Predictors of Length of Hospital Stay and Readmission in Hematologic Stem Cell Transplant Recipients

Abstract

Background: Hematopoietic stem cell transplantation (HSCT) is an established treatment for patients with various malignant and non-malignant diseases. The intense nature of the transplant experience mandates the presence of a consistent caregiver throughout the inpatient phase and after discharge. Increased length of hospital stay (LOS) and unplanned readmission increase health care costs. Identifying patients at risk for prolonged LOS and readmission and the caregiver factors that might contribute to these outcomes has the potential to reduce health care costs and patient's suffering. Caregiver characteristics are examined as predictors of these outcomes in cancer patients, but data are lacking on whether they predict LOS and readmission in HSCT recipients. Purpose: To examine caregiver characteristics as predictors of LOS and readmission risk for HSCT recipients. Methods: A secondary data analysis obtained in a study of the effectiveness of problem solving education in caregivers and patients during allogeneic HSCT at National Institutes of Health (Bevans et al., 2013) was conducted. Generalized linear mixed method models were used to examine predictors of LOS and generalized estimated equation models were used to examine predictors of 30-day readmission. Independent variables included: patient factors [age, disease condition, disease stage, comorbidity index, infection, psychological distress, and performance status] and caregiver factors [age, gender, relationship to patient, health problems, self-efficacy, burden and psychological distress]. Results: Reduced intensity conditioning allogeneic transplant recipients (N=60) and their caregivers (N=72) were included. Patients had a mean age of 46 years, were largely males (63%), and white (68%). Caregivers had a mean age of 52 years, were predominantly females (75%) and white (73.6%). Average LOS was 25 days and 35% of the patients were readmitted within thirty days after discharge. Patient factors that predicted LOS were disease stage (p=0.01), infection status (p=<.001) and comorbidity index (p=0.03). Infection status predicted readmission risk (p=<.001). Caregiver factors did not predict LOS or readmission. Caregiver psychological distress tended to predict LOS (p=0.09). Conclusions: Caregiver variables did not add meaningful information beyond traditional patient factors in predicting LOS and readmission in this population. More longitudinal, prospective studies are needed to understand the influence of caregiver factors on these outcomes.