Association between Transplant Volumes and 30-Day Readmissions Following Allogeneic Hematopoietic Cell Transplantation (allo-HCT) in the US

Background: Readmissions within 30 days after index hospitalization is a quality and cost-containment metric. Financial penalties to hospitals with high rates of risk-adjusted readmissions have been expanded beyond medical conditions like heart failure and pneumonia. Published data show significant heterogeneity in readmission rates and recent data from elderly Medicare beneficiaries reported a 17.8% readmission rate for targeted conditions. Allo-HCT is a widely used therapeutic strategy in the management of various hematologic disorders like acute myelogenous (AML) and lymphoblastic leukemia (ALL). However, allo-HCT readmission rates are poorly described, and limited to single center studies only. The association between institution HCT volume and 30-day readmission metric has not been examined.

Methods: In this observational study, we used the 2012-2014 Nationwide Readmission Database (NRD) to identify hospitals with established allo-HCT programs. Patients ≥18 years of age, discharged from hospital following an allo-HCT (identified using ICD-9 procedure code of 41.02, 41.03, 41.05, 41.06, or 41.08) were included. Annual hospital case volume was calculated as the sum of all discharges with allo-HCT within the calendar year; low, medium, and high annual case volume groups were created based on (survey weighted) tertiles of patients (pts.) in the analytic data domain (Figure 1). Rates, causes, and costs of 30-day readmissions were compared between low-, medium-, and high-volume hospitals. The analysis was limited to urban teaching hospitals and pts. admitted during month of December were excluded. The primary outcome, was the unplanned 30-day re-admission following allo-HCT. Multiple logistic regression was used to model each 30-day readmission outcome including hospital case volume with other predictors (age, sex, disease type, stem cell source, co-morbidity index, primary insurance, length of stay, infection and acute graft-versus-host-disease (aGVHD) at index admission, discharge disposition and median income quartile).

Results: A total of 17,214 (weighted) allo-HCTs were performed during the time period. Baseline characteristics of pts. in low (<58 allo-HCTs/yr.)-, medium (58-158 allo-HCTs/yr.)- and high-volume (>158 allo-HCTs/yr.) hospitals were comparable as shown in Table 1. The overall rates of readmissions were significantly higher in low volume centers (24.7.4%; SE, 1.5) compared to medium (21.4% (1.7) and high volume (9.5% (1.8), centers (p=0.03). The mean time to readmission in low vs. medium vs. high volume centers was, 11.6 [0.39] days vs. 12 [0.26] days vs. 11.5 [0.57] days respectively, (p <0.001). The length of readmission stay was significantly longer in low volume centers (mean [SD], 12.8 [0.64] days vs. 12.3 [0.91] days vs. 10.6 [0.80] days; p=<0.001) respectively. Consequently, cost per readmission was significantly higher in low volume centers (mean [SD], $164,349 [12,328] vs. $140,327 [15,297] vs. $107,362 [11,665]; p<0.001).

Readmission rates in low volume and medium volume centers compared to high volume centers were: adjusted odds ratio (aOR) 1.39, 95% CI 1.08-1.77; p =0.01 and 1.18, 95% CI, 0.89-1.55; p=0.23, respectively. Other significant predictors of readmission included disease type (ALL vs. AML): aOR 1.32, 95% CI 1.07-1.63; p= 0.009), type of primary insurance (Medicare vs. private): aOR 1.17, 95% CI 1.01-1.35; p=0.02; Elixhauser co-morbidity index (≥1 vs. 0): aOR 1.4, 95% CI 1.2-1.7; p= 0.001 and stem cell source (cord blood vs. peripheral blood; aOR 2.4, 95%CI 1.85-3.2, p<0.001). Patients with any infection and the presence of aGVHD at index admission did not have an effect on readmission rates. Neutropenia, fever, viral infection, sepsis, acute renal failure, and pneumonia were the most common reasons for readmission.

Conclusions: The likelihood of readmission after allo-HCT is elevated in centers performing <58 allo-HCTs/year, in those pts. with ≥1 co-morbidities, cord blood transplants, in ALL pts. and in Medicare beneficiaries. Lower readmission at higher-volume centers was associated with significantly lower cost to the health care system. There are important limitations with the use of data from NRD particularly the lack of information on donor status and conditioning regimen. Despite these shortcomings, the information may aid health care when developing quality-of-care metric for allo-HCT.

View largeDownload slide

View largeDownload slide

Close modal