Figure 2.

Temporal Patterns of Relative Rates of Emergency Department and Primary Care Utilization (Returning and Newly Insured vs. Continuously Insured); Notes: Rates were estimated by GEE Poisson Regression and adjusted for propensity class, sex, age, income, Primary Care Service Area, distance to nearest hospital, psychological disorders other than depression and anxiety, and alcohol and drug dependency. Rates in newly and returning insured are presented relative to rates in continuously insured to control for temporal trends in utilization that were unrelated to time on insurance. Additional graphs of the temporal patterns of MBHC and total specialist care visit rates are provided in the Supplement.