Abstract
Introduction:
Residential instability (RI) during adolescence is associated with poor health outcomes. Also, extraversion has been shown to be a moderator of these associations. However, the associations between RI, extraversion, and age at onset of psychosis (AOP) remain unknown.
Methods:
Data were collected from patients with first-episode psychosis (FEP). Linear regression models assessed the association between RI during adolescence and AOP. Extraversion was tested as a moderator using the interaction term RI-by-extraversion.
Results:
Among 89 participants with FEP, both RI (adjusted β =−0.278, p=0.006) and the interaction term RI-by-extraversion (adjusted β=0.290, p<0.001) were associated with earlier AOP. Stratified analyses showed that RI was only significantly associated with earlier AOP among those with low extraversion (adjusted β =−0.598, p<0.001).
Conclusions:
RI predicted earlier AOP and this association was moderated by extraversion. These findings suggest that extraversion may buffer the negative relationship between RI and AOP. Future research should replicate these findings.
Keywords: Age at onset, Extraversion, Psychosis, Residential instability, Schizophrenia
1. Introduction
Residential instability (RI), defined as a greater number of moves, during childhood, has been shown to be associated with lower levels of well-being and greater mortality rates (Oishi & Schimmack, 2010). RI during childhood and adolescence is also associated with a greater risk of developing a non-affective psychotic disorder even after controlling for genetic liability (Paksarian et al., 2020; Price et al., 2018). Additionally, living in an area with a higher percentage of people who moved (referred to as area-level RI), is also associated with a higher likelihood of developing a non-affective psychotic disorder (Ku, Addington, et al., 2021; Ku, Compton, et al., 2021; Silver et al., 2002). A possible explanation for these findings is that the disruptions of social relationships as a result of RI would lead to social maladjustment (Zammit et al., 2010), which has been shown to increase risk for psychosis onset (Tarbox et al., 2013). Recently, we showed that area-level RI predicted earlier age at onset of psychosis (AOP) among individuals with FEP who provided addresses to be geo-coded to area-level characteristics; participants were from the Atlanta Cohort on the Early course of Schizophrenia (ACES) project (Ku et al., 2020). The current study extends that previous study by investigating the relationship between individual-level RI and AOP among individuals with FEP.
Earlier AOP is associated with a more severe course of illness and poor disease prognosis, including more hospitalizations, a greater burden of negative symptoms, more relapses, poorer social/occupational functioning, and worse global outcome (Immonen et al., 2017). Given the significance of AOP to the prognosis of schizophrenia, identifying risk factors could ultimately inform potential interventions to improve outcomes.
This study investigated the relationship between individual-level RI during adolescence and AOP among individuals with first-episode psychosis (FEP). We also tested extraversion as a potential moderator of this relationship because a previous study showed that the negative association between the number of residential moves and well-being was observed only among introverts and not among extraverts. In addition, this significant association was explained by the relative lack of close social relationships (Oishi & Schimmack, 2010). Prior studies have also found that low extraversion correlated with psychotic experiences (Shi et al., 2018) and onset of schizophrenia (van Os & Jones, 2001). In this study, we hypothesized that RI predicts earlier AOP and that extraversion would serve as a moderator such that RI would be associated with earlier AOP only among those with lower extraversion scores, but not among those with higher extraversion scores.
2. Methods
2.1. Subjects
This study included a subset of participants from the ACES project who were admitted for a first episode of a schizophrenia-spectrum disorder in inpatient psychiatric units in Atlanta, Georgia and Washington, D.C. The age range for inclusion into the study was 18 to 40. We included participants for whom we had sociodemographic and clinical data and who were also born in the US. Since we are studying RI (as measured by number of moves between 12 and 18) as a predictor of AOP, we excluded those whose AOP was less than or equal to 18. A sensitivity analysis was conducted to assess the association between RI and AOP that did not restrict sample based on AOP. The study was approved by the Georgia Department of Human Resources Institutional Review Board (IRB), the Grady Health System Research Oversight Committee, and the Emory University IRB.
2.2. Instruments
Sociodemographic and clinical variables were obtained from interview-based measures, chart review, and informant/family member collateral reports.
Cannabis use disorder diagnosis was determined using the Structured Clinical Interview for Diagnostic Statistical Manual-IV Axis I Disorders (First et al., 1998). Individual-level general SES was created by averaging the z-scores of five variables: the patient’s highest level of education, the patient’s mother’s and the patient’s father’s highest level of education, and reverse-coded Hollingshead Redlich Index Scores for both parents, which is an indicator of the highest occupational level ranging from 1 (high executives and major professions) to 9 (chronically jobless) (Ku et al., 2020).
AOP was determined using the Symptom Onset in Schizophrenia (SOS) inventory (Perkins et al., 2000). The earliest date of onset of either hallucinations, delusions, or both was determined by team consensus following a thorough review of all available information, including the patient’s in-depth, semi-structured SOS interview, as well as informants’ SOS interviews when available (Compton et al., 2008).
In a subset of the sample, extraversion was measured using the Neuroticism-Extraversion-Openness Five-Factor Inventory (NEO-FFI), which includes a 12-item abbreviated questionnaire for each “Big Five” personality trait. The items on the NEO-FFI (e.g., “I like to have a lot of people around me”) are rated on a 5-point Likert scale (coded as 0=strongly disagree, 1=disagree, 2=neutral; 3=agree; 4=strongly agree) and the scale has been used in psychosis studies (Compton et al., 2015; Costa PT & McCrae RR, 1992).
2.3. Data Analyses
A multivariate linear regression with robust standard errors tested the association between residential instability and AOP. In a multivariable model, we then controlled for known predictors of earlier AOP and potential confounders by including the following four independent variables: male gender, family history of psychosis, cannabis use disorder, and general SES. The moderating effect of extraversion was tested using the interaction term, RI-by-extraversion, in predicting earlier AOP. If significant, subgroup analyses by low and high extraversion scores based on the median value would be used to interpret the moderating effect of extraversion. IBM SPSS v28 was used for analyses.
3. Results
3.1. Participant Characteristics
This study included 89 individuals. Sociodemographic and clinical characteristics are summarized in Table 1.
Table 1.
Sociodemographic and clinical characteristics, n=89
| n (%) | |
|---|---|
| Gender | |
| Male | 65 (73.0) |
| Female | 24 (27.0) |
| Race | |
| Black/African American | 77 (86.5) |
| White | 8 (9.0) |
| Other | 4 (4.5) |
| Family history of psychosis | |
| Yes | 16 (18.0) |
| No | 73 (82.0) |
| Cannabis Use Disorder | |
| Yes | 57 (64.0) |
| No | 32 (36.0) |
| Highest education of mother (n=84) | |
| Less than high school graduate/GED | 14 (16.7) |
| High school graduate/GED or above | 70 (83.3) |
| Highest education of father (n=67) | |
| Less than high school graduate/GED | 18 (19.4) |
| High school graduate/GED or above | 77 (82.8) |
| Residential instability | |
| 0 | 2 (2.2) |
| 1 | 48 (53.9) |
| 2 | 24 (27.0) |
| 3 | 11 (12.4) |
| 4 | 4 (4.5) |
| Diagnosis | |
| Schizophrenia | 51 (57.3) |
| Schizophreniform disorder | 12 (13.5) |
| Brief psychotic disorder | 2 (2.2) |
| Psychotic disorder, not otherwise specified | 13 (14.6) |
| Schizoaffective disorder, bipolar type | 2 (2.2) |
| Schizoaffective disorder, depressive type | 9 (10.1) |
| Mean (SD) | |
| Age | 23.7 (4.0) |
| Age at onset of psychosis | 22.9 (3.0) |
| Extraversion score (n=48) | 30.6 (7.1) |
3.2. RI, AOP, and the moderating effect of extraversion
In the main analysis (n=89), RI was associated with earlier AOP (unadjusted β=−0.215; 95% CI=−0.395 to −0.035; p=0.020) even after controlling for four covariates (adjusted β=−0.278; 95% CI=−0.473 to −0.083; p=0.006). Among those who completed the NEO-FFI (n=48), the correlation between RI and AOP remained significant, but the correlations between extraversion and RI as well as AOP were not significant (Table 2). The interaction RI-by-extraversion was significantly associated with earlier AOP (adjusted β=0.290; 95% CI=0.160 to 0.420; p<0.001). Subgroup analyses showed that among those with low extraversion (≤30.6; n=26), RI was significantly associated with earlier AOP (adjusted β=−0.598; 95% CI=−0.882 to −0.314; p<0.001). However, among those with high extraversion (>30.6; n=22), RI was not significantly associated with AOP (adjusted β=0.402; 95% CI=−0.155 to 0.958; p=0.146) (Table 3). Sensitivity analysis showed that the association between RI and earlier AOP remained (unadjusted β=−0.194; 95% CI=−0.333 to −0.056; p=0.006) without limiting the sample based on an AOP >18.
Table 2.
Correlations among sociodemographic and clinical characteristics
| Gender | Family history of psychosis | Cannabis use disorder | General socioeconomic status | Residential instability | Age at onset of psychosis | |
|---|---|---|---|---|---|---|
| Family history of psychosis | 0.380** | |||||
| Cannabis use disorder | 0.274 | −0.214 | ||||
| General socioeconomic status | −0.081 | 0.044 | −0.134 | |||
| Residential instability | −0.024 | −0.059 | −0.298* | −0.024 | ||
| Age at onset of psychosis | 0.078 | 0.276 | −0.192 | 0.138 | −0.333* | |
| Extraversion | −0.286* | −0.060 | 0.146 | −0.108 | 0.242 | −0.283 |
= Correlation is significant at p<.05
= Correlation is significant at p<.01
Note: Correlations among sociodemographic and clinical characteristics were tested among individuals with FEP who had all the variables available for analysis (n=48). Pearson’s correlation was used between continuous variables, Cramer’s V was used between categorical variable (e.g., gender, family history of psychosis, cannabis use disorder), and point-biserial correlation was used between continuous and categorical variables.
Table 3.
Linear regression models predicting age at onset of psychosis
| Univariate | Model A | Model B | Model C | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| β | 95% CI | p | β | 95% CI | p | β | 95% CI | p | β | 95% CI | p | |
| Male gender | −0.090 | −0.301 to 0.121 | 0.398 | −0.041 | −0.262 to 0.180 | 0.716 | 0.109 | −0.257 to 0.475 | 0.542 | 0.307 | −0.159 to 0.774 | 0.182 |
| Family history of psychosis | 0.038 | −0.198 to 0.274 | 0.752 | −0.010 | −0.227 to 0.207 | 0.928 | 0.208 | −0.202 to 0.619 | 0.303 | 0.115 | −0.160 to 0.390 | 0.387 |
| Cannabis use disorder | −0.275 | −0.507 to −0.043 | 0.021 | −0.315 | −0.545 to −0.084 | 0.008 | −0.401 | −0.769 to 0.033 | 0.034 | 0.290 | −0.379 to 0.958 | 0.372 |
| General socioeconomic status | 0.021 | −0.189 to 0.231 | 0.842 | 0.022 | −0.188 to 0.232 | 0.837 | −0.152 | −0.510 to 0.206 | 0.387 | 0.336 | −0.027 to 0.700 | 0.067 |
| Residential instability | −0.215 | −0.395 to −0.035 | 0.020 | −0.278 | −0.473 to −0.083 | 0.006 | −0.598 | −0.882 to −0.314 | <0.001 | 0.402 | −0.155 to 0.958 | 0.146 |
Univariate model tested associations between individual predictor variables and age at onset of psychosis.
Model A tested associations between all variables shown and age at onset of psychosis.
Model B tested associations between all variables shown and age at onset of psychosis among the subgroup of participants who had lower extraversion scores.
Model C tested associations between all variables shown and age at onset of psychosis among the subgroup of participants who had higher extraversion scores.
All significant associations (p < 0.05) are shown in bold.
4. Discussion
We found that RI during adolescence was associated with earlier AOP among individuals with FEP. These results build upon existing literature that have reported the associations between RI during childhood and adolescence and worse health outcomes, including psychosis (Oishi and Schimmack, 2010; Paksarian et al., 2015; Price et al., 2018). Our findings are consistent with the theory that RI may lead to disrupted social networks and relationships, predisposing vulnerable youth to greater stress. And in concert with other genetic and environmental factors, the social stress associated with frequent moves would contribute to the earlier development of psychosis.
In addition, less extraverted youth may be more susceptible to this stress as they may experience more difficulty forming and maintaining social networks and relationships. In fact, research suggests that personality traits including low extraversion may contribute to features of affect, cognition, and behavior that may, in turn, lead to social isolation and reduce opportunities for disconfirmation of psychotic experiences (Shi et al., 2018). While prior research has shown that the associations between RI during childhood and wellbeing and mortality are moderated by extraversion, our findings extend that work by demonstrating that the association between RI and earlier AOP was moderated by extraversion. Among those with lower extraversion, RI was significantly associated with earlier AOP. However, RI was not significantly associated with AOP among those with higher extraversion. Perhaps higher extraversion buffers the stressful effects of RI during adolescence that may contribute to earlier onset of psychosis.
This analysis has several limitations. First, due to the cross-sectional design, this study was limited in exploring the causal mechanisms and directionality of the association between RI during adolescence and earlier AOP. Second, the design only included individuals with FEP and did not include individuals without psychosis. Third, the sample size was limited, with only 48 of 89 participants having been administered the measure of extraversion.
This study suggests the importance of understanding the developmental role of RI in psychosis as well as the potential moderating effect of personality traits like extraversion. Future research should replicate these findings in larger samples of individuals experiencing FEP and elucidate the mechanisms of these associations, which may include characterizing the social support networks.
Supplementary Material
Funding
Research reported in this publication was supported by National Institute of Mental Health grants, R25 MH101079 and K23 MH129684 to the first author and R01 MH081011 to the last author. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or National Institute of Mental Health. The authors report no financial relationships with commercial interests.
References
- Compton MT, Bakeman R, Alolayan Y, Balducci PM, Bernardini F, Broussard B, Crisafio A, Cristofaro S, Amar P, Johnson S, & Wan CR (2015). Personality domains, duration of untreated psychosis, functioning, and symptom severity in first-episode psychosis. Schizophrenia Research, 168(1–2), 113. 10.1016/j.schres.2015.06.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Compton MT, Chien VH, Leiner AS, Goulding SM, & Weiss PS (2008). Mode of onset of psychosis and family involvement in help-seeking as determinants of duration of untreated psychosis. Social Psychiatry and Psychiatric Epidemiology, 43(12), 975–982. 10.1007/s00127-008-0397-y [DOI] [PubMed] [Google Scholar]
- Costa PT, & McCrae RR. (1992). Revised NEO personality inventory (NEO-PI-R) and NEO five factor inventory (NEO-FFI) professional manual.
- First M, Spitzer R, Gibbon M, & Williams J (1998). Structured Clinical Interview for DSM-IV Axis I Disorders (SCID). New York State Psychiatric Institute, Biometrics Research. [Google Scholar]
- Immonen J, Jääskeläinen E, Korpela H, & Miettunen J (2017). Age at onset and the outcomes of schizophrenia: A systematic review and meta-analysis. Early Intervention in Psychiatry, 11(6), 453–460. 10.1111/eip.12412 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ku BS, Addington J, Bearden CE, Cadenhead KS, Cannon TD, Compton MT, Cornblatt BA, Keshavan M, Mathalon DH, Perkins DO, Stone WS, Tsuang MT, Walker EF, Woods SW, & Druss BG (2021). Association between residential instability at individual and area levels and future psychosis in adolescents at clinical high risk from the North American Prodrome Longitudinal Study (NAPLS) consortium. Schizophrenia Research, 238, 137–144. 10.1016/J.SCHRES.2021.09.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ku BS, Compton MT, Walker EF, & Druss BG (2021). Social Fragmentation and Schizophrenia: A Systematic Review. The Journal of Clinical Psychiatry, 83(1), 0–0. 10.4088/JCP.21R13941 [DOI] [PubMed] [Google Scholar]
- Ku BS, Pauselli L, Manseau M, & Compton MT (2020). Neighborhood-level predictors of age at onset and duration of untreated psychosis in first-episode psychotic disorders. Schizophrenia Research, DOI: 10.1016/j.schres.2019.12.036. 10.1016/j.schres.2019.12.036 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Oishi S, & Schimmack U (2010). Residential mobility, well-being, and mortality. J Pers Soc Psychol, 98(6), 980–994. 10.1037/a0019389 [DOI] [PubMed] [Google Scholar]
- Paksarian D, Trabjerg BB, Merikangas KR, Mors O, Børglum AD, Hougaard DM, Nordentoft M, Werge T, Pedersen CB, Mortensen PB, Agerbo E, & Horsdal HT (2020). Adolescent residential mobility, genetic liability and risk of schizophrenia, bipolar disorder and major depression. The British Journal of Psychiatry, 1–7. 10.1192/bjp.2020.8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Perkins DO, Leserman J, Jarskog LF, Graham K, Kazmer J, & Lieberman JA (2000). Characterizing and dating the onset of symptoms in psychotic illness: the Symptom Onset in Schizophrenia (SOS) inventory. Schizophrenia Research, 44(1), 1–10. [DOI] [PubMed] [Google Scholar]
- Price C, Dalman C, Zammit S, & Kirkbride JB (2018). Association of residential mobility over the life course with nonaffective psychosis in 1.4 million young people in Sweden. JAMA Psychiatry, 75(11), 1128–1136. 10.1001/jamapsychiatry.2018.2233 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shi J, Yao Y, Zhan C, Mao Z, Yin F, & Zhao X (2018). The Relationship Between Big Five Personality Traits and Psychotic Experience in a Large Non-clinical Youth Sample: The Mediating Role of Emotion Regulation. Frontiers in Psychiatry, 9, 648. 10.3389/FPSYT.2018.00648/BIBTEX [DOI] [PMC free article] [PubMed] [Google Scholar]
- Silver E, Mulvey EP, & Swanson JW (2002). Neighborhood structural characteristics and mental disorder: Faris and Dunham revisited. Social Science & Medicine (1982), 55(8), 1457–1470. [DOI] [PubMed] [Google Scholar]
- Tarbox SI, Addington J, Cadenhead KS, Cannon TD, Cornblatt BA, Perkins DO, Seidman LJ, Tsuang MT, Walker EF, Heinssen R, McGlashan TH, & Woods SW (2013). Premorbid functional development and conversion to psychosis in clinical high-risk youths. Development and Psychopathology, 25(4 PART 1), 1171–1186. 10.1017/S0954579413000448 [DOI] [PMC free article] [PubMed] [Google Scholar]
- van Os J, & Jones PB (2001). Neuroticism as a risk factor for schizophrenia. Psychological Medicine, 31(6), 1129–1134. 10.1017/S0033291701004044 [DOI] [PubMed] [Google Scholar]
- Zammit S, Lewis G, Rasbash J, Dalman C, Gustafsson JE, & Allebeck P (2010). Individuals, schools, and neighborhood: A multilevel longitudinal study of variation in incidence of psychotic disorders. Archives of General Psychiatry, 67(9), 914–922. 10.1001/archgenpsychiatry.2010.101 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.