The prevailing opinion holds that the newer classes of antidepressants, while tending to be more expensive to acquire than the tricyclics and monoamine oxidase inhibitors, may ultimately be more cost-effective. Most of the evidence comes from outpatient settings, health maintenance organizations, or f rom health insurance claims files. Little has been done to evaluate antidepressant use in hospitals. We plan to do this by merging billing data from this hospital's administrative database with clinical outcome data from the hospital's quality management database for the 3 year period between July 1, 1994 and July 1, 1997. We previously reviewed computerized billing and pharmacy data bases for all hospitalizations to the UCLA Neuropsychiatric Hospital between July 1, 1994 and July 1, 1996. Our preliminary findings suggest that the total cost of treatment is different across classes of antidepressant. SSRIs and the atypical antidepressants (bupropion, nefazodone, and trazodone) are the most expensive in overall costs, which include room charges, laboratory charges, other medications, and other diagnostic or procedural charges). This finding was independent of the cost of drugs, themselves, patient age, gender, length of stay, inpatient unit, date of admission, order in which the drug was given, and whether it was given alone or in combination. When the analysis was restricted to patients with severe mood disorder, the atypical antidepressants were significantly more costly than the TCADs, MAOIs and venlafaxine; and MAOIs were significantly less costly than SSRIs, the atypicals, and venlafaxine. Additional study is now needed to evaluate the relative effectiveness of each class of antidepressant and each drug within each class. We will expand our preliminary analysis to include the most recent billing data, and we will add data from the hospital's Outcome Evaluation Program of patient's status at admission, discharge, and 6-month follow-up. We will also evaluate other outcomes including use of other medications or electroconvulsive therapy, length of stay, discharge to the medical center, readmission within 30 days, or readmission at any time during the 3 year period.
