Abstract
Objective
The purpose of this study was to compare health-care resource utilization and outcomes among children treated for low-risk febrile neutropenia (FN) in a hospital-based setting with those treated in a home-care-based setting.
Methods
The perspective of this retrospective, cohort study was the health payer. We collected health-care utilization and treatment outcome data from medical records of 63 children (26 boys and 37 girls) with low-risk, chemotherapy-induced FN who were treated at the University of Arizona (27 children, the hospital-based group) and University of New Mexico (36 children, the home-care-based group). We identified 144 FN episodes (72 episodes in each group). Health-care utilization included physician visits, home-care visits, laboratory visits, outpatient visits, hospital days, intensive care unit days, medical tests and studies, and medications used to manage FN (e.g., filgrastim, antimicrobials, and ancilliary drugs and supplies). We applied uniform charges, based on those used at the University of New Mexico in 1998. We collected outcomes of the FN treatment (success vs. failure and time to resolution, defined as number of days of antibiotic therapy). Rates of positive blood cultures during treatment were also compared. Data were analyzed using nonparametric Mann–Whitney U tests for continuous data and chi-square analysis for categorical data. Sensitivity analyses were conducted by varying the amount of total resource utilization, as well as utilization of specific health-care resources.
Results
There was no difference in outcome; all episodes of treatment in both groups resulted in successful recovery from FN. Time to resolution of FN was 8.3 ± 2.7 days for home-care FN episodes versus 7.3 ± 3.6 days for hospital FN episodes (P =.064). Median charge per FN episode was significantly (P .001). Positive blood cultures were more common among the hospital-based FN treatments (30.6 vs. 11.1%, P=.012).
Conclusions
We found that management of low-risk FN in a home-care-based setting was associated with significantly lower median total charges with no differences in outcome.
Authors
Dennis W. Raisch Mark T. Holdsworth Stuart S. Winter John J. Hutter Michael L. Graham