Economic consequences of not recognizing bipolar disorder patients: A cross-sectional descriptive analysis

Background: This retrospective study compared treatment patterns and costs for patients with recognized and unrecognized bipolar disorder with those of depressed patients without a bipolar disorder claim. Method: Claims data for 7 large national employers covering 585,584 persons aged less than 65 years were used to identify patients diagnosed with depression and initially treated with antidepressants. Data on employees, as well as spouses and dependents, for the period 1998 to mid-2001 were used. Patients were identified as bipolar based on the criteria of a bipolar diagnosis claim (ICD-9 codes: 296.0, 296.1, 296.4-296.8) and/or a mood stabilizer prescription claim. Of the patients identified as bipolar, unrecognized bipolar disorder (unrecognized-BP) patients met the criteria after antidepressant initiation, while recognized bipolar disorder (recognized-BP) patients met the criteria at or before initiation. The remaining patients in the sample were non-bipolar depressed (non-BP) patients. Outcome measures included treatment patterns and monthly medical costs in the 12 months subsequent to initiation of antidepressant treatment. Results: Of the 9009 patients treated for depression with antidepressants, there were 8383 non-BP patients (93.1%), 293 recognized-BP patients (3.3%), and 333 unrecognized-BP patients (3.7%). Use of combination therapies varied among the non-BP (11%), unrecognized-BP (32%), and recognized-BP patients (44%) (all pairwise p < .01). Use of mood stabilizers was less frequent among unrecognized-BP patients (14%) than recognized-BP patients (34%) (p < .0001). Unrecognized-BP patients incurred significantly greater (p < .05) mean monthly medical costs ($1179) in the 12 months following initiation of antidepressant treatment compared with recognized-BP patients ($801) and non-BP patients ($585). Monthly indirect costs were significantly greater (p < .05) for unrecognized-BP ($570) and recognized-BP ($514) employees compared with non-BP employees ($335) in the 12 months following antidepressant initiation. Conclusions: Patterns of medication treatment for bipolar disorder were suboptimal. Accurate and timely recognition of bipolar disease was associated with lower medical costs and lower indirect costs due to work loss.