In persons with severe psychiatric disorders, distinct neurocognitive profiles hold differential associations to positive, negative and disorganized symptom dimensions of psychosis. These patterns portend specific functional outcomes, treatment efficacy, and prognoses. Similar associations have not been established in multimorbid samples in which persons present with a complex array of psychiatric symptoms. The objective of this study was to (1) establish neurocognitive profiles in a multimorbid, marginalized sample and (2) investigate their pattern(s) of association with psychiatric symptom dimensions and psychosocial outcomes.

Participants (n=370; Mage = 45 years; 74% male) were precariously housed, substance-using adults with multimorbidity, recruited from Single-Room Occupancy hotels and a community court within the Downtown Eastside of Vancouver, BC, Canada. Data were collected as part of a longitudinal examination consisting of annual, bi-annual, and monthly neurocognitive, psychosocial, and psychiatric assessments. Neurocognitive scores were combined into five cognitive domains (Attentional Control [AC]; Processing Speed [PS]; Fluid Reasoning [Problem Solving and Reversal Learning; Gf]; Encoding and Retrieval [ER]; and Decision Making [DM]) and submitted to a latent profile analysis. The resulting profiles capturing neurocognition were validated on sociodemographic and clinical variables. Finally, the profiles were compared across previously validated, population-distinct factors derived from the Positive and Negative Syndrome Scale (PANSS), as well as on measures of psychosocial functioning.

An optimal goodness-of-fit was reached for a three-profile model (BLRT=127.86, p=.01). Profile 1 (n=207, 55.9%) showed stronger neurocognition (all p<.05), with a within-profile strength in Gf (p<.001). With the exception of ER, Profile 2 (n=109, 29.5%) exhibited inferior neurocognition across all indicators compared to Profile 1 (all p <.05); yet showed a relative, within-profile strength in Gf (p < .01). Profile 3 (n=54, 14.6%) generally displayed comparable impairments to Profile 2. Additionally, their performance on Gf was remarkably low compared to Profiles 1 and 2 (p<.001). Psychiatrically, compared to Profile 1, Profile 2 exhibited more positive/disorganized symptoms and general psychopathology, as well as higher total PANSS (all p <.05), whereas Profile 3 showed the poorest insight/awareness (p<.01). Profiles 2 and 3 had lower levels of adaptive functioning and work productivity compared to Profile 1 (all p<.01).

Three neurocognitive profiles were detected in a sample of precariously housed adults with multimorbidity: one profile of comparatively higher neurocognitive capacity, with less symptoms of psychosis and better psychosocial functioning; a second profile of comparatively poorer neurocognition and psychosocial functioning, with more symptoms of psychosis; and a third profile with a severe deficit in fluid reasoning and poor insight and awareness. Given their poor insight, the third profile may be comprised of particularly vulnerable persons at greater risk of unmet healthcare needs. Interventions to improve these individuals' understanding of their personal health risks might facilitate their capacity to access services. Conversely, individuals from Profile 2 may benefit from outreach programs focusing on medication access and adherence to address their symptoms of psychosis. In sum, our findings suggest that the confluence of neurocognition and psychiatric symptoms may implicate unique treatment approaches and outcomes in precariously-housed persons with multimorbid conditions.