Collections > Scholarly Posters and Presentations > Is the length of time uninsured prior to gaining coverage associated with changes in relative utilization of ED and primary care?
pdf

Introduction: Utilization of hospital emergency departments (ED) as a safety net provider for routine and non-emergent care by the uninsured is an oft cited problem. Expansion of health insurance coverage under the Affordable Care Act (ACA) to those previously uninsured allows for a potential reallocation of care away from more expensive settings (e.g., ED to office-based primary care). If familiarity with the health care system and connection with a primary care provider is important, those with longer spells of being uninsured prior to gaining coverage may be less likely to shift their utilization of services towards primary care. This study seeks to assess whether length of time uninsured is associated with changes in relative utilization of ED and primary care. Methods: This study uses the Medical Expenditure Panel Survey (MEPS), a comprehensive survey of health insurance, healthcare utilization, and medical expenditures in the United States Population – Adults (18 years or older) who were fully insured (covered for all 12 months) in 2014 [from longitudinal data file (HC 172) for those participating in both 2013 and 2014 (Panel 18)] Policy variable – Insurance status in 2013 – Fully insured - covered for all 12 months – Transiently uninsured – covered for 1-11 months – Persistently uninsured -uninsured for all 12 months Outcomes – Change in relative utilization of primary care to ED visits from 2013 to 2014 (primary) – Relative utilization is defined as the proportion of the total number of office-based physician, office-based physician assistant, office based nurse or nurse practitioner visits, and ED visits in a year that were not ED visits – If utilization was zero for both visit types in a given year, relative utilization was set to zero (i.e., no primary care used) – Change in utilization of primary care and ED visits (separately) from 2013 to 2014 (secondary) Model – Ordinary least squares controlling for insurance status, age, gender, race/ethnicity, education, employment, and family income (all in 2013) – Weighted using AHRQ-provided longitudinal survey weights. Results: Being transiently uninsured in 2013 is associated with a 6.3 percentage point increase (p<0.01) in relative utilization of primary care in 2014 (compared with those who were fully insured). Being persistently uninsured in 2013 was associated with a 9.0 percentage point increase (p<0.01) in relative utilization of primary care in 2014 (compared with those who were fully insured). The restricted model, which excludes education and employment due to missing values, yields similar findings. Conclusions: Longer spells of being uninsured were associated with larger shifts in relative utilization of primary versus ED care after gaining coverage. Implications: The potential for substitution away from ED utilization towards primary care by the persistently uninsured could help ease ED overcrowding and encourage earlier detection and treatment of disease. Further research is needed to determine whether selection bias may play a role in these findings (e.g., are persistently uninsured more likely to have a pre-existing condition that would encourage utilization of primary care when insured?)