OBJECTIVES: To evaluate the likelihood of employment termination among patients with bipolar disorder (BPD) treated with different classes of psychotropic medications. METHODS: Patients with BPD (classified according to ICD-9-CM codes) were identified from the Human Capital Management Services Research Reference Database. Patients with continuous eligibility six months before and 12 months after their initial prescription treatment for BPD were categorized into those using: atypical antipsychotics only (ATYP); conventional antipsychotics, mood stabilizers (including lithium, divalproex, lamotrigine, and carbamazepine), and specific anticonvulsants only (OTHER); medications from both categories (BOTH); and no study-specified psychotropic medications (NONE). The index “prescription” date for the NONE group was defined as six months after the initial diagnosis. Both voluntary and involuntary terminations of employment were included. Regression models controlled for possible confounding factors (age, gender, location, salary, race, marital status, full-time/part-time status, prior comorbidity index, prior medical costs related to BPD, prior other medical costs, index date, and medical services related to BPD). Treatment groups were compared over a follow-up period ranging from 12 to 24 months after the index prescription date. RESULTS: Six hundred ninety-nine patients with BPD were classified into the ATYP (n=25), BOTH (n=190), NONE (n=170), and OTHER (n=314) treatment groups. The ATYP group demonstrated the lowest rate of employment termination (1.5%; 95% CI -3.3%, 6.3%) followed by the BOTH (5.8%; 95% CI 2.5%, 9.1%), NONE (8.9; 95% CI 4.7%, 13.2%), and OTHER (9.3%; 95% CI 6.1%, 12.5%) groups. Differences between treatment groups were not significant. The numerical difference between ATYP and OTHER did not reach significance (p=0.058). CONCLUSIONS: The ATYP group demonstrated the lowest employment termination rate in the follow-up period of 12 to 24 months after the index prescription date. Further research is warranted to examine the influence of specific patient variables and treatment regimens on employment termination in patients with BPD.