Economic appraisal of citalopram in the management of single-episode depression

A retrospective intent-to-treat analysis (N = 1,339) was conducted to disce rn the natural course of antidepressant use and direct health service expen ditures for the treatment of single-episode depression (DSM-IV code 296.20) among patients initiating antidepressant pharmacotherapy with either a tri cyclic antidepressant (TCA) (amitriptyline, N = 237) or a selective seroton in reuptake inhibitor (SSRI) (citalopram, N = 71; fluoxetine, N = 411; paro xetine, N = 334; or sertraline, N = 286). Data were derived from the comput er archive of a network-model health maintenance organization for the perio d of January 1, 1996, through April 30, 1999. Comparisons at the end of the g-month post-period (180 days) were undertaken between cohorts initiating antidepressant pharmacotherapy with citalopram and each SSRI or TCA. Consis tent with the intent-to-treat design, all accrued health service expenditur es were assigned to the pharmacotherapeutic option initially prescribed. Mu ltivariate models were adjusted for patient's age, gender, number of concom itant disease state processes, use of health services in the 6-month time f rame (180 days) before initiating antidepressant pharmacotherapy, specialty of physician recording a diagnosis of single-episode depression, and the p resence or absence of a previous diagnosis of single-episode depression and receipt of antidepressant pharmacotherapy. Patients initiating antidepress ant pharmacotherapy with citalopram were far more Likely to (1) have been d iagnosed by a psychiatrist (37%; p less than or equal to 0.05); (2) continu e with the original pharmacotherapeutic option (79%) compared with patients originally prescribed amitriptyline (52%; chi(2) = 17.29, df = 1, p less t han or equal to 0.05) or sertraline (65%; chi(2) = 36.91, df = 1, p less th an or equal to 0.05); no significant difference was found compared with pat ients initiating antidepressant pharmacotherapy with paroxetine (72%; p = n ot significant [NS]) or fluoxetine (83%; p = NS); (3) obtain 90 days or mor e of antidepressant pharmacotherapy (86%) compared with those prescribed am itriptyline (69%; chi(2) = 8.09, df = 1, p less than or equal to 0.05); no significant difference was found compared with sertraline (77%), paroxetine (81%), or fluoxetine (84%); and (4) obtain 6 months (180 days) of antidepr essant pharmacotherapy (68%) compared with those prescribed amitriptyline ( 39%; chi(2) = 18.26, df = 1, p less than or equal to 0.05) or sertraline (5 1%; chi(2) = 6.02, df = 1, p less than or equal to 0.05); no significant di fference was found compared with paroxetine (56%) or fluoxetine (59%). Rece ipt of amitriptyline or sertraline as initial medication was associated wit h a per capita increase (p less than or equal to 0.05) in health service ut ilization (17% and 9%,respectively) relative to citalopram. No significant difference (p > 0.05) in health service utilization was discerned between c italopram and either fluoxetine or paroxetine. Multivariate models adjusted for nonrandom assignment to the initial pharmacotherapeutic option confirm ed these findings. Further research over a longer time course is warranted.