2059. Does Age or Patient Condition Determine Allocation of Medical Resources in the Last Year of Life?
BG Fincke, Center for Health Quality, Outcomes, and Economic Research, AK Rosen, Center for Health Quality, Outcomes, and Economic Research, M Amuan, Center for Health Quality, Outcomes, and Economic Research, ME Montez, Center for Health Quality, Outcomes, and Economic Research

Objectives: In non-VA patients, health care costs in the last year of life decline with each decade of age over 65.  Our objective was to  determine if this is also true in the VA and, if so, to examine whether medical resources are being allocated on the basis of age or patient condition.

Methods: Our study population  consisted of all 49,357 patients >= 65 years old in VISN 20 who received health care services in FY 1998-99. Data came from the VISN 20 data warehouse, Decision Support System (DSS), and BIRLS. Data elements included total costs, hospital days, and month of death. The total costs (TotC) of patients not hospitalized were used to estimate the outpatient costs (OutC) of those who were.  Hospital costs (HospC) were then estimated as TotC minus OutC. We pro-rated the costs of patients who died. We calculated average costs for patients who lived and for those who died according to their decade of life above age 65. Linear regression models were developed to examine whether age was a significant predictor of costs after controlling for patient diagnoses, measured using Diagnostic Cost Groups (DCGs).

Results: For patients who died, TotC, HospC, and OutC each declined by decade of increasing age from 65 to >= 85 (TotC $13,612 to $9,594; HospC $11,934 to $7,191; OutC $4,887 to $3,115). Costs were 3-4x lower for those who lived and changed little by decade (TotC $3,942 to $4,637; HospC  $2,187 to $3,135; OutC $1,755 to $1,502).  Model results showed that age was not an independent predictor of any of the outcome measures (TotC, OutC, or HospC), with p-values varying from  0.11 to 0.71 depending on which costs were used as the dependent variable.

Conclusions: TotC, OutC, and HospC are much greater in patients who die compared to those who live. Although these costs decrease with each decade of age in the former (but not in the latter), the decline in those who die was related to patient diagnoses, not age.

Impact: Our data do not support concern that elderly patients  may be receiving lower quality care due to advancing age.