Factors Associated with Outcomes of Hematopoietic Stem Cell Transplants in Taiwan: A Population-Based Study Using Nationwide Claims Data

Abstract

Background: Advances in conditioning regimens and stem cell sources and processing has made it possible for hematopoietic stem cell transplantation (HSCT) to be conducted in broader patient populations and in more hospitals. While there are some reports about the effectiveness of these procedures from large databases in the US and internationally, most are descriptive and focus on clinical characteristics. The majority of studies reflect programs at individual hospitals and provide a narrow picture of the population of patients undergoing transplant. Few studies have examined the variation in effectiveness by transplant program characteristics. Purposes: 1) Describe the recipient and program/organizational factors of the HSCT patients with hematological cancers in Taiwan between 2001 and 2005; 2) Determine the influence of HSCT program and organizational factors on 100-day readmission and overall survival. Methods: This population-based retrospective cohort study with longitudinal follow-up used Taiwan's National Health Insurance Research Database from the years 2001 to 2006. All adult patients who underwent an HSCT procedure from 2001 to 2005 were identified using ICD-9-CM procedure codes. The PROC GENMOD and PROC PHREG of SAS programs were used to model 100-day readmission and overall survival, adjusting for the nested design. Results: The number of HSCT recipients steadily increased from 2001 to 2005. Mean age increased from 32.6 in 2001 to 39.5 in 2005 and then decreased to 36.8 in 2005. The majority of recipients were male and the mix of diseases varied across the five years. In the ten hospitals studied, there were more private nonprofit hospitals than pubic hospitals. Most hospitals were medical centers except one district hospital. The mean number of HSCT beds was stable and the mean length of stay for the entire sample was 43.7 days. Despite the long lengths of stay, 52% of recipients were readmitted within 100 days of discharge. Ownership status, the number of HSCT hematologists, and the ratio of HSCT procedures to HSCT hematologists independently predicted 100-day readmission after adjusting for clustered data and controlling for recipient characteristics. In survival analysis, ownership and volume of HSCT procedures were not independent predictors of overall survival, after controlling for recipient characteristics. Conclusion: This population-based study of HSCT procedures for hematological conditions found that program characteristics influenced 100-day readmission but not overall survival. Appropriately, recipient and transplant characteristics strongly influenced both outcomes. While a registry would be ideal, this study demonstrated the feasibility of using claims data to study an entire population of patients undergoing HSCT.