Objective: The Internal State Scale (ISS) is a self-report instrument that has been validated for discriminating mood states in patients with bipolar disorder. This study a) extends investigation to a multisite public sector sample and b) tests a revised scoring algorithm that formally identifies patients in mixed states.
Methods: Eighty-six patients with bipolar disorder from four Veterans Affairs medical centers were assessed in a cross-sectional design. Physician-conducted semi-structured interviews used DSM-IV criteria to identify subjects as meeting criteria for euthymia, mania or hypomania, depression, or mixed state (mania or hypomania plus depression). A revised ISS scoring algorithm independently assigned mood state. Mean subscale scores were analyzed across groups. Receiver-operating characteristic (ROC) curve analysis was conducted to determine optimal algorithm structure.
Results: Analysis of mean scores for the ISS subscales replicated original results for Activation, Well-Being, and Perceived Conflict, but indicated differences from the original results for the Depression Index. The ROC curve analysis identified optimal cut-off scores for the revised algorithm. The overall kappa score indicated moderate agreement between ISS and physician ratings of mood state, including mixed states.
Limitations: The study used a sample consisting primarily of male veterans. Mood state was assigned by experts using expert clinician diagnosis, not structured interviews.
Conclusion: The performance of the ISS in this multisite, public sector sample was similar to the performance in the initial research clinic sample. This finding confirms the validity of the ISS as a discriminator of mood states in bipolar disorder. The development of a revised scoring algorithm makes feasible formal identification of mixed episodes with the ISS.