This study aimed to explore:(1)

How people with lived experience of psychosis in Trinidad conceptualise and experience their neighbourhood;

Thirty-two participants were purposively sampled from the INTREPID II cohort (Roberts et al., 2020); stratified by municipality. INTREPID II is a multi-site research programme in India, Nigeria and Trinidad to investigate the epidemiology of psychotic disorders across diverse settings. As part of this programme, 212 individuals with a psychotic disorder in Northern Trinidad (Diego Martin, Port of Spain, San Juan/Laventille, Tunapuna/Piarco, Arima, Chaguanas, and Sangre Grande) were recruited and interviewed between 2018 and 2020, then followed up for 2 years.

All participants had a diagnosis of psychotic disorder (with no identifiable organic cause) confirmed by a local psychiatrist following a diagnostic interview, were aged 18–64 at the time of recruitment, resided within the catchment area, and had given permission to be re-contacted for further research. Potential participants were excluded if they lacked capacity to give informed consent due to their current symptom severity, intoxication, or any other reason. We monitored the age and gender profile of the sample during recruitment and attempted to ensure a range of ages and a balance of genders within each municipality.

Potential participants were first approached by a member of the INTREPID II research team who was known to them. Those who expressed interest were then contacted by the lead researcher if they agreed for their contact details to be shared, to invite them to participate in the current study.

Trinidad and Tobago is a dual-island nation in the southern Caribbean, with a population of approximately 1.37 million. The country is characterised by major social and economic inequalities, which intersect with historical ethnic divisions between the two major groups (Afro-Trinidadians and Indo-Trinidadians). These inequalities are rooted in the colonial legacy of slavery and indentured labour, which has led to ongoing racial segregation and economic hierarchies – with Afro-Trinidadians typically holding least resources – and political divisions along ethnic lines. It is religiously mixed, with large Catholic and Hindu populations in addition to various other Christian denominations and followers of Islam, Orisha and Rastafarianism. The Trinidadian economy is based around the oil industry, with little tourism, unlike much of the rest of the English-speaking Caribbean. Its relatively strong economy has driven immigration from other countries in the region, including increasing numbers from nearby Venezuela in recent years.

The catchment area for INTREPID II (see Fig. 1), from which participants were sampled, comprises seven municipalities with a diverse population. It includes both densely populated urban areas (e.g. Port of Spain; 3090 persons/sq.km) and rural areas (e.g. Sangre Grande; 82 persons/sq.km). Table 1 describes some characteristics of the municipalities included in the study. The municipalities in the catchment area are described in more detail in the supplementary material.Fig. 1

Map of Trinidad, showing INTREPID II catchment area in pale green (from which participants for the current study were recruited). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Data were collected between January and March 2022. Participants were invited to take part in walking interviews (Evans & Jones, 2011) with the lead author (TR), using a walk-along approach in which participants define the route and answered questions while moving through their neighbourhood. This approach shifts power from interviewer to interviewee by placing the interviewee in the role of “tour guide” and the interviewer as the visitor who follows and learns from the guide, prompts observations that may not be elicited within an interview room, and provides spatial context for the data (Jones et al., 2008). Interviews were semi-structured, using a topic guide designed to address the research questions of the study. In the topic guide, “recovery” was left for participants to interpret as relevant to them. Interviews were audio recorded and subsequently transcribed. The walking route was also captured using a GPS tracker app (Bogdanovich, 2013); which provides timestamped location data that can be linked with the audio transcripts to locate the data geographically.

This study was approved by the Health Faculties Research Ethics Sub-Committee at King's College London (HR/DP-21/22–22378) and the Campus Research Ethics Committee at St. Augustine campus, University of the West Indies (CREC-SA.1237/11/2021). Participants were sent an information sheet in advance, and given the opportunity to ask questions before deciding whether to participate. Participants provided written informed consent and were given hard copies of the information sheet to keep. It was emphasised that they did not have to answer any question that they preferred not to answer, and could end the interview at any time. Data were held confidentially on a password-protected computer and backed up on a secure server. No identifying details have been included in this report or any other research outputs.

Data were analysed using an inductive approach to thematic analysis (Braun & Clarke, 2006); facilitated by NVivo software (Ltd QIP, 2020). The analysis process proceeded in three stages: describing, organising, and connecting. In the first stage, immersion and familiarisation, the first author summarised the narrative of each interview and took preliminary notes on the main points raised. In the second stage, the lead author organised the data into meaningful themes, in consultation with the co-authors, focussing on neighbourhood-level factors linked to recovery, concepts of neighbourhood and concepts of recovery. These codes were iteratively refined through comparison across interviews, and progressively reorganised to group similar ideas and avoid duplication, until a codebook was developed that incorporated all emergent themes of relevance to the research questions. The full transcripts were then coded using this codebook. Stage three involved exploring and identifying inter-relationships between codes. The themes identified were presented back to participants remotely for feedback and to guide the interpretation of results. For those transcripts with accompanying GPS data, codes and associated quotes were then linked to the GPS route and displayed geographically using ArcGIS (CESRI, 2011). These maps were consulted when synthesising and writing up the findings, and are commented on below. Double-coding was not possible due to resource constraints; this is discussed in the limitations section below.

Interviews were conducted by a white British female researcher (TR). This identity provided an “outsider” status, with both benefits and drawbacks. Being an outsider enabled questions to be posed that might seem odd for a Trinidadian researcher to ask. However, it limited participants' willingness to walk in some areas, due to concerns about safety and potential gossip. Being interviewed by a white researcher in areas where white people rarely go may have also influenced the content of the conversations in unanticipated ways; for instance, issues of race and racism were frequently raised despite not being part of the topic guide. The interviewer's gender may have also influenced how participants discussed their experiences, particularly as issues of violence and safety are highly gendered in this context. Female participants may have felt more able to speak freely about sexual assault, for example, and it is possible that male respondents could show more vulnerability than with a male interviewer. For younger participants, the interviewer's age (mid-thirties) may have generated greater formality or deference.

The analysis was led by the same researcher, which brings a specific perspective to the data. While this researcher has lived in various settings internationally and was based in Trinidad for the duration of this project, she was educated in the UK and her worldview is influenced by her public health training and her life experiences as a white British woman. The findings were discussed at various stages with the Trinidadian research team (GH, JLP, CD) and presented back to participants to protect against misunderstandings and contextualise the results, but the interpretation of findings and relative salience of themes may nonetheless have been influenced by the lead researcher's relationship to the data. A female perspective will also have brought a particular lens to the analysis, which may have differed from that of a male researcher exploring the same issues, given the gendered nature of some of the themes.

Table 2 summarises study participants’ characteristics. 22 of the 32 participants agreed to walk during the interview. Those who refused agreed to be interviewed in their home. Reasons for refusing were (a) concerns about gossip, (b) considering neighbourhood irrelevant to their recovery because they did not spend time outside of the home locally, and (c) adverse weather (typically extreme heat/intense sun). Those who agreed to walk generally chose short routes lasting for a minority of the total interview time, due to heat but also in some cases because the places that participants regularly visited were all within a short distance of their home.Table 2

Participant characteristics.

Many participants discussed recovery in terms of symptomatic remission. The symptoms that most concerned interviewees were not always psychotic symptoms. Several participants saw their primary problem as depression or anxiety (although the term “depression” may be used to refer to mental health problems more generally, as it is less stigmatised than psychosis). These participants often described recovery as being free from or able to manage these problems. Other participants had a single psychotic episode which they attributed to substance use, and considered themselves to have recovered after this episode.

There was a common perception that recovery is incompatible with continued medication use. Those who had been told by medical professionals they would need to take medication for life understood this to mean that recovery was impossible.

“What makes you say you can’t recover?”

This perception was detrimental to the outlook and self-esteem of such participants, particularly those with a schizophrenia diagnosis, which is highly stigmatised.

Others rejected their diagnosis entirely and therefore either saw recovery as irrelevant or as a process of proving to themselves and others that the professionals were wrong and they were “not mad”.

Many participants discussed issues affecting quality of life that were particularly important to them, e.g. getting a (better) job, moving into better accommodation, and claiming their rights. However, it was not clear whether they considered these broader dimensions of their life to constitute part of their “recovery” per se, as the term is typically used to refer to clinical recovery. Due to the clinical associations of the word “recovery”, during interviews we often asked questions in broader terms such as “overcoming the problems that you have experienced” and “achieving your goals” in addition to asking what helps or hinders their recovery.

When asked to define their neighbourhood, many interviewees described small areas, often limited to their own street or a couple of neighbouring streets. Places that were over 10 minutes’ walk were described as “far”, and several participants drew sharp distinctions between their neighbourhood and the next, despite being adjacent and within the same ward.

“The racialists is on my side, not out the back on that side. On the back on that side is more African, but here it’s like 50:50 African-Indian … They’re totally, totally different.” (26-year-old male, peri-urban area)

On the other hand, some participants drew much wider boundaries; in one case including the whole island:

“Although I grew up here … I live plenty of places … where people will consider bad areas … So when I think of neighbourhood I don’t think of anywhere specific, I just think of the whole of Trinidad, because I always all over the place in Trinidad, I don’t really have no place I don’t feel safe.” (20-year-old male, rural area)

Defining neighbourhoods was further complicated by several participants reporting multiple addresses (e.g. splitting time between separated parents' houses), or living elsewhere to their official registered address. Other settings were also relevant to interviewees’ recovery, e.g. prisons and the state psychiatric hospital (both typically described in extremely negative terms) and workplaces outside their local area. Non-location0based forms of community were also frequently mentioned, such as virtual/remote relationships or communities based around common interests (e.g. karaoke, Dungeons and Dragons – a fantasy role-playing game).

A substantial proportion of participants described staying at home most of the time. For some this was due to fear of crime (described below); some, who described themselves as introverts, preferred isolation. Experiences of neighbourhoods therefore varied widely among interviewees.

Neighbourhood factors that participants linked to recovery are summarised in Fig. 2, organised around the themes identified by Doroud and colleagues (Doroud et al., 2018).Fig. 2

Neighbourhood characteristics and their relationship to Doroud's four dimensions of place and recovery.

In this study, the spatial aspect of transcripts added little to the analysis beyond the traditional approach using NVivo. No patterns were noted regarding the places in which people were located when mentioning particular themes.

Our findings suggest that living in a neighbourhood with higher rates of violent crime may hinder recovery. For context, Trinidad has experienced a dramatic rise in violent crime over the past two decades, and now has one of the highest homicide rates in the world (Baird, 2020; Caribbean, 2012). There appears to have been a parallel rise in the incidence of psychosis over this period (Bhugra et al., 1996; Morgan et al., 2023). The urban municipalities that have been most affected by gang violence also have the highest rates of psychosis (Roberts et al., 2023).

Our findings are consistent with previous evidence linking violence with the onset of psychosis. Exposure to violence during childhood is an established risk factor for psychosis (Morgan & Gayer‐Anderson, 2016). Preliminary evidence suggests that violence and perceived threat at both the individual and neighbourhood levels may influence mental health outcomes throughout the life course (Baranyi et al., 2020; Conley et al., 2022); and victims of violent crime report more psychotic experiences than controls across diverse settings (DeVylder et al., 2018). The evidence on neighbourhood crime and psychosis is limited, but a positive association has been reported (Baranyi et al., 2021). A UK longitudinal study found that the cumulative effects of personal crime victimization and social cohesion and disorder may explain higher rates of psychotic experiences in more urban neighbourhoods (Newbury et al., 2018).

Moving further upstream, there is evidence of a relationship between economic inequality and rates of violence (Wilkinson, 2004); which may at least partially explain the associations observed elsewhere between inequality and rates of psychosis (Burns & Esterhuizen, 2008). Our findings lend some support to the hypothesis that rising inequality may contribute to increasing rates of psychosis by corroding social relationships within communities and increasing the likelihood of violence.

Prior literature on personal recovery emphasises the importance of both objective and perceived safety (Dell et al., 2021; Dixon et al., 2021). Safety is also emphasised under “place for being” in Doroud et al.‘s framework (Doroud et al., 2018) (see Fig. 2). The findings from this and other studies suggest that ensuring the safety of people with psychosis – in services and the community – is essential to enable recovery.

Our finding that greater social cohesion may buffer the effect of dangerous neighbourhoods is in line with research on subclinical psychotic experiences (Marsh et al., 2022) and mental health in general (Latham et al., 2022). Some studies report a negative association between the incidence of psychosis and social cohesion (Boydell et al., 2002); while others suggest that high social cohesion may increase the risk of psychosis (Kirkbride et al., 2008); potentially because more cohesive majority groups exclude minorities more effectively. The double-edged nature of social cohesion was also highlighted in our findings. Our findings suggest that, in Trinidad, engagement in social activities is less relevant than the security and sense of identity provided by knowing and trusting one's neighbours.

Existing evidence on social capital and the course of psychosis is limited (Rotenberg et al., 2020); but social cohesion and connection fit within “place for belonging” (Doroud et al., 2018); and feature prominently in many models of personal recovery (Leamy et al., 2011; Vera San Juan et al., 2021). The current study found that while neighbourhood social cohesion might promote recovery, individual experiences vary substantially within neighbourhoods.

Previous studies from Europe and North America have documented deliberate withdrawal from urban life as a strategy for coping with psychosis (Corin, 1990) in response to the stress induced by crowding and noise (Söderström, 2019). Our findings suggest that social withdrawal and the creation of safe spaces are also relevant coping strategies for people living with psychosis in Trinidad, but suggest that in this context staying within one's local area is not solely driven by managing over-stimulation, but also the threat of violence.

Substance use was also identified as important. Cannabis use is associated with worse outcomes for psychosis (Hasan et al., 2020) and may contribute to the high incidence of psychosis in Trinidad (Pow et al., 2023). Greater availability of cannabis – particularly high-potency strains and synthetic cannabinoids (Murray et al., 2016) – is likely to worsen outcomes for psychosis. The same may apply to other illicit substances that contribute to psychosis risk. If access to such substances varies by neighbourhood, this may lead to local inequalities in rates of clinical and personal recovery.

Some themes were notably absent from participant interviews. Firstly, access to services was not raised; this may be because Trinidad has a relatively accessible public mental health system in all municipalities. While formal public transport is limited, “maxi taxis” (private minibuses) operate in all areas, enabling those without cars to travel relatively easy. Secondly, neighbourhood deprivation and inequality were not directly mentioned. Deprivation may however be associated with reduced social cohesion and higher levels of violent crime (Maguire-Jack et al., 2022); with these factors mediating an association between psychosis and neighbourhood deprivation. Thirdly, no interviewee raised the visual appearance of the neighbourhood as relevant to their recovery (O'Brien et al., 2019). However, we did find evidence that fear of crime can lead to social retreat, and that living in stressful contexts impacts negatively on mental health, which are two mechanisms by which neighbourhood disorder is hypothesised to generate mental health inequalities (Massey, 2004).

Finally, we found that research that uses administrative boundaries to define neighbourhoods fails to reflect residents’ own conceptualisations of their neighbourhoods. Frequent references to small-scale geographic areas (similar to “urban bubbles” in London (Whitley et al., 2005) suggest that more fine-grained and flexible analyses are needed to fully capture neighbourhood processes in Trinidad, respecting the varied ways in which participants think about their neighbourhoods, and acknowledging that neighbourhoods are to a large extent experienced as social rather than spatial. This has potentially important implications for the delivery of mental health services, which are organised by geographic catchment area, and might benefit from shifting away from a place-based approach to a more community-based approach that builds from existing social structures. Asset mapping techniques (Lightfoot et al., 2014) could be a promising avenue in mapping community networks and building support systems around these.

An important neighbourhood variable implicated by this analysis is local rates of violent crime, and future neighbourhood studies should analyse this at a small scale (e.g. police station catchment areas, rather than municipality), while being mindful of the limitations of such boundaries in capturing the variable experiences of residents.