Quarterly, state level data on prescription drug utilization are available from the Health Care Financing Administration(now known as Centers for Medicare and Medicaid Services(CMS)) (HCFA(now known as CMS)). These data are reported at the National Drug Code (NDC) level. These data include all pharmaceuticals reimbursed through State Medicaid pharmacy programs. However, these data exclude prescriptions dispensed to Medicaid recipients enrolled in managed care plans where pharmacy benefits are included in the capitation arrangement. Therefore these data will miss drug utilization within those plans. As most Medicaid Managed Care plans are limited to the AFDC or TANF populations, therefore, one would expect that this deficiency would affect the analysis of antidepressant utilization more than it would antipsychotic utilization.
While these data are representative of the US Medicaid population, state-level analyses are difficult in some States due to missing data. Arizona and Tennessee fail to report Medicaid drug utilization data at all while Texas has not reported since the second quarter of 1995, and Colorado has not reported since the fourth quarter of 1996. Connecticut reported no data for 1998. Other States with questionable data include Indiana, Kansas, and Wyoming (1995 and 1996 data appear to be off by a factor of 10).
A total of 19 States report data for all 16 quarters of 1995-1998. Data imputation was accomplished for 13 other States using a linear or exponential growth interpolation model in Microsoft Excel. Missing data were imputed for States missing up to three quarters of data over the 1995-1997 period. Imputation of missing data allows analysis of data from 44 States and the District of Columbia (i.e., "45 States"). Data from Connecticut is included for years 1995-1997, but not 1998. Based on information from HCFA(now known as CMS) form 2082 (1998), these 46 States accounted for over 80% of Medicaid recipients in 1998. Only Arizona, Tennessee, Texas, Colorado, Kansas, and Indiana are excluded because of missing or suspect data.
Because the Medicaid data do not report diagnosis or dosage associated with each prescription, we were unable to analyze use of antidepressants and antipsychotics as a function of disease treated. Rather, we summarize all Medicaid use of these agents regardless of diagnosis. It should be emphasized that these data, therefore, include the use of antidepressants and antipsychotics for the treatment of illnesses other than schizophrenia or major depressive disorder. For antipsychotics, these illnesses may include behavioral disturbances in various forms of dementia as well as treatment-refractory depression and bi-polar illness. Antidepressants may be used to treat Obsessive-Compulsive Disorder, Attention Deficit Disorder, Generalized Anxiety Disorder, Bulimia, chronic pain, and sleeplessness to name a few.
In the following discussion, the word "prescription" should be understood to refer to, on average, a month's supply of medication. Using this definition requires claims data for clozapine to be divided by a factor of 4.3 to achieve comparability to prescriptions for other agents (this method accounts for Federal supply limitations of 7 days).
The monetary figures quoted for drug expenditures are net of rebate (i.e., the rebate amount returned to the State Medicaid agencies, as required by OBRA 1990, has been factored out of the expenditure data). As a result, Medicaid reimbursements represent a lower bound estimate of actual costs.