Where I lay my head is home: residential instability and earlier onset of psychosis

skyline of houses in oxford, UK

Having a secure home during childhood and adolescence is known to benefit future wellbeing. There’s emerging evidence that residential instability – moving house many times – can cast a long shadow over mental health, particularly if it happens repetitively during adolescence. Residential instability is known to increase the likelihood of future psychosis, however we can anticipate that this is due to a complex interplay between neurodevelopmental, social, and economic and educational factors. The authors of the paper discussed in this blog (Ku et al, 2023) have previously shown that living in a more transient neighbourhood is associated with earlier-onset and potentially more severe psychotic illness. In this paper, they develop this work by asking two important questions:

  1. Is this also true of one’s own residential instability, i.e. at the individual level?
  2. If the critical factor is disruption of social networks, then does the impact of residential instability depend on your level of extraversion?

Extraversion is an interesting way to look at the social question: this trait will govern a young person’s tendency to seek out contact with others, boosting their opportunities for reality checking, or having early symptoms recognised. All this speaks to potentially modifiable factors that could mitigate the risk of transition to illness, its eventual severity and lifetime impact.

Lots of moving boxes gathered in an empty room

What makes moving house so stressful? What is the mental health impact when our reality is inherently unstable?


This cross-sectional study recruited US-born people aged 18-40 from Atlanta, Georgia and Washington D.C. who were admitted to hospital with a first episode of a schizophreniform psychosis. They were a subset of the project Atlanta Cohort on the Early Course of Schizophrenia (ACES), from which their previous work was derived. A sample of 89 addressed the main hypothesis relating adolescent residential instability to age of onset of psychosis, however this almost halved to 48 for the instability-by-extraversion interaction. To measure residential instability and age of onset of psychosis, they made the best use of available (albeit non-contemporaneous) data, drawing from patient, collateral and medical records. Extraversion was measured using the Neuroticism-Extraversion-Openness Five Factor Inventory (NEO-FFI) during their current admission. They adjusted for confounders of male gender, family history of psychosis, cannabis use disorder and general socioeconomic status (a summary of the highest education level of the patient and their parents, and the occupational level of the parents).


The participants were broadly representative of first-episode psychosis cohorts, with mean age of 23.7 years (SD 4 years), and 73% male. A high proportion (86.5%) of the participants were Black or African American.

Greater residential instability during adolescence is associated with earlier onset of psychosis

This had a relatively modest negative impact on the age at onset of psychosis (see Table 1 below). Adjusting for their four a priori nuisance covariates (gender, family history of psychosis, cannabis use and socioeconomic status) strengthened their main finding, suggesting a specific impact of repeatedly moving home that transcended these confounds.

Higher extraversion may mitigate the association between residential instability and earlier onset of psychosis

Within the smaller sample (n = 48) of those who had completed the NEO-FFI, there was a significant interaction between extraversion and residential instability on age of onset of psychosis (see Table 1 below). To understand this interaction, they divided the sample into low and high extravert groups, according to the sample mean of 30.6 (which is in keeping with normative data). Only those with low extraversion demonstrated a tendency to develop psychosis earlier according to an increasing number of house moves. The same effect was nonsignificant in the high extravert group (p = 0.146).

Table 1: Effect of residential instability and its interaction with extraversion on the age at onset of psychosis. Covariates adjusted for were gender, family history, cannabis use, and socioeconomic status.

β of the effect 95% confidence interval p-value
Residential instability

(n = 89)

Unadjusted -0.215 -0.395 to -0.035 0.020
Adjusted -0.278 -0.473 to -0.083 0.006
Residential instability x Extraversion

(n = 48)

Adjusted 0.290 0.160 to 0.420 0.001
Low extraversion group (n = 26) -0.598 -0.882 to -0.314 0.001
High extraversion group (n = 22) 0.402 -0.155 to 0.958 0.146
a girl with brunette hair reaches her hands out towards the camera

Adolescents with low extraversion scores (as measured using the Neuroticism-Extraversion-Openness Five Factor Inventory) demonstrated a tendency to develop psychosis earlier according to an increasing number of house moves.


Higher residential instability during adolescence hastened the onset of psychosis, even after accounting for key confounding factors. When considering the personality trait of extraversion, this relationship was only seen in those with low extraversion, although the sample for this analysis was small. The authors concluded:

Our findings are consistent with the theory that residential instability may lead to disrupted social networks and relationships, predisposing vulnerable youth to greater stress.

Strengths and limitations

This is a well-conducted study that has made the most of the information available, within the confines of a cross-sectional design. In terms of bias, recall bias for such a marked event as moving home will likely be relatively low. The authors also accounted for several key potential confounds of interest. Authors also excluded people who became unwell before age 18, and those with incomplete data. Those excluded are likely to have a more severe illness, or be subject to particularly chaotic adolescent periods, so it could be argued that these findings underestimate the impact of residential instability overall.

The most novel finding is the interaction residential instability has with extraversion – though extraversion data was only available for approximately half their original sample. The authors could also have assessed the impact of extraversion within a formal mediation analysis, rather than a 50:50 split as they have done, however this would likely have been underpowered.

Most participants in this study were Black or African American: on the one hand, it’s refreshing to see a study that reports on a group that are often neglected in research, and are more likely to be detained, medicated and incarcerated with psychosis than other racial/ethnic groups. However, African Americans face a specific and complex interplay between socioeconomic factors, residential instability and the onset of psychosis that may not generalise to all populations.

This US study focuses on people who were born there: it therefore excludes migrants, who intrinsically have residential instability, and who face an established increased risk of psychosis, especially if they migrate prior to age 18. The authors therefore cannot comment on the complex interplay between migration and risk of psychosis, but it is interesting to establish that residential instability is an important risk factor even for native populations.

Finally, the authors do not describe whether their sample are care-experienced (i.e. have been in foster or residential care), but it is known that this group face a particularly high risk of mental illness, and are sadly often excluded from research studies, in part because of their tendency to move around in fraught circumstances. This paper highlights how important it is that we include and appropriately describe care-experienced people in future studies of psychosis.

Many houses are seen from above, some have terracotta roof tiles, others have swimming pools in their back gardens

It could be argued that, due to the exclusion of adolescents who are potentially more unwell and in more chaotic residential circumstances, these findings underestimate the impact of residential instability overall.

Implications for practice and research

The authors argue that there is something about a person’s natural inclination to seek contact with other people that helps stave off the risk – or at least the speed – of developing psychosis, in the face of particular stressors. There are many reasons why this might be, but specifically highlighting residential instability is useful in terms of improving early interventional and risk reduction approaches. Clinicians may wish to be particularly proactive in bolstering social opportunities and support for people who are less extravert, and are currently facing a move of residency.

What is it about being extravert that is potentially protective? Seeking contact with others, and communicating your inner states more vividly, could elicit the help people need when facing stressors such as moving home. More introverted young people facing stressors may therefore benefit from being gently supported more proactively, in all educational, social, and health settings, rather than being allowed to fly under the radar and only come to anyone’s attention once illness is established. This may be particularly relevant to people with autistic spectrum disorder, who have higher rates of both introversion and psychosis – and may be especially sensitive to the environmental upheaval of moving house. Clinicians could respond by viewing moving house as a potential precipitant of mental illness, and establish supports to make the transition as planned and predictable as possible.

This study focusses on the key life phase of adolescence. It would be helpful to expand on this and explore whether there are specific time windows across childhood and adolescence where extraversion-moderated residential instability exerts its greatest impact. These risk windows could highlight important causal factors – perhaps moves co-occurring with school transitions, or the onset of puberty, are especially ill-timed thanks to their social and neurodevelopmental sequalae.

Extraversion also raises interesting questions about dopamine function, (a core component of the neurobiology of psychosis), as responses to dopamine-modulating medications are known to depend on trait extraversion. Further research in this area could help personalise optimal medication choices for people with first episode psychosis.

Finally, it is worth asking why moving home is so stressful. The authors speculate that being an extravert and seeking more social support means having more opportunities to reality-check. But what if reality is inherently unstable? Bayesian Brain Theory proposes that the brain is constantly trying to predict its sensory input and minimise surprise. When this process goes wrong, the divergence between perceived and objective reality is believed to underpin psychosis. This process is made harder by residential instability, placing the adolescent in a constant state of uncertainty, and so aggravates the very neural mechanisms that may lead to psychosis.

A group of young black men are gathered around a phone and they are laughing

What is it about being extravert that is potentially protective? Extraverts seek contact with others, and communicate their inner states vividly, which could elicit the help they need when facing stressors, such as moving home.

Statement of interests

The author has no interests to declare.


Primary paper

Ku BS, Walker EF, Druss BG. et al (2023) Residential instability during adolescence predicts earlier age at onset of psychosis: The moderating role of extraversion. Early Interv Psychia, 1-5. https://doi.org/10.1111/eip.13375

Other references

Oishi, S., & Schimmack, U. (2010). Residential mobility, well-being, and mortality. Journal of Personality and Social Psychology, 98(6), 980–994. https://doi.org/10.1037/a0019389

Li M, Li WQ, Li LMW. Sensitive Periods of Moving on Mental Health and Academic Performance Among University Students. Front Psychol. 2019 Jun 13;10:1289. doi: 10.3389/fpsyg.2019.01289. PMID: 31263436; PMCID: PMC6585164.

Egan, V., Deary, I. & Austin, E. The NEO-FFI: emerging British norms and an item-level analysis suggest N, A and C are more reliable than O and E. Pers Indiv Differ 29, 907–920 (2000).

Thompson, R. G. & Hasin, D. Psychiatric Disorders and Treatment Among Newly Homeless Young Adults With Histories of Foster Care. Psychiatr Serv 63, 906–912 (2012). https://pubmed.ncbi.nlm.nih.gov/22706986/

Maura, J. and Weisman de Mamani, A., 2017. Mental health disparities, treatment engagement, and attrition among racial/ethnic minorities with severe mental illness: A review. Journal of clinical psychology in medical settings24, pp.187-210. https://pubmed.ncbi.nlm.nih.gov/28900779/

Anderson, K. K. & Edwards, J. Age at migration and the risk of psychotic disorders: a systematic review and meta-analysis. Acta Psychiat Scand 141, 410–420 (2020). https://pubmed.ncbi.nlm.nih.gov/31903545/

Patel, P. K., Leathem, L. D., Currin, D. L. & Karlsgodt, K. H. Adolescent Neurodevelopment and Vulnerability to Psychosis. Biol Psychiat 89, 184–193 (2021). https://pubmed.ncbi.nlm.nih.gov/32896384/

Ribolsi, M. et al. Recognizing Psychosis in Autism Spectrum Disorder. Frontiers Psychiatry 13, 768586 (2022). https://pubmed.ncbi.nlm.nih.gov/35295770/

Wacker, J. & Smillie, L. D. Trait Extraversion and Dopamine Function. Soc Personality Psychology Compass 9, 225–238 (2015). https://compass.onlinelibrary.wiley.com/doi/full/10.1111/spc3.12175

Sterzer, P. et al. The Predictive Coding Account of Psychosis. Biol Psychiat 84, 634–643 (2018). https://pubmed.ncbi.nlm.nih.gov/30007575/

Photo credits

Share on Facebook Tweet this on Twitter Share on LinkedIn Share on Google+