapy course. Bacteremia follow-up began 3 days after course-specific chemotherapy completion and continued until the earliest of death, absolute neutrophil count recovery to greater than 500 cells/uL, or initiation of subsequent chemotherapy course. First occurrence of bacteremia in each course was defined as a single positive blood culture result for a bacterial pathogen unless the bacterium was considered a common commensal organism by the National Healthcare Safety Network. For common commensals other than Viridans group streptococci, 2 positive cultures within 3 consecutive days were required for classification as bacteremia.
Course-specific mortality was a secondary medical outcome. Inpatient resource utilization and intensive care unit (ICU)-level requirements were additional secondary outcomes ascertained from the Pediatric Health Information Systems (PHIS) database for the subpopulation of patients treated at PHIS institutions. 24,25 The primary PCO was HRQOL, measured using caregiver proxy responses to acute PedsQL 4.0 Generic Core Scales. Secondary PCOs included patient sleep disturbance, measured using the SDCS-DIMS 22 ; parental stress, measured with the PIP-D 21 ; and financial distress, evaluated with the modified COST. 23

Statistical Analysis
Analyses were restricted to post-induction I chemotherapy courses, an a priori decision based on prior data that discharge during the first course is rare even at institutions where outpatient management is standard practice. Analyses at each treatment course were restricted to patients Linear regressions compared follow-up PIP-D scores and change in COST scores, and generalized linear models with γ distribution compared SDSC-DIMS scores.

JAMA Network Open | Pediatrics
Control for confounding was accomplished through adjustment for propensity score quintiles as well as any remaining unbalanced patient-or hospital-level confounders. Propensity scores were

Medical Outcomes
The full abstraction cohort included 610 patients treated during the study periods at each contributing site. Of these patients, 554 (90.8%) met early discharge-eligibility criteria in at least 1 course and contributed a total of 1196 post-induction I courses (eFigure 1 in the Supplement). Table 1 presents distributions of patient and hospital characteristics for the induction II study population of Course-specific differences in incidence of bacteremia for outpatient vs inpatient management were not statistically significant ( Table 2). Times to the next course ranged from 31 to 40 days overall.
Differences in time-to-next course varied by course, but was only statistically significant for outpatient management compared with inpatient management in induction II (difference, −3.1 [95% CI, −5.2 to −1.0] days; P = .003) (    Table 3. Mean PedsQL Generic Core scores were low and did not differ for outpatient

Patient and Caregiver Perceptions
We interviewed 86 respondents (32 children and 54 caregivers) from 57 families with a child who received AML chemotherapy. Of these 57 families, 39 received inpatient management and 18 received outpatient management. Exemplar quotes demonstrating key qualitative findings are presented in Table 4.

Inpatient Management
A total of 57 respondents who experienced inpatient management (86.4%; 32 caregivers, 25 children) expressed satisfaction with hospitalization for the duration of neutropenia. Respondents perceived the hospital to be the safest place for their child owing to its close proximity to emergency care, the reassurance of constant medical surveillance, and a perceived lower risk of infection. They change. Data presented are the change in COST scores from baseline to follow-up assessment. Measure of association reflect the mean difference in change scores between outpatient and inpatient such that positive differences suggest worse financial distress among the inpatient management group. Adjusted for propensity score quintile. noted that prolonged hospitalization could be difficult, but positive aspects included engagement with Child Life Service staff, meaningful relationships with clinical caregivers, and the perception that the hospital was a safe space where children did not have to explain their illness to others.

JAMA Network Open | Pediatrics
A total of 9 respondents (13.6%; 7 caregivers and 2 children) who experienced inpatient management were dissatisfied. Both dissatisfied children (100%) cited family separation and 6 dissatisfied caregivers (85.7%) cited a belief that the hospital was riskier with respect to infection.
Four children of the 7 caregivers dissatisfied with inpatient management preferred the hospital.
These children perceived their home environment differently than their caregivers, describing it as stressful and chaotic.

Outpatient Management
A total of 20 respondents (85.0%; 12 caregivers and 5 children) who experienced home management were satisfied with their experience. They emphasized the psychosocial benefits for the child. Three respondents (15.0%; all caregivers) dissatisfied with outpatient management found caring for a neutropenic child at home to be very stressful owing to lack of support from family, challenges of managing infection risk, and anxiety about monitoring their child's health.

Discussion
This mixed-methods cohort study reports multiple medical and PCOs associated with management of patients receiving AML chemotherapy. Notably, pediatric patients with AML who were discharged to outpatient management did not have higher rates of bacteremia or delays in progression to subsequent treatment courses compared with those who received inpatient management. While the mortality rates overall were generally low, mortality in intensification II was higher in patients Neutropenia management strategy Hospital perceived to be safer for the child At the hospital, he got the help he needed, if things go downhill. There was a point where he was done with his chemo[therapy] and his ANC went down as expected and then started climbing up. He was feeling just fine. He was eating. He was drinking. He was playing. And then [snaps fingers] a fever.…He went to bed 1 night totally fine and woke up with a 104[°F] fever just a few hours later. It's just that quick, within 4 hours quick. And he was on the verge of going to the ICU. So yeah, now think of your child being at home, sitting on the couch and that happens? Do you call an ambulance? Do you get in the car and run stop signs to get to the hospital? Or would you rather already be at the hospital?

Father Inpatient
Home perceived to be cleaner and safer for the child Pretty much, I can keep his bedroom cleaner than the hospital room. To be honest with you, some days, I come here [to the hospital] and his floor is gross, and I'm like, "What the heck is going on? Can you please send environmental up here to clean his floor? It's disgusting." It's just disgusting. Then, he caught, each time he was here [at the hospital], he caught C. difficile, which I know he probably wouldn't have caught if he was at home, because it's a hospital-borne illness, and that was a really tough time for him too, going through that and being treated for it, bad stomach cramping. He's like balled up in knots. He can't eat because he's running to jump up, and he has to have a commode by his bed.

Mother Inpatient
Patient preference for inpatient management owing to home instability I was at Children's Hospital for 9 months. The first month or two, maybe the first month and a half, it wasn't-it didn't feel like home, but once I got used to it, you know, being around the environment and the sweet people and the nurses and doctors and everything and knew around the place, it was okay. It felt like I was home…I stopped noticing that I was in the hospital and I just felt like I was at home and, you know, like I was around sweet people, like I don't have to worry about yelling anymore, I don't have to worry about people lying, I don't have to worry about nothing bad at all. It was all good vibes and good things.

Patient Inpatient
To be honest, after going through it all, I actually would've rather stayed at the hospital, because the noise at home, after being perfectly quiet in the hospital room and the only thing happening was the beeping of the pump. That doesn't match how many kids are in our house yelling and hitting each other daily. So I would actually kind of rather stayed-because the noise itself just caused such a lot of stress for me, because I just hated that, compared to all the nice, perfectly quiet hospital room.

Patient Inpatient
Home perceived as closer to family and normal life The best part was just the comfort of home, just having her own bed and privacy and stuff like that. Being around family members, it was like a refresher. Even though you were scared to death, it was a refresher to go home, be normal for a while, be in our own home with our own stuff. These results align with prior studies demonstrating low mortality rates, 16 fewer days of febrile neutropenia, less antimicrobial use, 26 and higher ICU resource utilization for outpatient management. 6 Work in other pediatric cancer populations has also demonstrated sleep disturbances in children receiving inpatient care, 27 and low HRQOL has been similarly described in pediatric AML. 28 These findings have multiple implications for the care of children with AML. First, these data support outpatient management for select patients in post-induction I chemotherapy courses except intensification II. Outpatient management, as presaged in the qualitative data, may be associated with a lower rate of bacteremia and less sleep disruption, and potentially lower stress. For many families, these benefits could lead to a strong preference for outpatient management. However, the increased need for ICU-level support in patients who received outpatient management, which may reflect more severe infectious complications, may sway some families to prefer inpatient management. This preference may be heightened in families who view outpatient management as inherently more stressful and less safe than inpatient management.
The rarity of mortality events in intensification II makes definitive conclusions regarding the safety of outpatient management in that course challenging. However, given extensive external data documenting increased toxic effects and mortality in intensification II, 5,29 outpatient management in that course should be approached with particular caution.
Importantly, no single end point reported in this study can determine the optimal discharge strategy for an individual patient and their family. Rather, information from each data domain must be assessed by the clinical care team and integrated into nuanced discussions with patients and families. Some of these assessments typically are performed by nonclinicians, such as social workers, and require careful discussions of complex topics, such as physical or financial security. Close collaboration between medical and nonmedical staff is essential to determining the most effective discharge strategy for an individual patient. Discrepancies in preferences between patient and caregiver will need particularly close attention and careful management in determining the optimal approach. The coordinated effort needed to personalize neutropenia management for an individual patient may exceed the efforts needed to personalize therapy based on AML molecular characteristics.
These findings also have broader implications for children receiving chemotherapy and their families. These results clearly demonstrate the diverse experiences and preferences of patients and families during cancer care. Such heterogeneity is assuredly present across pediatric cancers and the processes needed to personalize AML hospital management will likely be applicable to multiple other supportive care practices. The remarkable breadth of information obtained from the mixedmethods approach would likely be reproduced in other pediatric cancer care questions. Moreover, the inclusion of patient and family voices directly in the research process provided an invaluable perspective that arguably should be included in all clinical evaluations that seek to provide a definitive care recommendation.

Limitations
Despite the multisite, mixed-methods approach, this study has limitations. The multi-center design was not able to overcome sample size limitations owing to the rarity of pediatric AML. While appropriately powered for the analyses of the primary medical outcomes, analyses of the secondary prospectively measured outcomes are based on small numbers. Additionally, course-specific comparisons of mortality in intensification II were based on very rare events. We sought to address these limitations with cautious interpretation of individual comparisons and careful evaluation of magnitudes of association and CIs. 30,31 Still, further evaluations of these prospectively measured PCOs within larger cohorts are warranted. Second, the manual health record abstraction, while extensive, did not include fungal infections or information required to enable systematic grading of bacteremia and sepsis. However, our secondary analyses of resource utilization aimed to provide additional context to the primary bacteremia comparison specifically related to relative overall acuity of patients who were discharged to outpatient management and those who were not. Future studies are warranted to identify potential mechanisms for the observed difference in ICU-level care requirements. Third, all patients were enrolled at US centers. The marked disparities in resources available to patients and families may limit direct generalizability of our findings to countries with national health insurance and strong social safety nets. Fourth, while we accounted for hospital-level infection control and central line management practices and use statistical methods to account for hospital-level clustering, there is a possibility for residual confounding by unmeasured or poorly measured factors, such as specific psychosocial or logistical factors considered by the institutions where standard practice was early discharge to outpatient management. Furthermore, 2 qualitative method limitations should be recognized. Despite focused efforts to include a diverse range of participants, interviewed patients and families may differ systematically from those declining interviews. Since demographic data on nonparticipants were not collected, an assessment of potential bias implications is not possible. However, this concern is partially mitigated by the large number and wide geographic distribution of interviewed patients and families. Additionally, perspectives of treating clinicians were not captured either in interviews or surveys; this is an important limitation that will need to be addressed in subsequent implementation studies.

Conclusions
This cohort study found that outpatient neutropenia management for pediatric AML may be undertaken safely for select patients and may be less burdensome to patients and families than inpatient management. However, not all chemotherapy courses may be amenable to outpatient management, given current supportive care and health care access in the US, and not all families will experience outpatient management as a preferred strategy. The combined use of medical outcomes, PCOs and qualitative interview data provides dramatically richer and more nuanced data to guide clinicians, patients, and families in the decision-making process around outpatient management.
These data will inform implementation studies to operationalize more personalized discharge practices for children receiving AML chemotherapy.