This abstract reports the first phase of a two-phase project that examines utilization and identifies patterns of antidepressants treatment (Phase-I), and compares physician and patient reasons for treatment patterns (Phase-II).

Prescription database of a US national pharmacy chain was queried to identify treatment-naïve patients receiving a new prescription for an SSRI or SNRI. Date of new prescription served as the index date. Patients were classified based on utilization patterns over a 3-month follow-up period post index date. A multinomial logit model was used to predict antidepressant treatment pattern. Statistical analyses were performed with two-tailed alpha 0.05.

The sample consisted of 108,229 patients, with mean age of 43 years, majority females (71%) and with 82% initiating a SSRI treatment at index. Average index copayment for the antidepressant was $23, and 90% of the sample had third-party insurance. Over the follow-up period, 40%, 58%, 2%, and <1% were classified as continuers, discontinuers, switchers and augmenters, respectively. Compared with continuers, augmenters were 34% less likely (95%CI=0.46-0.95) and discontinuers and switchers were 4 and 29% more likely (95%CI=1.00-1.07 and 1.12-1.48) to have an index SNRI vs. SSRI. Discontinuers were 62% more likely than continuers to be cash-paying vs. third-party-paying (95%CI=1.55-1.69). Compared to continuers, augmenters/discontinuers/switchers were more likely (19-79%) to have received their index-prescription from a psychiatrist vs. an internist (p<.05).

Patient, physician, drug and economic factors predicted change in the utilization of antidepressant prescription, discontinuation being the most prevalent. Determinants of discontinuation (lack of efficacy/tolerability/feeling better) will be further explored.