OBJECTIVE: The poor relationship between subjective and objective cognitive impairment in bipolar disorder (BD) is well-established. However, beyond simple correlation, this has not been explored further using a methodology that quantifies the degree and direction of the discrepancy. This study aimed to develop such a methodology to explore clinical characteristics predictive of subjective-objective discrepancy in a large BD patient cohort.

METHODS: Data from 109 remitted BD patients and 110 healthy controls were pooled from previous studies, including neuropsychological test scores, self-reported cognitive difficulties, and ratings of mood, stress, socio-occupational capacity, and quality of life. Cognitive symptom 'sensitivity' scores were calculated using a novel methodology, with positive scores reflecting disproportionately more subjective complaints than objective impairment and negative values reflecting disproportionately more objective than subjective impairment ('stoicism').

RESULTS: More subsyndromal depressive and manic symptoms, hospitalizations, BD type II, and being male positively predicted 'sensitivity', while higher verbal IQ predicted more 'stoicism'. 'Sensitive' patients were characterized by greater socio-occupational difficulties, more perceived stress, and lower quality of life.

CONCLUSION: Objective neuropsychological assessment seems especially warranted in patients with (residual) mood symptoms, BD type II, chronic illness, and/or high IQ for correct identification of cognitive deficits before commencement of treatments targeting cognition.