Although baseline sleep patterns did not predict conversion to psychosis, our findings demonstrate that disturbed sleep is strongly related to increased severity of CHR symptoms over time. This association held in half the sleep characteristics when explored independently. Adjusting for depression attenuated the association between sleep and symptoms considerably. Furthermore, while effect sizes were similar bidirectionally, baseline sleep was a significant predictor of CHR symptoms two months later but not vice versa. These findings advance the current literature in several important ways. First, our findings correspond with previous CHR studies in which sleep difficulties at baseline did not predict conversion.This may suggest that sleep operates as an indicator of conversion only in conjunction with several other important markers. This hypothesis is supported by a predictive risk model of conversion in which “sleep disturbance” was one of the six factors that, jointly, yielded a markedly high positive predictive value. Sleep disturbances may have predictive value when they occur close to time of conversion, since sleep alterations are early indicators of psychosis relapse.Poor sleep may also predict conversion among specific groups, such as those with prolonged sleep disturbance or those who only recently developed sleep problems. The role of disturbed sleep as a potential catalyst for conversion in our study remains speculative because 30% of converters completed only one sleep assessment and 27% transitioned to psychosis 10 months post-baseline, at which point sleep was no longer tracked longitudinally. Future studies would benefit from looking at sleep metrics tracked longitudinally in large samples followed through the time of conversion. Second, our findings reveal long-term and robust correlations between a wide range of sleep disturbances and symptoms across all four CHR domains. Although poor sleep did not predict conversion,plant racks its relation with symptom exacerbation is clinically important considering that even non-converting CHR youth often have poor functional outcome and persistent symptoms.
Our current findings using self-reports are supported by electrophysiological studies relating REM latency and REM density to positive symptoms in psychotic disorders, reduced delta EEG activity and reduced REM latency to negative symptoms in schizophrenia, and reduced sleep spindle activity to both positive and negative symptoms,although these findings vary.In further concordance with our findings on disorganized and general symptoms, individuals with non-affective psychoses and comorbid sleep disorders had greater cognitive disorganization, depression, anxiety and tension/stress levels than those without the latter morbidity.Likewise, EEG delta activity in early-course non-affective psychoses were inversely related to disorganization syndrome.Although associations between sleep parameters and clinical symptoms vary largely across studies, individuals, and stages of illness, sleep characteristics of psychosis-afflicted individuals unequivocally differ from those who are unaffected, and these differences are present prior to illness onset. Third, symptoms of depression showed a strong attenuating effect, such that relationship strengths between sleep and all CHR symptom domains were reduced by roughly half or more. Disturbed sleep is a symptom of and risk factor for depression,and unipolar depressive disorders are common comorbidities among CHR youth. In 744 CHR individuals, >40% met DSM-IV criteria for current depression and almost 20% for past depression.Similarly, depression mediated the association between sleep and suspiciousness in CHR youth, and consistently partially mediated the association between sleep and psychotic experiences.The attenuating role of depression in the current study served to strengthen existing cross-sectional evidence and demonstrate its validity over time. Lastly, we aimed to address the directionality of the relationship between sleep and CHR symptoms. Between baseline and 2-month follow-up, total sleep score was a significant predictor of total, positive, negative, and general CHR symptoms. While general symptoms were the only domain with a statistically significant bidirectional association with sleep, bidirectional effect sizes, as indicated by regression coefficients , were all nearly equal.
These findings suggest that sleep is a driving force in symptom exacerbation, and that promoting healthy sleep may be a useful target for the maintenance of CHR symptoms. Evidence has been accumulating in support of cognitive behavioral therapy for insomnia, which not only reduced insomnia but also improved symptoms of paranoia and hallucinations in a large sample of university students and in a small CHR group.Adjacent research has found that sleep quality is malleable and thus can be improved even in clinical populations.These findings offer promising evidence for advancement in clinical staging models and future sleep-related therapies for both CHR and overt psychosis, such as an upcoming trial by Waite et al.Another budding direction of research involves computational advances in causal discovery analysis, offering innovative approaches to addressing causality in the context of observational data.Such methods require careful consideration of assumptions and properties underlying the data at hand,but could be utilized in future analyses of these and other prospective longitudinal datasets involving CHR and other clinical populations. Limitations of this study include the use of a self-report to assess sleep. The PSQI and the RU-SATED have high validity and reliability and subjective sleep problems have been the primary focus of sleep treatments. Such assessments, however, are fundamentally different from electrophysiological sleep characteristics that may be differentially associated with conversion and symptom levels. Therefore, the use of both self-report and electrophysiological sleep measures over time may inform clinical risk models and constitute targets for intervention. Other notable limitations include the significant dropout rates, the absence of ongoing sleep assessments succeeding the 8-month follow-up, and the relatively small sample size of our converting group . Furthermore, we acknowledge that pre-registering our study hypotheses would have strengthened our findings.The 1.6 to 3.5 million homeless youth in the United States may be staying in shelters or temporary housing , or living on the street – those between the ages of 18-24 years – who are experiencing homelessness are challenging to quantify due to the transiency of homelessness and the lack of consistent definitions of homelessness within the scientific literature . In San Francisco, approximately one in five homeless individuals is a TAY . Young people who experience homelessness have tumultuous lives and must continuously prioritize basic needs within a context of limited resources .
Thirty-four percent of homeless youth in San Francisco reported trading drugs to help meet basic needs such as shelter . Instability and risky behaviors inevitably place youth in situations that may cause physical or emotional harm, including exposure to social and structural violence such as racial discrimination, sexism, homophobia or living in a community where violence occurs frequently . Housing instability and the accompanying exposure to external stressors and violence have an impact on mental health and patterns of substance use. Several studies indicate the presence of psychiatric disorder in more than 48.4% of homeless youth . Homeless youth also experience higher rates of poor mental health symptoms and higher consumption of alcohol and drugs compared to their housed peers . These youth may use substances as a coping mechanism for mental health symptoms and for the daily challenges experienced while homeless ; Stress and trauma are common risk factors for substance use . TAY are also at a greater risk of experiencing violence, physical injuries, and psychological consequences over peers who are housed . Past histories of abuse, transphobia, dangerous living situations, limited financial and emotional resources, engagement in substance use and high-risk sexual activity,multi-tiered growing and irregular patterns of sleep and eating all contribute to poor mental health and substance use in youth . Evidence of the impact of trauma and substance use on health outcomes, specifically with TAY who are experiencing homelessness, is growing. This paper investigates relationships between past traumatic experiences and current substance use and mental health symptoms. We conducted a cross-sectional study of 100 homeless TAY in San Francisco, California, in close collaboration with a local community-based organization that serves this population. Each year, the CBO delivers housing, employment, and education services for 2,500- 3,000 youth aged 12-24 years who are experiencing homelessness or at risk of becoming homeless. Clients include individuals who are actively living on the street, in temporary housing or shelters, living in single-room occupancy housing, living with friends, in foster care, or otherwise unstably housed. The organization offers a broad range of programs, which includes referral centers, education and employment training programs, temporary emergency shelter, residential programs, medical care, behavioral health, and case management services. Recruitment occurred at multiple service sites of the CBO, including drop in-centers, medical clinics, agency community meetings, CBO events, transitional housing sites, and at the CBO’s housing site that exclusively serves clients with an HIV diagnosis. Study flyers were posted at CBO sites, announcements were made at site meetings, and CBO staff members referred clients to approach research staff for inclusion in the study. TAY clients who were interested in participating contacted study personnel onsite or via a secure Google Voice telephone number to schedule an appointment with one of the trained research assistants. At the meeting, potential participants were screened for eligibility. Individuals were eligible if they were between 18 and 24 years old and were recipients of services provided by the CBO. Eligible individuals then proceeded through an informed consent process. Following the informed consent process, the research assistant conducted a one-on-one survey interview with the participant, reading each question aloud and marking responses in a Computer-Assisted Survey Information Collection system via an iPad tablet. The CBO specifically recommended that the research assistants, rather than the clients, navigate the CASIC system due to client differences in reading and concentration levels that could affect their ability to complete the questionnaire by themselves. Each interview was conducted in a private room and lasted 45-120 minutes.
Exposure to traumatic events prior to the age of 18 was measured with the Adverse Childhood Experiences instrument , which has been used extensively including among young adults . The number of reported ACEs was dichotomized into less than 4, or 4 or greater. Scores of 4 or greater indicate a greater level of household dysfunction, abuse, and neglect and have been shown to be associated with morbidity and poor school performance in TAY . PTSD symptoms were measured with the PTSD Checklist for DSM-5 , which has demonstrated internal consistency, reliability, and validity . The PCL-5 can be used for screening and for monitoring change in symptom intensity. A total score was calculated , with a score of 33 or above indicating potential PTSD and the need for a more thorough psychiatric evaluation . The CES-D was used to determine symptoms of depression. The CESD has shown good reliability and validity for assessment of depression in various populations including adolescents . A total score was calculated , with a score of 16 or above indicating risk for clinical depression . Anxiety symptoms were assessed using the GAD-7 which has been shown to have strong validity and reliability including among adolescents . A total score of 10 or above on a scale of 0-21 indicates at least moderate anxiety . To evaluate substance use, participants first answered the NIDA Quick Screen V1.0 , which identifies participants who have reported alcohol binge drinking in the past year , and report past year use of tobacco products, illicit drugs, or prescription drugs for non-medical reasons. Responses were dichotomized . Participants who reported use of illegal drugs or prescription drugs for non-medical reasons in the past year then completed the NIDA-Modified ASSIST V2.0 . The original ASSIST was developed by the World Health Organization and has demonstrated strong reliability and validity . In the NIDA-Modified ASSIST, participants are asked about past three-month’s use of cannabis, cocaine, prescription stimulants, methamphetamine, inhalants, sedatives, hallucinogens, street opioids, and prescription opioids. If a substance used was not listed, participants could specify it in the “other substances” option . For each substance reported, a Single Substance Involvement Score was calculated. A mean SSIS between 0-3 indicates low-risk of harmful consequences for the user; a score of 4-27 indicates moderate-risk; a score greater than 27 indicates high-risk. Moderate-risk level scores indicate that the participant may be misusing substances but may not currently meet diagnostic criteria for a substance use disorder. Study participants ranged from 18-24 years of age; the majority were male , and 52% identified as lesbian, gay, bisexual, transgender, or queer . More than two-thirds were persons of color. At the time of the survey, 23% of participants were living with HIV, 50% were experiencing literal homelessness, and almost one-third had been previously incarcerated for more than three days. Over three-quarters of participants reported 4 or more ACEs, 80% reached the diagnostic threshold for PTSD, 74% for depression, and 51% for at least moderate anxiety.