(PS5-1) A Rasch Analysis of the Patient Reported Outcomes Information System Sleep Disturbance (PROMIS-SD) Short Form Scale Among Individuals with Bipolar Disorder Undergoing Specialty Clinic Treatment

Changes in sleep are associated with bipolar disorder mood fluctuations. Often, impacted individuals exhibit excessive sleep during depression and a decreased need for sleep in mania. Because effective sleep regulation is a protective factor in managing bipolar disorder, measuring sleep disturbances can be an important preventive intervention for emerging mood episodes. While the Patient Reported Outcomes Information System Sleep Disturbance (PROMIS-SD) short form scale was developed by the NIH using state of the art psychometric methods, it has not been widely adopted in clinical practice, raising questions of its utility. Additionally, psychometric properties of the PROMIS-SD have not been examined in individuals with bipolar disorder (Yu et al., 2012). Therefore, we conducted a unidimensional item response theory (IRT) Rasch analysis of the PROMIS-SD among 193 people with bipolar disorder undergoing six weeks of specialty clinic treatment to assess how well it measures sleep disturbance and its feasibility in a clinical venue.

The PROMIS-SD short form consists of eight, psychometrically high-performing items from a larger item bank. Questions in this measure ask individuals if they had trouble sleeping or found their sleep refreshing with higher scores indicating greater sleep disturbance. A Rasch analysis uses a mathematical model that specifies the form of the relationship between bipolar individuals and the items that operationalize a single trait, in this case, sleep disturbance. The model assumes that item responses are governed by a person’s position on the underlying trait, sleep disturbance, and item difficulty. Specifically, the likelihood of higher scores increases as individuals with bipolar have more sleep disturbance and decreases as they have less. Item responses are modeled rather than sum total responses.

We hypothesized that the PROMIS-SD does not adequately measure sleep disturbance in people with bipolar disorder, underlying its rare use in clinical settings. In this poster, we report a summary of our findings supporting this hypothesis. While the measure may assess characteristics of sleep disturbance in other populations, it does not appear to capture elements of sleep disturbance adequately in individuals with bipolar disorder. Clinical implications are discussed for clinicians utilizing the PROMIS-SD when treating individuals with bipolar disorder.