Abstract
Introduction: Multiple myeloma (MM) requires hospitalization for chemotherapy, stem cell transplantation, and for disease or treatment-related complications. Although there is data regarding overall incidence and mortality of MM, less is known about the patterns of hospital utilization and inpatient mortality. The purpose of this study was to describe the characteristics of patients hospitalized in 2014 for a primary diagnosis of MM, and factors associated with length of stay (LOS) and inpatient mortality.
Methods: We performed a retrospective cohort analysis of the National Inpatient Sample 2014 Database (HCUP-NIS). Patients were included in the study if they had a principal diagnosis of MM and were aged 18 years or older. We used descriptive statistics to characterize the cohort in terms of personal demographic factors (i.e., age, race, sex, insurance type, community-level income level), hospital characteristics (i.e., size, region, teaching status, and urban or rural location), and admission timing (i.e., weekend or weekday). We performed univariate analyses and multivariate analysis using these variables to determine the associations with LOS and mortality. All analyses apply the HCUP-NIS weights.
Results: The cohort comprised 16,890 patients. Most were white (63.7%), males (54.7%), and aged 61 years or older (64.8%). Nearly half (49.2%) were insured by Medicare. Hospitalizations were uniformly distributed across socioeconomic groups based on median household income by zip code. Care was delivered most often in large hospitals (69%) and urban teaching hospitals (81.5%). The greatest proportion of patients received care in the South (37.2%) and the least in the West (15.7%). Mean hospital charge was $ 113272 (95% CI $104651 to $121893) and net total hospital charge was $1.9 billion. The mean LOS was 11.4 days (95% CI 10.87 to 12.015). On multivariate analysis LOS was longer with increased Charlson index (AMD 0.77, 95% CI 0.56 to 0.98, p<0.01), BMT (AMD 9.25, 95% CI 8.30 to 10.20, p<0.01) and type of hospital (urban AMD 1.48, 95% CI 0.27 to 2.68, p=0.017, teaching AMD 1.42, 95% CI 0.63 to 2.22, p<0.01). Shorter LOS was observed for patients from areas with median household incomes in the 26th-50th and 51st-75th percentiles (AMD -0.97 95% CI -1.70 to -0.25, p=0.008 and AMD -0.99, 95% CI -1.74 to -0.24, p=0.009 respectively) compared to the lowest quartile, and in the Midwest (AMD -1.77, 95% CI -2.8 to -0.75, p=0.001). Mortality was associated with patient age (OR 1.03, 95% CI 1.02 to 1.05, p<0.01) and Charlson index (OR 1.24, 95% CI 1.14 to 1.35, p<0.01), being uninsured or self pay compared to Medicare (OR 3.14, 95% CI 1.36 to 7.25, p=0.007 and OR 6.43, 95% CI 1.31 to 31.54, p=0.022), and hospital type (urban OR 1.48 95% CI 0.27 to 2.68, p=0.017, teaching OR 1.42, 95% CI 0.63 to 2.22, p<0.01).
Discussion: Socioeconomic factors appear to be associated with LOS and inpatient mortality in patients with MM, while patient factors such as age and frailty play a smaller role. Urban and teaching hospitals are associated with longer LOS and higher mortality, but this may reflect, in part, the referral of patients for more intensive treatment in these sites. Indigent patients tend to have more comorbidities which may have confounded our findings of increased mortality in this subgroup. Further elucidation of these relationships could help inform health resource planning in the future.
Marks:Seattle Genetics: Equity Ownership; Heron: Membership on an entity's Board of Directors or advisory committees; Lilly: Membership on an entity's Board of Directors or advisory committees; Odonate: Membership on an entity's Board of Directors or advisory committees; UPMC: Employment.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal