Management of Febrile Neutropenias in Adolescents and Young Adults: A Cost-Minimization Analysis Between Adult Versus Pediatric Units

Abstract

Purpose:

Management of adolescents and young adults (AYAs) differs between adult and pediatric units, especially regarding febrile neutropenia (FN). In our previous study, we found that AYAs treated in adult units were significantly less hospitalized for FN than in pediatric units, without difference in morbimortality. The objective of this work was to assess the economic impact of these practices.

Methods:

This study retrospectively collected data from the medical records of AYAs treated at the Comprehensive Cancer Center Léon Bérard, Lyon, France, in the Euro-E-W-I-N-G99 protocol between September 1, 2000 and May 31, 2013. We focused on FN occurring after VIDE (vincristine, ifosfamide, doxorubicin, etoposide) courses. Costs were calculated using a micro-costing technique from the hospital's perspective (in 2014-Euro); the time horizon was the induction period. Multivariate analyses were performed on the total cost and cost of FN. Uncertainty was captured by sensitivity analyses.

Results:

Forty-four AYAs (18 in the adult sector, 26 in the pediatric sector) received 260 courses of VIDE. Mean cost of care was €37,544 in the pediatric sector, including €11,948 (32%) for FN (€11,851 in hospitalization), versus €34,677 in the adult sector, including €6,143 (18%) for FN (€5,789 in hospitalization). Cost for FN was significantly higher in pediatric units (difference in mean cost of €5,830 per patient, 95% bootstrapped confidence interval [1,939.1; 10,028.9]). In multivariate analysis, the only factor significantly influencing this cost difference was the sector of care. The most sensitive parameter was the unit cost of conventional hospitalization.

Conclusion:

These results support the adult sector strategy, in agreement with the results of our first work showing comparable effectiveness.

Introduction

In France, about 1700 new cases of cancer occur each year in adolescents and young adults (AYA, 15–25 years).¹ The management of these patients includes biological, pharmacological, emotional, psychological, and social specificities.² Disparities in survival have been reported, with a worse prognosis compared to younger or older patients, for identical diseases, especially for sarcomas and leukemia.³ Dedicated care programs were set up in 2012 in France to promote optimal management of these patients, improving collaboration between adult and pediatric oncologists.

Febrile neutropenia (FN) is the second cause of hospitalization in pediatric oncology for patients during treatment.⁴ We noted in our center that the management of FN was different between the units. After initial evaluation by their physician or in the emergency unit, adults are most often treated at home, while patients treated in the pediatric unit are routinely hospitalized for a minimum of 3 or 4 days. In a previous study, we showed that treatment in adult unit is as safe as in pediatric unit for these episodes, since we did not observe any difference in morbidity and mortality between the two strategies.⁵ In particular, there was no significant difference in survival, number of documented infections, transfusions, dose modifications, chemotherapy delay, and need for intensive care.⁵ These results were concordant with the data reported in the literature, which showed no difference in effectiveness between homecare and hospitalization for low-risk FN.^6–10

As the two options are equally effective, the objective of this study was to perform a cost-minimization analysis measuring the difference in costs between both options.

Methods

Study population

Patients aged 15–25 years, treated at the Comprehensive Cancer Centre Léon Bérard, Lyon, France, either in an adult or a pediatric unit for Ewing's sarcoma in the Euro-E.W.I.N.G.99 protocol (EE99) between September 1, 2000 and May 31, 2013 were retrospectively included.⁵ The study focused on the management of FN occurring in the induction phase, delimited between the first medical contact in our center and the day of post-induction surgery, or the first day of adjuvant chemotherapy for inoperable patients. This period consists in six VIDE courses (vincristine, ifosfamide, doxorubicin, etoposide).¹¹ Per protocol, all patients were supposed to receive granulocyte colony-stimulating factors (G-CSF) after each course.

In our center, AYAs have the choice to be hospitalized either in an adult or a pediatric unit whatever their age, depending on their psychosocial profile. The caregivers of the mobile unit dedicated to AYAs see them in their rooms. In both sectors, aplasia is defined by blood neutrophil counts below 0.5 g/L. Fever is defined by temperature measured once greater than 38.5°C, or twice between 38°C and 38.5°C. However, the management of FN is different between pediatric and adult units.

In pediatrics, blood count is monitored biweekly. In case of fever, AYAs are routinely hospitalized for complete evaluation and empiric antibiotic therapy. If FN is not documented, antibiotics are usually stopped after 24 hours of apyrexia, and patients are discharged 24 hours later. In case of bacteriological documentation, antibiotics and length of treatment are adapted.

In the adult sector, blood count is monitored weekly. In case of FN, patients are examined in the emergency center and discharged if the MASCC score (Multinational Association of Supportive Care in Cancer) is greater than 21 (defining low risk of complications); they receive intravenous (IV) antibiotics at home.

Due to the retrospective nature of our study, we could not consider quality of life or patients' preferences.¹² However, we concluded in a previous work that there was no significant difference between the two strategies, in terms of effectiveness.⁵

The study was conducted in accordance with the ethical principles for medical research involving human subjects developed according to Declaration of Helsinki principles by the World Medical Association. The study received approval in France from the National Committee for Protection of Personal Data (no. 1860448). Full details of the study design have been published elsewhere.⁵

Cost assessment

Cost-minimization analysis is a method for the economic evaluation of healthcare programmes,¹³ in which relative costs of strategies showing equivalent outcomes are simply compared.¹⁴ On the basis of our previous results,⁵ we assumed that the effectiveness of the two strategies (adult vs. pediatric unit) was equivalent. Therefore, a cost-minimization analysis was conducted, based strictly on a microcosting approach.^15–17 We performed the study from the hospital perspective.¹⁸

Dates, duration of hospitalizations, and expenses of homecare were collected using the medical records of each patient.

Cost of hospitalization

For each patient, length of stay was multiplied by the daily unit cost, distinguishing conventional hospitalization (CH) and day hospitalization (DH).

Cost of nurse care and drugs at home

Nurse shifts, extra charge for Sunday shifts, all nursing cares (blood collections on central venous access, subcutaneous [SC] or IV injections, wound care, setting up, monitoring and withdrawal of infusion, and change and monitoring of patient-controlled analgesia devices), and all the SC and IV medications were considered. We did not consider other paramedical care (physical, occupational, and physiotherapies), or oral medication because our retrospective study did not permit to be comprehensive. The nursing care and drugs dedicated to FN management were differentiated from other expenses. The cost of nursing care was calculated using the General Nomenclature of Professional Acts defining the pricing of each act¹⁹; the cost of drugs was calculated from the French reference price according to the website Vidal.fr.²⁰ These elements were considered somewhat variable over time and therefore identical for all patients.

All costs were expressed in 2014 Euros and were not discounted.

Statistical analyses

Consumptions and costs (hospitalizations, nursing care, and drugs for homecare) were summarized using descriptive statistics. Consumptions were compared between the groups using the nonparametric Mann–Whitney–Wilcoxon test. The mean costs of care (including hospitalizations, nurse care, and drugs at home) and mean costs of FN management were compared between the groups calculating the 95% nonparametric bootstrap confidence interval (CI), using the percentile approach for the mean difference. Thus, 1000 simulated bootstrap samples were generated by independent draws.²¹ Multivariate analyses by linear regression were then performed to compare the total cost and the cost for FN between sectors of care, adjusting the analysis for patients and tumor characteristics. The significance threshold α used for statistical analyses was 5%.

Univariate analyses of deterministic sensitivity as Tornado diagrams were performed to illustrate the change in total cost of care, depending on the variation in unit costs and quantities of resources consumed, highlighting the most sensitive parameters.²² The variation factor used was 20%. Statistical analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC).

Results

Patients and clinical data

Forty-four AYAs were included in the study EE99 during this period: 26 treated in pediatric and 18 in adult units. Patients' characteristics are described in Table 1. The two groups were comparable (location of the primitive tumor, stage, and histological response after the induction phase…) except for age (median 24.1 years for the adult group and 17 for the pediatric group, p < 0.0001). Ten patients older than 18 years were treated in pediatric units.

Table 1.

Characteristics of Patients and Tumors

	Number of patients (%)
Characteristics	Treated in adult units (n = 18)	Treated in pediatric units (n = 26)	p value
Gender			1.000
Female	6 (33)	9 (35)
Male	12 (67)	17 (65)
Age at diagnosis in years			<0.0001
Median (min–max)	24.1 (18.5–25.8)	17.0 (15.1–25.6)
Under 18 years	0 (0)	16 (58)
Localization of tumor			0.553
Abdominal/pelvis/hip bone	5 (28)	7 (27)
Vertebral/paravertebral/sacrum	3 (17)	3 (12)
Thorax/costal	2 (11)	4 (15)
Limb	6 (33)	12 (46)
Head	2 (11)	0 (0)
Stage at diagnosis			1.000
Localized operable	12 (67)	16 (62)
Localized inoperable	2 (11)	4 (15)
Metastatic	4 (22)	6 (23)
Histologic evaluation after induction			0.508
Good response (necrosis >90%)	9 (50)	10 (38)
Poor response (necrosis <90%)	3 (17)	8 (31)
Not evaluable	6 (33)	8 (31)

Of the 260 VIDE courses administered, 122 were complicated by FN (42 occurring in adult units and 80 in pediatric units). We observed a median of 2 FN per patient in adult units and 3 per patient in pediatric units (p = 0.2). Further clinical results published elsewhere showed no significant difference for death (p = 1), need for intensive care (p = 1), documented infection (p = 0.64), post-operative infection (p = 1), chemotherapy delay (p = 0.41), dose reduction (p = 0.40), and number of transfusions (p = 0.46 for red cells and p = 0.94 for platelet transfusions).⁵

Length of hospitalization

The average length of hospitalization was 41.6 ± 17.2 days in the adult sector 48.1 ± 14.9 days in the pediatric sector (Table 2). Average duration of CH for FN was significantly longer in the pediatric sector (p = 0.011) and DH for FN was larger in the adult sector (p = 0.009). The average lengths of stay at hospital for other reasons (diagnosis, insertion of central venous access, chemotherapy, management of pain, and transfusions) were not significantly different between the groups.

Table 2.

Mean Consumption of Resources Depending on the Strategy (Standard Deviation)

	Patients treated in
Cost items	Adult units (n = 18)	Pediatric units (n = 26)	p value
Mean length of hospitalization (in days)
CH for FN	8.0 (8.4)	16.8 (12)	0.011
DH for FN	0.4 (0.8)	0.0	0.009
CH for other reasons	32.3 (11.4)	29.6 (7.1)	0.38
DH for other reasons	0.9 (1.1)	1.7 (1.5)	0.06
Total hospitalization (in days)	41.6 (17.2)	48.1 (14.9)	0.068
Mean number of nurse care acts at home
Number of cares for FN
Nurse shifts	6.8 (8.6)	1.2 (4.2)	0.005
Sunday shifts	1.0 (1.3)	0.3 (1.0)	0.013
Antibiotic infusions	6.8 (8.5)	0.3 (1.1)	0.003
Monitoring/changing of antibiotic PCA	0.0	0.9 (7.6)	0.003
Number of acts for other reason
Nurse shifts	63.1 (30.1)	43.6 (12.6)	0.033
Sunday shifts	5.8 (3.1)	4.0 (2.8)	0.13
Injection of G-CSF (SC)	24.8 (15.7)	29.1 (11.8)	0.65
Infusion of antiemetics (IV)	9.7 (25.5)	0.0	0.033
Infusion of anticoagulants (SC)	5.3 (27.7)	4.2 (14.7)	0.71
Blood collection on CVA	11.3 (2.8)	17.9 (5.1)	<0.001
Wound care	0.0	1.2 (3.8)	0.09
Setting up/withdrawal of infusion	28.8 (28.8)	0.2 (0.8)	<0.001
Monitoring/changing of morphine PCA	0.9 (5.6)	0.5 (2.6)	0.52
Drug mean consumption at home
For FN
Ceftriaxone 1 g	10.2 (15.1)	0.6 (2.3)	0.003
Vancomycin 1 g	0.0	0.9 (7.6)	0.41
Ertapenem 1 g	0.7 (3.3)	0.0	0.23
Imipenem 1 g	0.0	2.2 (32.9)	0.41
Cefepime 2 g	1.0 (9.0)	0.0	0.23
For other reasons
Polyionic 2 L	14.4 (14.4)	0.1 (0.8)	<0.001
Pegfilgrastim 6 mg	2.3 (2.8)	1.1 (2.3)	0.73
Lenograstim 34 MUI	15.9 (18.7)	19.1 (17.0)	0.80
Filgrastim 30 MUI	6.7 (15.1)	9.0 (16.2)	0.73
Methylprednisolone 60 mg	2.4 (9.2)	0.0	0.033
Ondansetron 8 mg	7.2 (24.9)	0.0	0.033
Tinzaparin 10,000 UI	5.3 (27.7)	2.0 (11.1)	0.71
Enoxaparin 2000 UI	0.0	2.2 (18.7)	0.41
Morphine 100 mg	0.9 (11.3)	0.5 (2.6)	0.77

Bold values indicate significant difference between the 2 groups.

CH, conventional hospitalization; CVA, central venous access; DH, day hospitalization; FN, febrile neutropenia; G-CSF, granulocyte colony-stimulating factor; IV, intravenous; MUI, millions of units; PCA, patient-controlled analgesia; SC, subcutaneous.

Resources mobilized for homecare

Regarding nursing procedures for FN, the average number of nurse shifts, extra charge for Sunday shifts, and nursing acts were significantly higher for patients treated in the adult sector (p = 0.005, p = 0.013, p = 0.003 respectively), for which FN is preferentially managed at home (Table 2).

Regarding nursing procedure for other reasons, the number of nurse shifts (p = 0.033), antiemetic infusions (p = 0.033), and setup and withdrawal of infusion (p < 0.001) were significantly higher in the adult sector. Conversely, the mean number of blood collections was significantly higher in the pediatric sector (p < 0.001).

Regarding drugs received at home, the average number of ceftriaxone injections was significantly higher in the adult sector (p = 0.003), as it is the current first-line antibiotic therapy for FN (Table 2). The average number of hydration infusions and antiemetics was also significantly higher for adult patients (p < 0.001 and p = 0.033).

All patients received G-CSF. Other drugs (other antibiotics, anticoagulants, and painkillers) were each used by less than three patients.

Unit costs

The unit cost of each element of management of patients is exposed in Table 3.

Table 3.

Unit Costs (in Euros 2014)

Cost items	Unit cost (€)
Hospitalizations (source DFA)
CH	700
DH	424
Nurse care at home (source GNPA)
Shift	2.5
Sunday shift	8
Injection on CVD	12.6
SC injection	3.15
Blood collection on CVA	3.15
Monitoring of infusion <1 hour	31.5
Withdrawal of infusion	15.75
Organization of infusion	12.6
Wound care	6.3
Drugs (source www.vidal.fr)
Ceftriaxone 1 g	6.5
Vancomycin 1 g	6.1
Ertapenem 1 g	45
Imipenem 1 g	27.7
Cefepime 2 g	19.3
Polyionic 2 L	2.56
Pegfilgrastim 6 mg	979
Lenograstim 34 MUI	97.1
Filgrastim 30 MUI	99.6
Methylprednisolone 60 mg	3
Ondansetron 8 mg	14.2
Tinzaparin 10,000 UI	10.1
Enoxaparin 2000 UI	3.3
Morphine 100 mg	2

DFA, direction of financial affairs of our center; GNPA, general nomenclature of professional acts.

Total cost

The mean cost of hospitalization (standard deviation) par patient was €28,764 (12,109) in adult units, including €5,789 (5,736) for FN management. This cost was €33,303 (10,460) in pediatric units, including €11,851 (8,369) for FN (Table 4).

Table 4.

Mean Costs (Standard Deviation) Depending on the Strategy (in Euro 2014)

	Patients treated in
Cost items	Adult unit (n = 18)	Pediatric unit (n = 26)
Hospitalizations (1)
CH for FN	5,600.8 (5,852.5)	11,767.1 (8,370.9)
DH for FN	188.4 (332.3)	16.3 (83.2)
Total hospitalization for FN	5,789.2 (5,735.5)	11,851.2 (8,369.1)
CH for other cause	22,597.7 (7,985.3)	20,733.7 (4,962.6)
DH for other cause	376.9 (457.3)	717.5 (643.2)
Total hospitalization for other cause	22,974.5 (7,869.7)	21,451.3 (4,871.7)
Total 1	28,763.8 (12,109.1)	33,302.5 (10,460.2)
Nurse care at home (2)
For FN
Shifts	16.9 (21.4)	3.1 (10.4)
Sunday shifts	8.0 (10.3)	2.2 (8.0)
Injection of antibiotics	213.5 (269.8)	9.7 (35.4)
Change/managing of antibiotic PCA	0.0 (0.0)	11.5 (150.6)
Total for FN	238.4 (301.3)	26.5 (81.2)
For other causes
Shifts	157.8 (75.2)	108.9 (31.5)
Sunday shifts	46.7 (25.2)	32.0 (22.3)
SC injection of G-CSF	78.2 (49.4)	91.7 (37.0)
IV injection of antiemetics	121.8 (322.1)	0.0 (0.0)
SC injection of heparin	16.6 (59.2)	13.1 (46.3)
Blood collection on CVD	35.5 (8.9)	56.5 (16.0)
Wound care	0.0 (0.0)	7.8 (23.6)
Setup/withdrawal of infusion	453.3 (454.2)	2.4 (12.4)
Change/managing of morphine PCA	11.2 (47.5)	5.9 (29.7)
Total other cause	921.1 (672.9)	318.3 (116.3)
Total 2	1,159.5 (800.5)	344.7 (160.2)
Drugs at home (3)
For FN
Ceftriaxone	66.8 (98.5)	4.0 (14.7)
Vancomycin	0.0 (0.0)	5.6 (28.8)
Ertapenem	30.0 (127.3)	0.0 (0.0)
Imipenem	0.0 (0.0)	60.7 (309.6)
Cefepime	19.3 (81.7)	0.0 (0.0)
Total drugs for FN	116.0 (173.8)	70.4 (310.5)
For other cause
Polyionic	36.8 (36.9)	0.2 (1.0)
Pegfilgrastim	2,230.1 (2,703.4)	1,054.4 (2,179.2)
Lenograstim	1,542.0 (1,796.6)	1,851.4 (1,690.5)
Filgrastim	663.9 (1,435.9)	892.5 (1,549.4)
Methylprednisolone	7.3 (23.0)	0.0 (0.0)
Ondansetron	102.6 (294.0)	0.0 (0.0)
Tinzaparin	53.5 (190.4)	20.3 (103.4)
Enoxaparin	0.0 (0.0)	7.1 (36.2)
Morphine	1.8 (7.5)	0.9 (4.7)
Total drugs for other cause	4,638.1 (1,460.1)	3,826.8 (1,053.0)
Total 3	4,754.1 (1,512.4)	3,897.2 (1,033.2)
Total FN	6,143.7 (5,768.9)	11,948.1 (8,558.3)
Total other cause	28,533.7 (6,776.2)	25,596.3 (5,320.9)
Total 1 + 2 + 3	34,677.4 (10,808.2)	37,544.4 (10,522.5)

CVD, central venous device.

The cost of nursing care at home per patient was significantly higher in the adult sector, either for FN (95% bootstrapped CI for mean difference [−355.4; −73.7]) or for other causes (95% bootstrapped CI [−930.4; −317.9]).

G-CSF accounted for more than 95% of the drug cost in the two groups.

The mean total cost of care for this treatment period was €34,677 (10,808) per patient in adult units and €37,544 (10,522) in pediatric units. The resulting difference in mean cost was not significant (bootstrapped 95% CI for mean difference [−3,413.58; 9,098.92]).

However, the mean cost of FN management per patient was significantly higher in pediatric units: bootstrapped 95% CI for the difference of the cost €5,830.2 was [1,939.15; 10,028.86].

Multivariate analyses

Covariables included in the regression analyses were as follows: sector of care, gender, stage at diagnosis, and localization of tumor. However, age of patient, which was highly correlated with the sector of care, was excluded from the regressions to avoid multicollinearity. None of the studied parameters significantly influenced the total cost of care in multivariate analysis. The only factor significantly influencing the cost for FN management was the sector of care (p = 0.023) (Table 5).

Table 5.

Multivariate Analysis by Linear Regression for the Cost of Febrile Neutropenia Management

Parameter (reference)	Coefficient	Standard error	p value
Intercept	10,430	2,280.76	<0.001
Adult unit (pediatrics)	−5,617.61	2,362.45	0.023
Gender (male)	4,276.63	2,419.06	0.085
Localized inoperable (localized operable)	−2,790.66	3,761.75	0.463
Metastatic stage (localized operable)	336.82	2,930.22	0.909
Abdominal/pelvic tumor localization (limbs)	1,653.99	2,985.32	0.583
Other tumor localization (limbs)	−1,231.78	2,932.10	0.677

Age was excluded from the regression analysis to avoid multicollinearity.

Bold value indicates significant difference between the 2 groups.

Deterministic and sensitivity analyses

Figure 1 shows the Tornado diagram, illustrating the impact of ±20% variation in the parameters taken into account in the calculation of the total cost of care. The calculated costs were most sensitive to the unit cost of CH, followed by the average length of hospitalization for other causes and by the average length of hospitalization for FN. A decrease of 20% in the unit cost of CH would reduce the total cost to €29,038 in the adult sector and to €31,044 in pediatric units. A decrease of 20% of the length of hospitalization for FN would reduce the total cost to €33,557 in adult units and €35,191 in pediatric units.

FIG. 1.

Tornado diagram of total cost of care with a 20% variation in parameters. Length of each bar represents the extent to which the mean cost was sensitive to this parameter. The graph is laid out so that the most sensitive parameter (the one with the longest bar) is at the top and the least sensitive is at the bottom. The vertical line represents the mean cost, when all parameters assume their base value. (A) Tornado diagram for adult unit. (B) Tornado diagram for pediatric unit.

Discussion

The objective of this study was to assess the cost of care for the AYA patients of our cancer center, especially the cost concerning FN, depending on the sector of care. Indeed, published work in this area refers to American or Canadian healthcare systems: there is no medico-economic study in France.

Costs of care

We showed that the difference in total cost of care between adult and pediatric sectors was not significant, while the cost of FN management was significantly higher in the pediatric sector. In multivariate analysis, the only factor explaining this difference was the treatment sector (pediatric or adult, i.e., systematic hospitalization or homecare).

The cost of FN management is almost twice as high in pediatrics: it represents 32% of total cost over the 6 months of induction period versus 18% of the total cost for adults.

About the length of hospital stays, the only significant difference was the length of CH for FN. The number of DH for FN was significantly higher in the adult side, with a very small difference (0.4 against 0 days): this corresponds to the initial evaluation before treatment at home. We noted a trend for a lower mean duration of CH for other causes in the pediatric sector. This is partly related to our healthcare network: underaged patients are initially managed in partnership with the pediatric hospital for surgical diagnostic biopsy, and sometimes for radiologic evaluation. In contrast, patients older than 18 years are generally sent to our center when cancer is suspected.

The expenses at home were four times higher in the adult sector. This difference is partly explained by each sector's own habits. One of the main expenses is represented by the post-chemotherapy hydration, IV antiemetics, and related nurse shifts (about €670). Since 2008, most of the adult sector patients receive nocturnal hydration and IV antiemetics (methylprednisolone and ondansetron) at home during 3–5 days after the chemotherapy courses, whereas this practice has been introduced more recently in the pediatric sector. For the pediatric sector, the cost at home is almost only represented by shifts, samples, and G-CSF injections. In this area, the blood count is monitored twice a week, explaining higher number of blood samples on central venous accesses. Drug cost is almost entirely constituted by G-CSF: 96% for adults and 99% in children. Antibiotics only have a tiny place in the total cost: the management of FN represents less than 10% of the cost at home (€354 on for the adult sector and €97 in pediatric sector).

According to the deterministic analysis (Tornado chart), the factors that have the greatest impact on the cost of care are the length of CH for other causes and the unit cost of CH. These two elements are not amenable since the minimum length of hospitalization for six VIDE courses is 24 days (4 days per course): the rest of CH corresponds to the initial examinations, management of pain or other complications, and leukapheresis for high-dose chemotherapy for some patients. The third factor affecting the total cost is the mean length of CH for FN; it is the only factor we can change by adapting our strategy. In terms of decision-making, when the alternatives are equally effective, the intervention with the lower cost should be adopted (Table 6).¹⁴ Consequently, our data support the adult unit strategy, which privileges homecare, as we showed in our previous study that there is no difference in terms of efficacy.⁵

Table 6.

Cost-Minimization Analysis

Strategy	Comparison of cost	Comparison of effectiveness	Cost-minimization analysis
Patients treated in pediatric unit	Higher	Comparable	The strategy “patient treated in adult unit” is superior, compared with the strategy “patient treated in pediatric unit”
Patients treated in adult unit

Limits

Our data are consistent with previously published studies on the cost of FN, showing that hospitalization represents the higher expense^23,24 and homecare divides the cost by at least two.²⁵ However, total costs are not comparable since these studies apply to American or Canadian health systems: organization and billing are widely different.

The main limit of this study is the small number of patients, limited to one disease and treatment protocol. This choice was made to study homogeneous and comparable groups, but it results in a possible loss of statistical power.

This work is also retrospective, resulting in possible data loss, especially for acts and drugs at home. We did not include rehabilitation, physiotherapy, other paramedical procedures, and oral medications because of the age of medical records. This induces a probable underestimation of homecare cost. However, we assume that these acts are similar in the groups since patients were comparable.

This study was conducted only from the hospital perspective. It does not include the health insurance point of view or a societal perspective. Some indirect costs could be impacted by the decision of hospital treatment or home management. Finally, our study is single center, making difficult to extrapolate to all centers. Our homecare structure already allows managing a lot of care at home, including chemotherapy, enteral and parenteral nutrition, antibiotics, and palliative care. A change of practice involves developed, structured, and organized homecare structure to monitor patients safely.

The results of this study cannot be extrapolated to all AYAs treated for a solid tumor because the depth of aplasia, the rate of complications, and hospitalization lengths depend on chemotherapy protocols. Nevertheless, this study is a pilot study for our cancer center, first step before conducting a prospective study on a larger scale.

Conclusion

Our economic study showed a significant difference in the cost of management of FN between adult and pediatric units. However, the total cost of care over the study period was not significantly different. These results support the adult unit strategy, which privileges homecare, in agreement with the results of our first work showing comparable effectiveness. Further multicentric studies are needed to assess the quality of life and the preference of patients related to a change of practices. Given the differences of practice highlighted by this study, supportive care should be homogenized. Dedicated care units for AYAs could be considered.

Footnotes

Acknowledgments

We acknowledge the Clinical Research Associates of our center, who participated on the collection of data. The authors thank the reviewers for their valuable comments.

Disclaimer

Characteristics of patients, clinical data, and comparison of the effectiveness of the two strategies were published in 2015.⁵ This study was conducted after this publication, with a new data collection.

Author Disclosure Statement

No competing financial interests exist.

References

Desandes

, Lacour

, Belot

, et al. Cancer incidence and survival in adolescents and young adults in France, 2000–2008. Pediatr Hematol Oncol. 2013; 30:291–306.

Ferrari

, Thomas

, Franklin

ARK

, et al. Starting an adolescent and young adult program: some success stories and some obstacles to overcome. J Clin Oncol. 2010; 28(32):4850–57.

Bleyer

. Cancer in older adolescents and young adults: epidemiology, diagnosis, treatment, survival, and importance of clinical trials. Med Pediatr Oncol. 2002; 38(1):1–10.

Gibson

, Chisholm

, Blandford

, et al. Developing a national “low risk” febrile neutropenia framework for use in children and young people's cancer care. Support Care Cancer. 2013; 21(5):1241–51.

Penel-Page

, Normand

, Bertrand

, et al. [Management of febrile neutropenias in adolescents and young adults: differences of practice between adult and pediatric units]. Bull Cancer. 2015; 102(11):915–22.

Ammann

, Hirt

, Lüthy

, Aebi

. Identification of children presenting with fever in chemotherapy-induced neutropenia at low risk for severe bacterial infection. Med Pediatr Oncol. 2003; 41(5):436–43.

Cherif

, Johansson

, Björkholm

, Kalin

. The feasibility of early hospital discharge with oral antimicrobial therapy in low risk patients with febrile neutropenia following chemotherapy for hematologic malignancies. Haematologica. 2006; 91(2):215–22.

Manji

, Beyene

, Dupuis

, et al. Outpatient and oral antibiotic management of low-risk febrile neutropenia are effective in children—a systematic review of prospective trials. Support Care Cancer. 2012; 20(6):1135–45.

Klaassen

, Goodman

, Pham

, Doyle

. “Low-risk” prediction rule for pediatric oncology patients presenting with fever and neutropenia. J Clin Oncol. 2000; 18(5):1012–19.

10.

Orme

, Babl

, Barnes

, et al. Outpatient versus inpatient IV antibiotic management for pediatric oncology patients with low risk febrile neutropenia: a randomised trial. Pediatr Blood Cancer. 2014; 61(8):1427–33.

11.

Euro Ewing Consortium—International clinical trials to improve survival from Ewing sarcoma. Accessed October 5, 2016 from: www.euroewing.eu

12.

Motheral

, Brooks

, Clark

, et al. A checklist for retrospective database studies—report of the ISPOR Task Force on Retrospective Databases. Value Health. 2003; 6(2):90–7.

13.

O'Brien

, Heyland

, Richardson

, et al. Users' guides to the medical literature. XIII. How to use an article on economic analysis of clinical practice. B. What are the results and will they help me in caring for my patients? Evidence-Based Medicine Working Group. JAMA. 1997; 277(22):1802–6.

14.

Higgins

, Harris

. Health economic methods: cost-minimization, cost-effectiveness, cost-utility, and cost-benefit evaluations. Crit Care Clin. 2012; 28(1):11–24.

15.

Frick

. Microcosting quantity data collection methods. Med Care. 2009; 47(7 Suppl 1):S76–S81.

16.

Tan

, Rutten

FFH

, van Ineveld

, et al. Comparing methodologies for the cost estimation of hospital services. Eur J Health Econ. 2009; 10(1):39–45.

17.

Glick

, Doshi

, Sonnad

SSPD

. Economic evaluation in clinical trials (handbooks in health economic evaluation). Oxford, UK: Oxford University Press; 2007.

18.

Haute Autorité de Santé. Construction d'un outil de micro-costing en chirurgie ambulatoire. 2015. Accessed October 5, 2016 from: www.has-sante.fr/portail/upload/docs/application/pdf/2015-03/methodes_et_outils_construction_dun_outil_de_micro-costing_en_chirurgie_ambulatoire.pdf

19.

Nomenclature Générale des Actes Professionnels. Accessed October 5, 2016 from: www.ameli.fr/fileadmin/user_upload/documents/NGAP.pdf

20.

Vidal.fr—La base de données en ligne des prescripteurs libéraux. Accessed October 5, 2016 from: www.vidal.fr

21.

Gray

, Clarke

, Wolstenholme

and Wordsworth

. Applied methods of cost-effectiveness analysis in health care. Oxford, UK: Oxford University Press; 2011.

22.

Eschenbach

. Engineering economy: applying theory to practice, 3rd ed. Oxford University Press; 2010.

23.

Raisch

, Holdsworth

, Winter

, et al. Economic comparison of home-care-based versus hospital-based treatment of chemotherapy-induced febrile neutropenia in children. Value Health. 2003; 6(2):158–66.

24.

Lathia

, Mittmann

, DeAngelis

, et al. Evaluation of direct medical costs of hospitalization for febrile neutropenia. Cancer. 2010; 116(3):742–8.

25.

Teuffel

, Amir

, Alibhai

SMH

, et al. Cost-effectiveness of outpatient management for febrile neutropenia in children with cancer. Pediatrics. 2011; 127(2):e279–86.