The fundamental need for relative value metrics in educational innovation

The effect of transfers between health care facilities on costs and length of stay for pediatric burn patients

Hospitals vary widely in the services they offer to care for pediatric burn patients. When a hospital does not have the ability or capacity to handle a pediatric burn, the decision often is made to transfer the patient to another short-term hospital. Transfers may be based on available specialty coverage for children; which adult and non-teaching hospitals may not have available. The effect these transfers have on costs and length of stay (LOS) has on pediatric burn patients is not well established and is warranted given the prominent view that pediatric hospitals are inefficient or more costly. The authors examined inpatient admissions for pediatric burn patients in 2003, 2006, and 2009 using the Kids' Inpatient Database, which is part of the Healthcare Cost and Utilization Project. ICD-9-CM codes 940 to 947 were used to define burn injury. The authors tested if transfer status was associated with LOS and total charges for pediatric burn patients, while adjusting for traditional risk factors (eg, age, TBSA, insurance status, type of hospital [pediatric vs adult; teaching vs nonteaching]) by using generalized linear mixed-effects modeling. A total of n = 28,777 children had a burn injury. Transfer status (P < .001) and TBSA (P < .001) was independently associated with LOS, while age, insurance status, and type of hospital were not associated with LOS. Similarly, transfer status (P < .001) and TBSA (P < .001) was independently associated with total charges, while age, insurance status, and type of hospital were not associated with total charges. In addition, the data suggest that the more severe pediatric burn patients are being transferred from adult and non-teaching hospitals to pediatric and teaching hospitals, which may explain the increased costs and LOS seen at pediatric hospitals. Larger more severe burns are being transferred to pediatric hospitals with the ability or capacity to handle these conditions in the pediatric population, which has a dramatic impact on costs and LOS. As a result, unadjusted, pediatric hospitals are seen as being inefficient in treating pediatric burns. However, since pediatric hospitals see more severe cases, after adjustment, type of hospital did not influence costs and LOS. TBSA and transfer status were the predictors studied that independently affect costs and LOS.

Where are lengths of stay longer and total charges higher for pediatric burn patients?

Treatment of pediatric burn patients is costly and may require long length of stay in the hospital (LOS). Establishing where these LOS and charges are highest is warranted. The current study investigated whether pediatric burn patients had higher total charges and longer LOS when seen at teaching hospitals, when compared with nonteaching hospitals. The study reviewed inpatient admissions for pediatric burn patients in 2003, 2006, and 2009 by using the Kids' Inpatient Database, which is part of the Healthcare Cost and Utilization Project. International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes 940-947 were used to define burn injury, LOS, total charges, and type of hospital. The authors tested for differences between the LOS and total charges between children seen at three types of hospitals (pediatric, nonpediatric/teaching, nonpediatric/nonteaching) while adjusting for traditional risk factors (eg age, total burn surface area) by using generalized linear mixed-effects modeling. A total of N=28,777 children had burn injuries (n=16,115, 56.0% seen at pediatric hospitals; n=9353, 32.5% seen at nonpediatric/teaching hospitals; and n=3309, 11.5% seen at nonpediatric/nonteaching hospitals). Pediatric burn patients seen at pediatric hospitals, unadjusted, have significantly longer LOS (5.54 days vs 4.25 days and 4.00 days, P<.001) and more total charges in 2009 dollars ($31,319 vs $24,413 and $21,499, P<.001). In addition, patients seen at pediatric hospitals had significantly more total burn surface area (P<.001), more comorbidities (P=.021), and were younger (P<.001). After adjusting for total burn surface area, number of comorbidities, and age, no differences existed between teaching and nonteaching hospitals for LOS (P=.481) or total charges (P=.758). Although pediatric burn patients may have increased LOS and total charges when seen at teaching hospitals, when taking an unadjusted perspective, this may be an artifact that teaching hospitals see pediatric burn patients who are younger, have more comorbidities, and have more total burn surface area. As such, after adjustment, type of hospital may have no influence on LOS and total charges.