The Effect of Medicaid Reimbursement Generosity on Physician Access: Diagnosing of Asthma in Children

Medicaid reimbursement levels have been traditionally lower than other insurances.  This discrepancy has driven researchers to assess whether the care received by Medicaid patients is different from other patients.  Results from the current literature are mixed.  However the state-level generosity measures used in these studies are generally indices representing overall generosity and are not directly linked to the access measures being evaluated.  Instead, in this study we estimate state-level generosity measures specific to the treatment of children with asthma and assess whether reimbursement generosity affects access to physician services for children with asthma across states.  Using Medicaid MAX databases across the US in 2007, we gather Medicaid claims for all children with an ICD-9 diagnosis code for asthma in the first diagnosis field.  We then identify the top 10 procedure codes billed on these claims at the national level and calculate the average cost of this “basket” of services by state, weighting by a national distribution in the frequency of these codes.  This became our measure of reimbursement generosity.  Using weighted regressions, we test three hypotheses: whether more generous reimbursement leads to higher rates of asthma diagnoses for children in the Medicaid program, measured as a rate per child-year of eligibility;  whether more generous reimbursements lead to a higher percentage of asthma visits occurring in physician offices (in contrast to community health clinics); and whether more generous reimbursements affect the percentage of all asthma-related visits that occur in physician office for a patient with asthma.  In each model, we specify the state-level asthma reimbursement measure, a squared term for the reimbursement measure, the percentages of beneficiaries in the state that are female and non-white, and a measure of the prevalence of childhood asthma in the state in 2007 (from the CDC’s BRFSS prevalence data).  Reimbursement levels for our basket of goods ranged from $20.13 in Rhode Island to $80.81 in Alaska with a median of $41.02.  Iowa had the lowest prevalence of childhood asthma (5.2%) and DC the highest (15.7%).  Our results suggest that access to physician services for Medicaid children with asthma increases with reimbursement generosity but that this effect is smaller at higher levels of reimbursement generosity.  We estimate that a $10 increase in reimbursement generosity at the mean ($40.73) would increase the rate of asthma diagnoses by 2 children per 100 child-years of eligibility (mean = 5.7, p-value = 0.022), increase the percent of asthma diagnoses occurring in a physician’s office by 6.8 percentage points (mean = 51.8%, p-value=0.003) and increase the percent of all asthma visits occurring in a physician’s office by 7.8 percentage points (mean=40.0%, p-value<0.001) for a given patient.  Our results show unambiguously that Medicaid reimbursement rates affect how often asthma is diagnosed in an area, whether physicians see asthma patients (compared with other healthcare providers) and whether they continue to follow up with them.  If the current rates of asthma diagnosing in physician offices are low then increasing Medicaid reimbursements would benefit patients.