Summary:
Children with cancer have high emergency department (ED) utilization, but little is known about their chief complaints. A retrospective chart review of ED chief complaints for children with cancer (actively receiving therapy) at Riley Hospital for Children from January 2014 to December 2015 was performed. Proportions of visits and disposition for top 5 chief complaints were determined. Multivariate logistic regression analyzed factors associated with admission. There were 598 encounters by 231 children with cancer. About half (49%) had > 1 complaint. The 5 most common primary chief complaints were: fever (60.2%), pain (6.5%), nausea/vomiting (5.0%), bleeding (3.9%), and abnormal laboratory values (3.3%). Admission rates varied, with the highest rates being for nausea/vomiting (66.7%). Risk factors for admission were: hospitalization in prior 4 weeks (odds ratio [OR], 2.67; confidence interval [CI], 1.77-4.02), chief complaint of fever (OR, 1.90; CI, 1.16-3.09). For each increase in number of chief complaints, odds increased by 1.45 (CI, 1.14-1.83). Black, non-Hispanic (OR, 0.44; CI, 0.22-0.88) as compared with white, non-Hispanic, younger age (OR, 0.53; CI, 0.29-0.99) or complaint of abnormal laboratory values (OR, 0.20; CI, 0.06-0.68) had lower odds of admission. Children with cancer present to the ED with multiple and varied complaints. Future interventions could aim to improve caregiver anticipatory guidance and ED visit preparedness.
Keywords: health outcomes, health care utilization, childhood cancer, emergency department
BACKGROUND
Emergency department (ED) utilization is higher among children with cancer than the general pediatric population.1,2 Children with cancer represent a unique population in terms of the reasons for seeking care in the ED. Previous publications have evaluated the reasons for ED visits by children with cancer via assessment of International Classification of Diseases, Ninth Revision, Clinical Modification discharge diagnoses.3,4 Unfortunately, discharge diagnoses are not necessarily representative of the chief complaints that brought the child with cancer to the ED. A classic example of the difference between chief complaints and discharge diagnoses is if an adult presents with a complaint of chest pain, they may be diagnosed with a potentially fatal heart attack or a benign condition such as gastroesophageal reflux.5 None of the currently available databases of ED utilization among children with cancer contain chief complaints, only discharge diagnoses, leading to a gap in our understanding of the complaints that prompt children with cancer to seek care in the ED.
Therefore, the purpose of this study is to identify the spectrum of chief complaints among children with cancer presenting to a free-standing children’s hospital ED and evaluate factors associated with admission among this population. Our hypotheses were that children with cancer present to the ED with > 1 chief complaint, with fever being the most common complaint and most associated with admission. Our investigation will lead to a better understanding of the signs and symptoms that cause children with cancer to seek care in the ED.
METHODS
Study Design and Setting
We performed a retrospective chart review of ED encounters for children with cancer at the Riley Hospital for Children Emergency Department at Indiana University Health from January 1, 2014 to December 31, 2015. The Riley Hematology-Oncology Clinic at Indiana University Health treats 80% of all new pediatric cancer diagnoses in Indiana and sees an average of 200 new cancer patients each year. Riley Hospital’s Emergency Medicine and Trauma Center is a Level 1 Pediatric Trauma Center that provides care for over 34,000 children annually.
Study Population/Database Development
In order to create a database of all children with cancer seen at the Riley ED, we first obtained ED visit logs for all ED visits to the Riley Hospital for Children from the electronic medical record. Second, children with cancer were extracted using a clinical patient database maintained by the pediatric oncology clinical research office. Inclusion criteria for this study required an ED visit by a child with cancer actively receiving treatment, including either chemotherapy and/or radiation. ED encounters were excluded if the visit was for a new diagnosis of cancer, a relapse of cancer, or occurred before the cancer diagnosis. We then created a database of ED visits by children with cancer by extracting patient characteristics and ED encounter information from the clinical patient database and the electronic medical record.
Patient demographic characteristics included sex, age category, race/ethnicity (white, non-Hispanic; Black, non-Hispanic; Hispanic; and Other), primary payer (public/governmental, private, self-pay). Patients were classified into one of the following types of cancer based on electronic medical record data: acute lymphoblastic leukemia (ALL), acute myelogenous leukemia, central nervous system (CNS) tumors, solid tumors (non-CNS), Hodgkin lymphoma, and non-Hodgkin lymphoma. Patients with nonspecified malignancies or rare cancers were placed into an “Other” category. Encounter characteristics included chief complaints (including a primary chief complaint and then all secondary complaints), whether or not there was an inpatient encounter within the prior 4 weeks, ED disposition (admit vs. discharge), and which unit the patient was admitted to (floor status vs. pediatric intensive care unit). We chose to include an inpatient encounter within the prior 4 weeks because readmissions have been the focus of quality improvement and cost containment in the era of the Patient Protection and Affordable Care Act.6 We chose not to evaluate the reason for the previous inpatient visit since our focus was primarily on the reasons for ED visits.
There were 41 encounters with missing primary payer information, which accounted for about 7% of encounters; these were included in all analyses except the logistic regression.
Outcome and Exploratory Variables
We reviewed all listed chief complaints and those with a similar clinical focus were condensed into chief complaint categories. Each encounter could have > 1 complaint associated with it. The highest number of complaints found on chart review were 5 complaints, so that was the maximum we collected for each encounter. For the logistic regression, the primary outcome of interest was whether an ED visit resulted in admission to Riley Hospital for Children versus discharged to home (patient treated then released from the ED).
Statistical Analyses
We summarized encounter characteristics using frequencies and percentages. We compared characteristics of those who were admitted versus discharged using χ2 analyses. We evaluated the number of ED visits per patient and the number of total chief complaints per visit. We separated primary chief complaints from secondary complaints and generated a rank list for each. Thereafter, the top 5 primary chief complaints were presented as proportion of visits, both overall and by type of cancer. Proportions of hospitalization rates were calculated, both overall and by the top 5 primary chief complaints.
A multivariate logistic regression model was used to estimate factors associated with admission for children with cancer. Variables were included based on our defined model: patient’s sex, age category, race/ethnicity, primary expected payer, whether there was an inpatient visit in the prior 4 weeks, the number of chief complaints as a continuous variable, and dichotomous variables for the presence or absence of each of the top 5 most common primary chief complaints. All analyses were performed using SAS 9.4 (Carey, NC), and P-values <0.05 were considered statistically significant.
RESULTS
Overall, there were 598 encounters by 231 children with cancer during the 2-year time frame investigated. Only 40.3% of children had just 1 ED visit, 23.8% had 2 visits, 12.1% had 3 visits, and 23.8% had ≥ 4 visits. We present the encounter level characteristics of the ED visits among children with cancer in Table 1, both overall and by ED disposition status. There were significant differences between age groups who were admitted compared with discharge, with higher proportions of younger patients discharged. Similarly, a higher proportion of black, non-Hispanic patients were discharged. About half of visits were by children with public/governmental primary insurance payer (51.9%). Other common characteristics included a diagnosis of ALL (41.3%) and having been hospitalized in the prior 4 weeks (53.2%). Among these encounters, 43.5% resulted in discharge to home and 56.5% were admitted to the hospital. Among the admissions, 93.8% were admitted to the inpatient floor and 6.2% were admitted to the pediatric intensive care unit.
TABLE 1.
Encounter Level Characteristics of Emergency Department Visits for Children With Cancer
| n (%) |
||||
|---|---|---|---|---|
| Overall N = 598 |
Admitted N = 338 (56.5) |
Discharged N = 260 (43.5) |
P | |
| Sex | ||||
| Female | 240 (40.1) | 138 (40.8) | 102 (39.2) | 0.69 |
| Male | 358 (59.9) | 200 (59.2) | 158 (60.8) | |
| Age category (y) | ||||
| 0-4 | 223 (37.3) | 118 (34.9) | 105 (40.4) | 0.040 |
| 5-9 | 172 (28.8) | 89 (26.3) | 83 (31.9) | |
| 10-14 | 108 (18.1) | 68 (20.1) | 40 (15.4) | |
| 15-20+ | 95 (15.9) | 63 (18.6) | 32 (12.3) | |
| Race/ethnicity | ||||
| White, non-Hispanic | 450 (75.3) | 258 (76.3) | 192 (73.9) | 0.003 |
| Hispanic | 77 (12.9) | 47 (13.9) | 30 (11.5) | |
| Black, non-Hispanic | 55 (9.2) | 20 (5.9) | 35 (13.5) | |
| Other | 16 (2.7) | 13 (3.9) | 3 (1.2) | |
| Primary payer* | ||||
| Public/governmental | 289 (51.9) | 153 (49.2) | 136 (55.3) | 0.30 |
| Private | 258 (46.3) | 153 (49.2) | 105 (42.7) | |
| Self-pay | 10 (1.8) | 5 (1.6) | 5 (2.0) | |
| Type of cancer | ||||
| Acute lymphoblastic leukemia | 247 (41.3) | 130 (38.5) | 117 (45.0) | 0.37 |
| Solid tumor | 206 (34.5) | 127 (37.6) | 79 (30.4) | |
| Central nervous system tumor | 96 (16.1) | 53 (15.7) | 43 (16.5) | |
| Non-Hodgkin lymphoma | 20 (3.3) | 14 (4.1) | 6 (2.3) | |
| Hodgkin lymphoma | 19 (3.2) | 10 (3.0) | 9 (3.5) | |
| Acute myelogenous leukemia | 5 (0.8) | 2 (0.6) | 3 (1.2) | |
| Other | 5 (0.8) | 2 (0.6) | 3 (1.2) | |
| Inpatient encounter in prior 4 wk | ||||
| Yes | 318 (53.2) | 207 (61.2) | 111 (42.7) | < 0.001 |
| No | 280 (46.8) | 131 (38.8) | 149 (57.3) | |
| Admit unit | ||||
| Floor | 317 (93.8) | |||
| Pediatric ICU | 21 (6.2) | |||
Bold value indicates statistically significant findings.
There were 41 encounters with missing primary payer information. The remainder of the variables were complete for each encounter.
ICU indicates intensive care unit.
ED Chief Complaints
About half of children with cancer (51%) had only 1 chief complaint while 30.9% had 2 complaints (Fig. 1). The 5 most common primary chief complaints prompting an ED visit for children with cancer are presented in Table 2. Fever was the most common primary chief complaint (60.4%); followed by pain (6.1%), nausea/vomiting (5.1%), bleeding (3.9%), and abnormal laboratory values (3.4%). When we evaluated the chief complaints by type of cancer, fever remained the most common ED visit chief complaint for all types of patients; ALL (66.0%), solid tumor (60.2%), CNS tumor (44.8%), non-Hodgkin lymphoma (70.0%), Hodgkin lymphoma (57.9%), and acute myelogenous leukemia (60.0%). The top 5 secondary complaints included: cough (16.4%), nausea/vomiting (11.9%), pain (8.2%), rhinorrhea (6.9%), and headache (4.2%).
FIGURE 1.
Number of chief complaints for children with cancer per ED encounter. ED indicates emergency department.
TABLE 2.
Top Primary Chief Complaints Prompting ED Visits Among Children With Cancer—Overall and by Cancer Type
| By Cancer Type (%) |
||||||||
|---|---|---|---|---|---|---|---|---|
| Overall (N = 598) (%) |
ALL (N = 247) |
Solid Tumor (N = 206) |
CNS (N = 96) |
NHL (N = 20) |
HL (N = 19) |
AML (N = 5) |
||
| Rank | ||||||||
| 1 | Fever | 60.2 | 66.0 | 60.2 | 44.8 | 70.0 | 57.9 | 60.0 |
| 2 | Pain | 6.5 | 6.1 | 6.8 | 4.2 | 5.0 | 10.5 | 0 |
| 3 | Nausea/vomiting | 5.0 | 4.1 | 5.3 | 8.3 | 5.0 | 0 | 0 |
| 4 | Bleeding | 3.9 | 3.2 | 5.3 | 4.2 | 0 | 0 | 0 |
| 5 | Abnormal Laboratory values | 3.3 | 2.8 | 2.4 | 6.3 | 10.0 | 0 | 0 |
ALL indicates acute lymphoblastic leukemia; AML, acute myelogenous leukemia; CNS, central nervous system; HL, Hodgkin lymphoma; NHL, non-Hodgkin lymphoma.
Admission Rates by Chief Complaint
Overall, 56.5% (N = 338) of ED encounters resulted in an inpatient admission. Admission rates varied by the top 5 primary chief complaints with nausea and vomiting having the highest (66.7%) and abnormal laboratory values with the lowest (20%) (Fig. 2).
FIGURE 2.
Emergency department admission rates among children with cancer, overall and by top 5 primary chief complaints.
Factors Affecting Admission From the ED Among Children With Cancer
In a multivariate analysis shown in Table 3, factors associated with significantly increased odds of admission included having a chief complaint of fever (odds ratio [OR], 1.90, confidence interval [CI], 1.16-3.09) and having been hospitalized in the prior 4 weeks (OR, 2.67; 95% CI, 1.77-4.02) compared with not having a recent admission. For each increase in the number of chief complaints, the likelihood of admission increased by 1.45 (95% CI, 1.14-1.83). Younger patients (age, 0 to 4 y) had lower odds of admission (OR, 0.53; CI 0.29-0.99) than older children (age, 15 to 20+ years). Black, non-Hispanic patients were less likely to be admitted (OR, 0.44; CI, 0.22-0.88) as compared with white, non-Hispanic. Patients with a primary chief complaint of abnormal laboratory values also had significant decreased odds of admission (OR, 0.20; CI, 0.06-0.68).
TABLE 3.
Multivariate Logistic Regression to Evaluate Factors Associated With Admission Versus Discharge From the ED Among Pediatric Cancer Patients
| Factors | Adjusted Odds Ratio (OR) | 95% CI | P |
|---|---|---|---|
| Patient characteristics | |||
| Sex | |||
| Female | 1.04 | 0.70-1.56 | 0.84 |
| Age (y) | |||
| 15-20+ | Reference | ||
| 10-14 | 1.01 | 0.50-2.0 | 0.99 |
| 5-9 | 0.55 | 0.29-1.06 | 0.07 |
| 0-4 | 0.53 | 0.29-0.99 | 0.046 |
| Race/ethnicity | |||
| White, non-Hispanic | Reference | ||
| Black, non-Hispanic | 0.44 | 0.22-0.88 | 0.021 |
| Hispanic | 1.76 | 0.97-3.21 | 0.07 |
| Other | 2.87 | 0.73-11.28 | 0.13 |
| Type of cancer | |||
| Acute lymphoblastic leukemia | Reference | ||
| Acute myelogenous leukemia | 0.52 | 0.07-4.02 | 0.53 |
| Solid tumors | 1.31 | 0.81-2.12 | 0.28 |
| Central nervous system tumor | 1.51 | 0.81-2.81 | 0.20 |
| Hodgkin lymphoma | 0.69 | 0.22-2.14 | 0.52 |
| Non-Hodgkin lymphoma | 1.74 | 0.56-5.45 | 0.34 |
| Other | 0.53 | 0.07-4.06 | 0.54 |
| Primary payer | |||
| Public/governmental | Reference | ||
| Private | 1.07 | 0.69-1.64 | 0.77 |
| Self | 0.71 | 0.17-2.86 | 0.63 |
| ED visit characteristics | |||
| No. chief complaints | |||
| 1.45 | 1.14-1.83 | 0.002 | |
| Top 5 primary chief complaint categories (in rank order) | |||
| Fever | 1.90 | 1.16-3.09 | 0.010 |
| Pain | 1.78 | 0.78-4.05 | 0.17 |
| Nausea or vomiting | 1.88 | 0.72-4.91 | 0.20 |
| Bleeding | 0.49 | 0.17-1.41 | 0.19 |
| Abnormal laboratory values | 0.20 | 0.06-0.68 | 0.010 |
| Inpatient in the last 4 wk | |||
| Yes | 2.67 | 1.77-4.02 | < 0.001 |
Bold indicates significant factors associated with admission versus discharge.
CI indicates confidence interval; ED, emergency department; OD, odds ratio.
DISCUSSION
In this retrospective study, children with cancer who are actively receiving therapy not only present to the ED with a wide range of complaints, but also tend to have > 1 complaint. The most common ED chief complaint was fever, followed by pain, nausea/vomiting, bleeding, and abnormal laboratory values. Importantly, the majority of ED visits resulted in admission, with the highest rates of admission for a chief complaint of nausea/vomiting or fever. Factors associated with increased odds of admission included a chief complaint of fever, having a recent hospitalization, and having > 1 complaint. This analysis highlights the undesired, but not unexpected side effects of cancer therapies leading to ED visits for children with cancer. This information provides specific focus areas for future interventions aimed at providing caregivers of children with cancer improved anticipatory guidance and preparedness when an ED visit is necessary.
Fever is a significant ED complaint and discharge diagnosis for children with cancer with increased odds of admission.3,4 Unfortunately, in the field of pediatric oncology we lack agreement on the exact definition of a fever.7,8 This leads to difficulties in creating evidence-based clinical practice guidelines and differences in triaging and management practices, as demonstrated in a previous study of pediatric oncology providers in Michigan.9 In order to unify our approach and move towards implementation of validated risk prediction models to evaluate the risk of serious infection in the setting of fevers,10,11 a consensus on the definition of fever among children with cancer would be beneficial. At a minimum, it would be prudent for each institution to develop a clinical practice guideline with an agreed upon definition of fever in order to decrease confusion among patients, caregivers, providers, and staff.
Chemotherapy-induced nausea and vomiting (CINV) was the third major chief complaint that brought children with cancer to the ED. Inadequate control of CINV are among the most severe and bothersome side effects experienced by children with cancer according to their parents.12 Several recent updates on guidelines for CINV outline key changes in recommendations for both acute, anticipatory, breakthrough and refractory CINV.13,14 An important consideration for CINV is to understand whether these new evidence-based recommendations are being implemented by oncologists in the clinical setting. Moreover, attention should focus on the ability of caregivers of children with cancer to manage CINV in the home setting. A lack of caregiver education or appropriate at-home medications for breakthrough nausea will inhibit optimal supportive care.
Pain is not only an unfortunately common experience for children with cancer,15,16 but also the one which leads to ED visits. Pain can come from a variety of sources for this unique population; from complications of their malignancy (ie, bone pain, tumor compression), postoperative pain, and the pain experienced during their treatment (ie, central line accessing, mucositis, neuropathy). Therefore, pain assessment has been identified as an important quality measure for the outpatient care of children with cancer.17 It is essential that we not only assess pain in this unique population, but also provide sufficient anticipatory guidance for caregivers along with at-home prescription medications. Improved at-home pain management strategies could decrease ED reliance and improve quality of life.
Secondary complaints of cough and rhinorrhea could represent a simple viral upper respiratory infection or be signs of a more serious bacterial or fungal infection such as sinusitis or pneumonia. Recently published guidelines for management of fevers in children with cancer highlight that further research is needed to evaluate if viral processes increase the risk for serious infections in this population.8 Of note, in a previous analysis of discharges for fever and neutropenia, encounters with a short length of stay (≤ 3d) were associated with a discharge diagnosis of a viral illness.18,19 This suggests that if patients with cough or rhinorrhea are admitted for evaluation and no serious bacterial infection is found, they are more likely to be discharged after only a few days. Further evaluation could investigate if there are associated signs or symptoms with rhinorrhea and cough that could aid in outpatient triaging and decrease ED visits.
Limitations
There are several important limitations to our study. First, this was a single institution study, but we have ~200 new diagnoses of childhood cancer per year at Riley Hospital for Children and are similar to many other large, quaternary, free-standing children’s hospitals. Second, due to the large geographical spread of our patient population across the entire state of Indiana, this analysis may not represent all ED visits by our patient population as those who live far from Riley Hospital may have received care at local hospitals. Third, our goal was to explore chief complaints as documented in the medical record and we did not evaluate the highest documented or observed temperatures, thus the definition of “fever” was based on chief complaint report only. Moreover, we appreciate that admissions to the intensive care unit represent a higher level of clinical severity, but we did not have a large enough sample to evaluate differences in factors of admission to the intensive care unit versus the inpatient floor. Moreover, we did not explore factors associated with whether the admissions were considered “unplanned,” but this line of quality improvement could be the focus of future investigations. Nevertheless, this study represented a large number of ED visits to our institution from a wide variety of cancer types over a 2-year time period.
CONCLUSIONS
In this retrospective review of ED visits among children with cancer who are actively receiving therapy, we revealed that children with cancer not only present to the ED with a wide range of complaints, but also tend to have > 1 complaint. The 5 most common complaints for ED visits were fever, pain, nausea/vomiting, bleeding, and abnormal laboratory values. Factors associated with increased odds of admission included a chief complaint of fever and having > 1 complaint. Future evaluations should focus on the adequacy of current management strategies through incorporation of patient and caregiver perspectives.
Footnotes
The authors declare no conflict of interest.
REFERENCES
- 1.O’Sullivan C, Dupuis LL, Gibson P, et al. Refinement of the symptom screening in pediatrics tool (SSPedi). Br J Cancer. 2014;111:1262–1268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.National Center for Health Statistics. Health, United States, 2015: With Special Feature on Racial and Ethnic Health Disparities. Hyattsville, MD. 2016. [PubMed] [Google Scholar]
- 3.Mueller EL, Sabbatini A, Gebremariam A, et al. Why pediatric patients with cancer visit the emergency department: United States, 2006-2010. Pediatr Blood Cancer. 2015;62:490–495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mueller EL, Hall M, Shah SS, et al. Characteristics of children with cancer discharged or admitted from the emergency department. Pediatr Blood Cancer. 2016;63:853–858. [DOI] [PubMed] [Google Scholar]
- 5.Chan S, Maurice AP, Davies SR, et al. The use of gastrointestinal cocktail for differentiating gastro-oesophageal reflux disease and acute coronary syndrome in the emergency setting: a systematic review. Heart Lung Circ. 2014;23:913–923. [DOI] [PubMed] [Google Scholar]
- 6.Centers for Medicare and Medicaid Services. Readmissions Reduction Program 2013. Available at: www.cms.gov/Medicare/Medicare-Fee-for-ServicePayment/AcuteInpatientPPS/Readmissions-Reduction-Program.html. Accessed March 1, 2018.
- 7.Haeusler GM, Phillips RS, Lehrnbecher T, et al. Core outcomes and definitions for pediatric fever and neutropenia research: A consensus statement from an international panel. Pediatr Blood Cancer. 2015;62:483–489. [DOI] [PubMed] [Google Scholar]
- 8.Lehrnbecher T, Robinson P, Fisher B, et al. Guideline for the management of fever and neutropenia in children with cancer and hematopoietic stem-cell transplantation recipients: 2017 update. J Clin Oncol. 2017;35:2082–2094. [DOI] [PubMed] [Google Scholar]
- 9.Mueller EL, Walkovich KJ, Yanik GA, et al. Variation in management of fever and neutropenia among pediatric patients with cancer: a survey of providers in michigan. Pediatr Hematol Oncol. 2015;32:331–340. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Haeusler GM, Thursky KA, Mechinaud F, et al. Predicting infectious complications in children with cancer: an external validation study. Br J Cancer. 2017;117:171–178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Phillips RS, Sung L, Amman RA, et al. Predicting microbiologically defined infection in febrile neutropenic episodes in children: global individual participant data multivariable meta-analysis. Br J Cancer. 2016;114:623–630. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dupuis LL, Milne-Wren C, Cassidy M, et al. Symptom assessment in children receiving cancer therapy: the parents’ perspective. Support Care Cancer. 2010;18:281–299. [DOI] [PubMed] [Google Scholar]
- 13.Flank J, Robinson PD, Holdsworth M, et al. Guideline for the treatment of breakthrough and the prevention of refractory chemotherapy-induced nausea and vomiting in children with cancer. Pediatr Blood Cancer. 2016;63:1144–1151. [DOI] [PubMed] [Google Scholar]
- 14.Patel P, Robinson PD, Thackray J, et al. Guideline for the prevention of acute chemotherapy-induced nausea and vomiting in pediatric cancer patients: a focused update. Pediatr Blood Cancer. 2017;64. [DOI] [PubMed] [Google Scholar]
- 15.Dupuis LL, Sung L, Molassiotis A, et al. 2016 updated MASCC/ESMO consensus recommendations: prevention of acute chemotherapy-induced nausea and vomiting in children. Support Care Cancer. 2017;25:323–331. [DOI] [PubMed] [Google Scholar]
- 16.Van Cleve L, Bossert E, Beecroft P, et al. The pain experience of children with leukemia during the first year after diagnosis. Nurs Res. 2004;53:1–10. [DOI] [PubMed] [Google Scholar]
- 17.Teichman J, Punnett A, Gupta S. Development of quality metrics to evaluate pediatric hematologic oncology care in the outpatient setting. J Pediatr Hematol Oncol. 2017;39:90–96. [DOI] [PubMed] [Google Scholar]
- 18.Mueller EL, Croop J, Carroll AE. Fever and neutropenia hospital discharges in children with cancer: a 2012 update. Pediatr Hematol Oncol. 2016;33:39–48. [DOI] [PubMed] [Google Scholar]
- 19.Mueller EL, Walkovich KJ, Mody R, et al. Hospital discharges for fever and neutropenia in pediatric cancer patients: United States, 2009. BMC Cancer. 2015;15:388. [DOI] [PMC free article] [PubMed] [Google Scholar]