“Better at home”: Mixed methods report of intricacies in pediatric febrile neutropenia management

Abstract Introduction Many febrile neutropenia (FN) episodes are low risk (LR) for severe outcomes and can safely receive less aggressive management and early hospital discharge. Validated risk tools are recommended by the Children's Oncology Group to identify LR FN episodes. However, the complex dynamics of early hospital discharge and burdens faced by caregivers associated with the FN episode have been inadequately described. Methods An adapted quality‐of‐life (QoL) survey instrument was administered by a convergent mixed methods design; qualitative and quantitative data from two sources, the medical record and the mixed methods survey instrument, were independently analyzed prior to linkage and integration. Code book was informed by conceptual framework; open coding was used. Mixed methods analysis used joint display of results to determine meta‐inferences. Results Twenty‐eight patient–caregiver dyads participated with a response rate of 87%. Of the 27 FN episodes, 51.8% (14/27) were LR and 40.7% (11/27) had an early hospital discharge. The LR and early hospital discharge groups had higher mean QoL scores comparatively. Meta‐inferences are reciprocal influencers and expand the complex situation; FN negatively affects the entire family, and the benefits of hospital management were outweighed by risks and worsened symptoms, so an individualized approach to management and care at home was preferred. Conclusion Early discharge of LR FN episodes positively impacts QoL, yet risk‐stratified management for FN is intricately complex. Optimal FN management should prioritize the patient's overall health; shared decision‐making is recommended and can improve care delivery. These results should be confirmed in a larger, more heterogeneous population.


| INTRODUCTION
Febrile neutropenia (FN) is a complication of chemotherapy in children with cancer. 1,2However, not all FN episodes present the same risk for severe outcomes and carefully identified patients can safely receive less aggressive management. 3,46][7][8] Febrile neutropenia guidelines from Children's Oncology Group (COG) suggested six different risk stratification tools for oncologists. 9,10Until recently, research focused on the safety and efficacy of various models and inadequate attention was given to the patient and caregivers' experience. 11hildren's hospitals have recently implemented riskstratified approaches to FN management. 12,13To deliver FN management at home, caregivers and providers agreed a family-oriented support infrastructure is ideal. 14,15aregivers are interested in FN management at home, recognize the value of provider continuity after discharge, 16 and envision improvements in their quality-of-life (QoL) as a result. 15Caregivers favor shared decision-making in FN management and early hospital discharge, but face difficulty during discussions with providers due to the tension between risk stratification value and anecdotal risk perception. 15,17Few qualitative evaluations of FN episodes describe the experiences and burdens of dyads and further in-depth descriptions, especially in a North American cohort, are needed. 15ur institution implemented a risk-stratified management approach for the early hospital discharge of eligible LR FN episodes per international and COG guidelines. 9,18In response to the call for caregivers' inclusion in active FN projects and studies, 14 this report expands the description of the experiences patient/caregiver dyads diagnosed with FN by the novel utilization of a mixed methods survey instrument.

| Study design and planning
We previously evaluated a COG recommended riskstratification tool 9,18 modified with additional high-risk (HR) criteria and procalcitonin. 19This observational, convergent mixed methods study used a joint display of integrated mixed methods data collection during study planning, informed by a theoretical framework. 20The joint display table illustrated how the mixed quantitative and qualitative data from medical chart abstraction and the mixed methods survey instrument were conceptually and thematically related.The study received ethical approval from the University of Michigan institutional review board.This report follows the standards for reporting qualitative research guidelines. 21

| Setting and definitions
Fever neutropenia was defined as an absolute neutrophil count (ANC) less than 0.5 K/μL, or 1 K/μL with anticipation that it would decrease; and a fever of ≥38°C sustained over 1 h or a single temperature >38.3°C. 22Our tertiary, referral center follows COG guidelines in the treatment of fever neutropenia by administration of empiric antipseudomonal parental antibiotics. 9A COG recommended risk tool 18,23 modified with more risk criterion and serial procalcitonin was used. 19Described in a prior publication, children were eligible for early hospital discharge after one night of hospital observation if they met the LR criteria; the decision for early discharge was shared between the patient/caregiver and medical team. 19Episodes were HR if they did not meet LR criteria and continued to receive standard admission. 19(eTable 1).

| Mixed methods survey instrument development and participants
The original survey instrument was modified 6 so the 13 quantitative items shared a common Likert scale of "far below average" to "far above average," and 8 qualitative items were added to thematically expand upon the quantitative items (eTable 2).General oncology dyads under age 21 were eligible after hospital discharge for FN where the first fever was observed in the outpatient setting approached via phone or in person during the study period.Over a 5-month period (January-May 2022) and early in the implementation of FN risk stratification, 31 patient/ caregiver (1:1) dyads were approached 7 days after hospital discharge with a response rate of 87%; dyads were instructed to collaboratively respond on their most recent FN episode.Recruitment continued to thematic saturation (n = 28), the point at which continued data collection stopped adding complexity for codes or themes. 24Verbal consent was obtained by standardized script.Surveys were emailed or administered by research assistants via a virtual platform (Qualtrics, Provo, UT).

| Mixed methods survey instrument analysis
The theoretical framework and joint display of integrated data collection 20 informed identification of reoccurring themes by an open coding method. 25Discrepancies were discussed, then the final coded transcript was uploaded into NVivo Pro 11 (QSR International, Melbourne, Australia).A thematic analysis 24,25 of the qualitative data was conducted.Categorical variables were summarized by means, while coefficient alpha 26 described the internal consistency reliability of quantitative items using Excel software v.2209.

| Medical chart analysis
REDCap secure database 27 stored the medical chart abstraction data.Descriptive statistical analysis was performed using Excel software v.2209.The sample size was determined by the estimated thematic saturation point for the mixed methods survey instrument.Therefore, a bivariate power analysis was not performed.

| Mixed methods analysis
After independent analyses of each data type, a joint display analysis 28 integrated both types of data to determine agreement and discrepancy, guided and informed by the integrated joint display of data collection. 20Joint displays of results illustrated the data, demonstrated linkage between data type and origin, and then interpretation produced meta-inferences and significance for each theme.A joint display figure was generated (Draw.iosoftware, v20.7.4) to depict the complex relationships of results in the mixed methods survey instrument.

| Quantitative results
Quantitative results were analyzed from two separate data sources: medical chart abstraction and mixed methods survey items.

| Medical chart abstraction
Analysis for the 27 FN episodes identified 51.8% (14/27) as LR episodes and 48.1% (13/27) as HR episodes with an early hospital discharge in 40.7% (11/27) of FN episodes (eTable 3 and eTable 4).Of the 11 early hospital discharges, 81.8% (9/11) were LR and 18.2% (2/11) were HR; there were 0 intensive care or blood stream infections in this group.The FN episodes in the standard admission group included 31.3% (5/16) LR episodes and 68.8% (11/16) HR episodes, with a median length-of-stay of 3 days (range 1-8 days).The standard admission group had one blood stream infection and one intensive care admission observed in separate HR FN episodes.

| Quantitative mixed method survey items
Results from the quantitative items of the mixed methods survey instrument are depicted in eTable 5. Coefficient alpha 26 was 0.782.The mean score for LR FN group (2.8) was close to the HR FN group (2.7), but the mean score for the early hospital discharge group (3) was higher than the standard admission group (2.6).All groups had mean scores below or far below average in all items referencing their child.The LR FN group kept up better with household tasks compared to the HR FN group (2.7 vs. 2.1) and spent time more time with their partner (2.8 vs. 2.1).Compared to the standard admission group, the early hospital discharge group was more satisfied with their care (4.6 vs. 4.1), kept up with household tasks better (3 vs. 1.9) and described their child as more independent (3.1 vs. 2.3).

| Qualitative results
Thematic analysis yielded qualitative results (n = 28) from responses to the 8 open-ended items in the mixed methods survey instrument. 21,22Responses were generally a few sentences in length.Three main themes were identified: the family unit experiences burden, the preference for home FN management, and the desire for a comprehensive approach to FN management (Table 1).

| Family units experience burden
Participants described that the global strain of caring for a child with cancer was worsened by FN and disrupted T A B L E 1 Thematic analysis of qualitative question items from mixed methods survey instrument.

Theme Subthemes
Family units experience burden

| Mixed method results
Independent analysis of the quantitative and qualitative data occurred before the merger and linkage of data in the integrated joint display tables of analysis (Tables 2-4).
The meta-inferences from the integrated mixed methods analysis are presented in the joint display tables, along with relevant qualitative quotations, quantitative graphs, and meta-inferences.The mixed methods analysis highlighted: (1) FN episodes negatively impact the child's entire family unit; (2) caregivers appreciate the benefits and risks acquired by the treatment location and ascertain that at-home treatment at home would be most beneficial when empowered by the treatment team; (3) pediatric cancer caregivers desire a comprehensive, global approach to FN management for their child and family.A joint display figure was then constructed to improve description of the complex relationships of the facets involved in a riskstratified management approach for FN (Figure 1).The meta-inferences were reciprocally related: the negative familial burden influenced the treatment location which informed the medical management of FN episodes.The dynamic, complex relationships between meta-inferences and the positive and negative factors in FN management are best illustrated in the visual joint display (Figure 1).The visual joint display described factors (e.g., "hospital" and "team") discussed in both negative and positive ways, and factors with either a negative sentiment (e.g., "non-specialty providers" and "harm") or positive sentiment (e.g., "home" and "global treatment").For example, within "treatment location," both favorable and unfavorable aspects of hospital admission were expressed.

| Summary of main findings
The mixed methods survey instrument results indicated that implementation of safe, effective risk-stratification for early discharge of LR FN patients is preferred and benefits patients, caregivers, and their families.The first meta-inference (Table 2) described the significant strain the entire family unit endured during FN episodes.Although their support system helped with logistics, caregivers described significant emotional burden.Further, the second meta-inference (Table 3) demonstrated the preference for home management due to improved patient symptoms, decreased family burden, and reduced exposure to hospital-based infections, which outweighed the benefits of hospital management (e.g., monitoring and access to immediate intervention).Somatic and psychological symptoms were exacerbated in the hospital, and optimal FN management should consider these comorbidities (Table 4).The risk stratified management approach and decision for early hospital discharge by shared decision making is intricately complex as facets of management seem to have dynamic relationships (Figure 1).The results of this report align with current literature.6]29 Another study evaluated the health-related quality-of-life (HRQoL) of patient/caregiver dyads, and correlated decreased HRQoL with FN episode onset. 3016]29 Our findings further support the inclusion of families, caregivers, and overall patient well-being in shared decision-making of FN episodes, 14,16,29 which may improve navigation of the complexities within a risk stratified management approach and individualization of FN management to best meet the child and family needs.
In contrast to our findings, one study's quantitative survey indicated that, despite the perceived improvement in HRQoL for early discharge and outpatient care, caregivers and patients preferred hospital care. 31However, early discharge of LR FN was not offered. 319][30][31]

| Strengths and weaknesses
This study expands current literature by its mixed methods research design, which facilitated illustration of the complex relationships between the positive and negative factors involved in a risk-stratified management approach for FN episodes.Prior studies analyzed either qualitative 14,16,31 or quantitative data, 30 with one study that analyzed both types of data separately without mixed methods analysis, which limited appreciation of the complexities. 15Second, incorporation of patient/caregiver dyads responded to the call to include caregivers in active FN projects and studies, 14 which were considered in the quality improvement study.Furthermore, this study evaluated the influence early discharge of LR FN had on T A B L E 3 Mixed methods analysis of treatment location.dyad experiences after the clinical implementation of risk stratification 18,19 and showed an overall trend of QoL improvement in the LR and early discharge group.9,30 A limitation of this study stems from the sample size of participants, which limited the quantitative analysis of medical data from medical chart abstraction.Further, the observational study design and single-site recruitment may restrict the study population's experiences due to possible selection bias.Survey responses were not proctored, and collaborative responses cannot be guaranteed.Respondents were mainly Caucasian and non-Hispanic, a general characterization of our institution's patient population but fails to represent a diverse cohort.We did not collect socioeconomic status, psychosocial factors, or level of education, which may impact result interpretation.Despite these limitations, our results build upon and align with existing literature which upholds the external reliability of our findings.

| Implications
The results of the mixed methods analysis indicate the multifaceted burden of FN management faced by families is complex and may be improved by a risk-stratified management approach.Early hospital discharge with oral antibiotics should be considered in LR FN management. 3,8,9,32The safe, effective implementation of riskstratified FN management may avoid prolonged hospital admissions for eligible children.
Effective communication between patient/caregiver and medical team would improve care delivery.A comprehensive approach to management was desired and the medical team should facilitate the dialogue of FN treatment strategies, which will positively impact both the patient and caregiver. 14,16Caregivers reported positive experiences about the care received during hospital admission but also desired increased communication about length of stay, medications, and FN treatment options.
The importance of caregiver emotional well-being in the optimization of patient health applies more broadly T A B L E 4 Mixed methods analysis medical management.to cancer care. 33Emotional and resilience-based interventions may enhance caregiver well-being, which reciprocally benefits patients' level of care and health. 33,34Fever neutropenia episodes distress caregivers; however, these emotional difficulties are not unique to FN episodes apply to other caregivers of patients with cancer. 33

| CONCLUSION
Fever neutropenia negatively impacts everyone within the family; despite the benefits of hospital care, caregivers recognized home management improved the burdens associated with the FN episode.While a risk-stratified management approach to FN is complex, an individualized approach with shared decision-making between caregivers and health care professionals will improve care delivery.Although this report focused on FN episodes, the results may apply more broadly; FN episodes represent just one of many arduous aspects of the cancer journey.
Our work highlights the need for further mixed methods descriptions of cancer care to include patients and caregivers in management decisions.Improved characterization of patient and family burden would influence shared decision-making.Future investigations may consider further evaluation of the subgroups by risk stratification (LR to HR) and disposition (early discharge to standard hospitalization) in a larger heterogeneous population.

T A B L E 2
Mixed methods joint display of family burden.

F I G U R E 1
Visual joint display of factors discussed in mixed method survey instrument.Overlapping colors indicate factors with multiple sentiments (i.e., close monitoring was discussed both negatively and positively within the treatment location and medical management constructs).
spare, wants to play.She does not like being contained in a box-like room for days on end.She has a great appetite at home, sleeps well, likes to laugh and be silly.(CG17; HR FN, Standard Admission) (CG 04; LR FN, Standard Admission) Due to an early hospital discharge for a LR FN episode, one caregiver highlighted family members were not negatively impacted: Because of this new rule, we did not have to stay in the hospital very long which was great!(CG 05; LR FN, Early When my child is admitted to the hospital, a lot of her activities that she likes to do get cut off.At home she is sunshine and rainbows, she has energy to Caregivers desired individualized care during FN episodes, the value of antipyretics, and preference for an oncologist at presentation: There needs to be different treatments for children who present with neutropenia fevers