Hanchate AD, CHOIR, VA Boston; Abdulkerim H, CHOIR, VA Boston; Stolzmann KL, CHOIR, VA Boston; Pizer SD, HCFE, VA Boston; Trivedi A, VA Providence; Linsky A, CHOIR, VA Boston; Mohr D, CHOIR, VA Boston; Frakt AB, HCFE, VA Boston; Kressin NR, CHOIR, VA Boston;
Objectives:
Strategy 1 in VHA's Blueprint for Excellence calls for improved anticipation of the needs of enrolled Veterans. We propose that models of VA healthcare demand should incorporate factors external to the VA; local economic conditions and non-VA insurance coverage may influence VA use, particularly among non-elderly Veterans. Using a national cohort, we measured sensitivity of demand for VA healthcare to these external factors.
Methods:
We identified all Veterans aged 18-63 who were enrolled for VA healthcare in 2008, and tracked longitudinal utilization until the end of FY2013 or age 64 or death if earlier. We defined use with a dichotomous indicator (1/0) and with a continuous measure of the value of annual use (2008 $) of inpatient, outpatient, and pharmacy care. To assess patients' health status, we grouped prior-year diagnoses into the 29 Elixhauser comorbidities. Based on Veterans' residence location we obtained corresponding area-level measures of external determinants of demand for VA care: county-level population employment rate; percent of state's employees in firms offering private insurance; and distances to nearest VA inpatient and outpatient care providers. We estimated logistic and generalized linear models of the association of VA use with external determinants, adjusting for patient demographics, comorbidities, and patient- and county- level fixed effects.
Results:
We identified 4.1 million non-elderly Veterans enrolled in FY2008 with an average follow-up of 5.4 years; 60% were aged 50-63 in FY2008. 59% and 60% of the Veterans used any VA healthcare in FY2008 and FY2013, respectively, with mean utilization valued at $4,146 (median $670). Among external determinants, average percent employed decreased from 94.1% (FY2008) to 90.3% (FY2010), then increased to 92.5% (FY2013), with sizable regional variation; a 1% increase in employment was associated with 0.9% and 0.8% decrease in probability and value of VA use, respectively. Residing > 25 miles from the nearest VA inpatient facility, identified among 28% Veterans, was associated with 5.9% and 1.9% lower probability and value, respectively, compared to those living < 5 miles from a VA facility.
Implications:
Local economic conditions and proximity to VA facilities influence VA use.
Impacts:
Projections of VA healthcare use need to account for the role of external determinants.