Variation in Inpatient Costs of Hematopoietic Cell Transplantation Among Transplant Centers in the United States

Transplant centers vary in their practices for evaluation, treatment, and follow-up for patients receiving hematopoietic cell transplantation (HCT). This variation among centers has the potential to cause variation in costs of HCT. To characterize differences in costs of autologous (auto) and allogeneic (allo) HCT by hospital, we conducted a retrospective cohort study using the Nationwide Inpatient Sample (NIS). The NIS captures hospital utilization and costs, on a stratified sample of US hospitals. ICD9 diagnosis and procedure codes were used to identify hospitalizations for HCT from 2008-2010. Costs were measured by applying a cost-to-charge ratio, were discounted to 2008 dollars and log-transformed due to high skewedness. Analyses were restricted to hospitals that performed ≥ 30 HCT/year for patients ≥ 18 years, and stratified by transplant type. Our final cohort included 32 hospitals that performed allo HCT (1,932 patients) and 22 hospitals that performed auto HCT (3,164 patients). Linear regression was performed to assess associations between costs and characteristics of patients (age, gender, race, payer type, diagnosis and Elixhauser comorbidity index) and hospitals (hospital volume, ownership, and teaching status). Adjusting for these characteristics did not change the rank ordering of hospitals, and differences in cost remained significant (p<0.001). Overall, adjusted mean cost among hospitals performing auto HCT was $47,990 and costs varied by a factor of three. Among hospitals performing allo HCT, adjusted mean cost was $86,580 with costs that varied by a factor of five (Figure 1). A limitation of the NIS is that data only includes initial HCT hospitalizations; hence, we could not account for costs associated with outpatient care and subsequent hospitalizations. In conclusion, we found significant variation in costs of HCT independent of patient demographics and case mix. Future research should investigate factors that influence cost variation and whether uniform practice standards reduce costs of HCT.

 

Figure 1. Adjusted mean costs of initial HCT hospitalizations in 2008-2010  <>

Figure 1.  Ranked hospital-mean autologous HCT costs (n = 32) and allogeneic HCT costs (n = 24), 2008 – 2010.  Hospital geometric means are shown with 95% confidence intervals, and the size of the plotting character is proportional to the hospital's total volume of HCT cases.  Hospitals in the highest and lowest quintile are indicated with squares.