Hospital geometric means are shown for autologous (n = 32) (A) and allogeneic (n = 22) (B) HCTs. 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. The lines across the graphs indicate the overall adjusted mean. Limit lines indicate 95% CI.
Thao V, Kozhimannil KB, Thomas W, Golberstein E. Variation in Inpatient Costs of Hematopoietic Cell Transplantation Among Transplant Centers in the United States. JAMA Intern Med. 2014;174(8):1409-1412. doi:10.1001/jamainternmed.2014.2302
Hematopoietic cell transplantation (HCT) using the patient’s own (autologous) or a donor’s (allogeneic) HCT progenitor cells is a highly effective but costly therapy for life-threatening blood disorders and cancers. Despite the small number of annual procedures (20 000 a year in the United States), HCT hospitalization spending increased from $684 million to $1.3 billion between 2004 and 2007, placing it among the procedures with the most rapid spending increases.1 Those increases reflect rising procedure volume and rising hospitalization costs.1 Prior research2 on HCT costs has been limited to single-institution analyses. To quantify variation in costs of HCT across hospitals, we analyzed data from the Nationwide Inpatient Sample (NIS),3 the largest hospital database with charge information on all patients admitted to the sampled hospitals regardless of payer.
International Classification of Diseases, Ninth Revision (ICD-9) procedure codes were used to identify HCT hospitalizations in the 2008-2010 NIS database (autologous: 41.01, 41.04, 41.07, and 41.09; allogeneic: 41.02, 41.03, 41.05, 41.06, and 41.08). We analyzed data for adults from large-volume hospitals performing 30 or more HCT procedures (autologous or allogeneic). Smaller-volume hospitals were excluded because bed size was confounded with hospital HCT volume. We retained less than 5% of hospitals in the NIS survey; therefore, weighting was not needed. Patient age was categorized as adults (18-49 years) and older adults (≥50 years) owing to differences in transplant practices by age. We used ICD-9 diagnosis codes to construct Elixhauser comorbidity index scores.4
Cost was discounted to 2008 dollars and measured using an all-payer cost to charge ratio (when unavailable, the hospital’s group mean ratio was used). Costs for autologous and allogeneic HCT were modeled separately because of differences in indications, care, and outcomes. Linear regression was used to assess associations between costs and patient and hospital characteristics and to produce age-, sex-, diagnosis-, and comorbidity index–adjusted means. Costs were log-transformed for analysis and then transformed back for reporting. Statistical analyses were conducted with SAS, version 9.3 (SAS Institute Inc).
The adjusted mean cost for autologous HCT was $47 990. Cost varied by a factor of almost 4, from $21 280 to $81 830 (Figure). The highest–cost quintile hospitals had higher proportions of patients who were younger, were non-Hispanic white, were privately insured, and had a diagnosis of lymphoma compared with the lowest–cost quintile hospitals (Table) (P < .01 for all comparisons).
Overall adjusted mean allogeneic HCT cost was $86 580, and hospital means varied 5-fold ($26 580 to $146 090) (Figure). The highest–cost quintile hospitals performing allogeneic HCT had a larger percentage of privately insured patients, patients with 1 or more comorbidities, and fewer non-Hispanic white patients (Table).
Twenty-one hospitals performed both autologous and allogeneic HCTs. There was a significant cost correlation between the 2 procedures in these hospitals (r = 0.54, P = .01).
The costs of HCT vary significantly across hospitals in the United States. We were unable to determine how costs correlate with outcomes because the NIS does not contain postdischarge survival data. Future research should explore whether higher HCT costs correlate with better patient outcomes. The major strength of the present study is that the NIS data set allows comparison of HCT costs across hospitals nationally, which has not previously been performed. Limitations of the present study include the use of charge data to calculate costs and the sample size of US hospitals analyzed; HCT is performed in relatively few hospitals. Because the NIS database contains information only on inpatient settings, we could not consider variation in rehospitalizations, which should be a future research priority. Our findings warrant attention to payment policy and price transparency, especially in the context of health care reform, for their potential to reduce cost variation across hospitals.
Corresponding Author: Viengneesee Thao, MS, c/o Katy B. Kozhimannil, PhD, Division of Health Policy and Management, School of Public Health, University of Minnesota, 420 Delaware St SE, MMC 729, Minneapolis, MN 55455 (firstname.lastname@example.org).
Published Online: June 16, 2014. doi:10.1001/jamainternmed.2014.2302.
Author Contributions: Ms Thao had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Thao, Golberstein.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Thao.
Critical revision of the manuscript for important intellectual content: Kozhimannil, Thomas, Golberstein.
Statistical analysis: Thao, Thomas, Golberstein.
Administrative, technical, or material support: Kozhimannil.
Study supervision: Kozhimannil, Golberstein.
Conflict of Interest Disclosures: Dr Kozhimannil’s work on this article was supported by the Building Interdisciplinary Research Careers in Women’s Health grant K12HD055887 from the Eunice Kennedy Shriver National Institutes of Child Health and Human Development, the Office of Research on Women’s Health, and the National Institute on Aging, at the National Institutes of Health, administered by the University of Minnesota Deborah E. Powell Center for Women’s Health. No other disclosures were reported.
Additional Contributions: Peiyin Hung, PhD candidate at the University of Minnesota, assisted with the discussion and review of the topic. There was no financial compensation.