Abstract
Background
People living in precarious housing or homelessness have higher than expected rates of psychotic disorders, persistent psychotic symptoms, and premature mortality. Psychotic symptoms can be modeled as a complex dynamic system, allowing assessment of roles for risk factors in symptom development, persistence, and contribution to premature mortality.
Method
The severity of delusions, conceptual disorganization, hallucinations, suspiciousness, and unusual thought content was rated monthly over 5 years in a community sample of precariously housed/homeless adults (n = 375) in Vancouver, Canada. Multilevel vector auto-regression analysis was used to construct temporal, contemporaneous, and between-person symptom networks. Network measures were compared between participants with (n = 219) or without (n = 156) history of psychotic disorder using bootstrap and permutation analyses. Relationships between network connectivity and risk factors including homelessness, trauma, and substance dependence were estimated by multiple linear regression. The contribution of network measures to premature mortality was estimated by Cox proportional hazard models.
Results
Delusions and unusual thought content were central symptoms in the multilevel network. Each psychotic symptom was positively reinforcing over time, an effect most pronounced in participants with a history of psychotic disorder. Global connectivity was similar between those with and without such a history. Greater connectivity between symptoms was associated with methamphetamine dependence and past trauma exposure. Auto-regressive connectivity was associated with premature mortality in participants under age 55.
Conclusions
Past and current experiences contribute to the severity and dynamic relationships between psychotic symptoms. Interrupting the self-perpetuating severity of psychotic symptoms in a vulnerable group of people could contribute to reducing premature mortality.
Slicing, sampling, and distance-dependent effects affect network measures in simulated cortical circuit structures
