Defining Urbanicity in the Context of Psychosis Research: A Qualitative Systematic Literature Review

Abstract

Background and Hypothesis

Numerous studies have found that being born or raised in urban environments increases the odds of developing psychosis in Northern and Western Europe. However, available research from Southern Europe, Latin America, and Asia has reported null results. A limitation in most studies to date is the inadequate characterization of urban and rural life components that may contribute to varying psychosis risk across regions.

Study Design

To deepen our understanding of the different concepts and measures of urbanicity and related factors in psychosis research, we conducted a qualitative systematic literature review extracting information from studies published between 2000 and 2024.

Study Results

Sixty-one articles met the inclusion and exclusion criteria and were used in the thematic analysis. The analysis revealed that urbanicity lacked a single, coherent definition across studies and regions. Three major categories of themes were developed from the analysis: (1) Urbanicity comprises several interconnected constructs, (2) Urbanicity measurements vary between countries from the Global North and the Global South, and (3) Urbanicity operates through key neighborhood-level mechanisms.

Conclusions

Future research on urbanicity and psychosis should consider the potential limitations of urbanicity’s conceptualization and operationalization and aim to address these limitations by focusing on contextual, historical, and community-level factors, utilizing locally validated measures, and employing mixed-method designs.

Introduction

Social determinants of mental health comprise the range of structural conditions encountered by individuals from conception to death, which influence mental health outcomes and contribute to disparities within and across populations.¹ Urbanicity, defined as the impact of living in urban areas at a given time, has been proposed as a major social determinant of mental health, particularly for psychosis.^2,3 Recent reviews, including a series of papers on social determinants of schizophrenia, highlight the role of urban environments in increasing the risk of nonaffective psychosis.^4–8 However, evidence from the Global South—referring to countries that encompass a wide range of economic, political, and cultural contexts, many of which are low-income and politically neglected, each with unique challenges and contributions to the understanding of urban living in Latin America, Asia, Africa, and Oceania—is only partially included, and the differences between the Global South and the Global North—primarily referring to Australia, North America, and Western Europe—are usually not discussed.

The association between urban environments and the risk of nonaffective psychosis has been explored for nearly a century, beginning with Ødegaard’s selective migration theory,⁹ which suggested that immigrants in Norway were more prone to psychosis. This idea was expanded upon by Faris and Dunham¹⁰ in Chicago, who found higher rates of schizophrenia in densely populated, socially deprived areas, although they argued against the “social drift” hypothesis, suggesting that urban environments themselves increased schizophrenia risk due to social isolation. In 1992, Lewis et al.¹¹ challenged the social selection hypothesis, showing in a Swedish study that men raised in cities had a 1.65 times higher schizophrenia rate than those from rural areas, even after controlling for confounders like cannabis use and family history. Despite a decline in interest during the mid-20th century, research in the late 1990s and early 2000s revived the urbanicity-psychosis association, particularly in Northern Europe, where studies showed a consistent link between urban birth and upbringing and an increased risk of schizophrenia.

The relationship between urban living and health extends beyond psychosis, encompassing a wide range of noncommunicable disorders such as cardiovascular disease, diabetes, and respiratory conditions. These disorders share common risk factors with psychosis, including environmental stressors, social isolation, and socioeconomic deprivation. Understanding the broader urban health landscape can provide valuable insights into the specific mechanisms by which urbanicity influences mental health. For instance, the WHO highlights how urban environments can exacerbate health inequalities, particularly in low-income populations, where access to resources is limited.¹²

The urbanicity-psychosis association is considered one of the fundamental epidemiologic findings on environment and psychosis.¹³ Nevertheless, recent studies suggest that the association might be specific to Northern Europe (or similar contexts) rather than universal.¹³ Recent studies in Southern Europe (eg, Italy, Spain) have not found this association.¹⁴ Likewise, recent studies from the Global South also do not support this association.^15–17 These studies from other regions have not used designs of comparable strength to those from Northern Europe. Limitations include small sample sizes and/or unreliable measures. With a few exceptions, they have not used samples representative of a national or regional population.¹⁸ Furthermore, inequities in healthcare access, including the availability and quality of early detection services and help-seeking behaviors, are salient factors that potentially contribute to the variability in the effect of environmental exposures such as urbanicity on mental health outcomes. For instance, in most of the Global South, limited access to mental health services delays diagnosis and treatment, exacerbating the challenges faced by individuals with early psychosis. Thus, whether or to what degree urbanicity is associated with psychosis outside contexts like Northern Europe remains an open question.¹³

A limitation in all studies thus far is that the components of urban and rural life that could lead to a difference in psychosis risk have not yet been well characterized. Studies from Northern Europe have investigated whether factors associated with high population density, such as social fragmentation or social deprivation, might represent aspects of the urban environment that partially explain the observed association. However, these concepts have been operationalized in various ways, and their measures tend to be incomplete.⁸ Furthermore, few people in Northern Europe live in truly rural as opposed to semirural areas, and as a result, little conceptual or empirical work has been done on factors associated with rural (rather than semirural) living that might decrease or increase the risk of psychosis.¹³ Outside Northern Europe, there are few theoretical or empirical studies of how population density corresponds to theoretical concepts of urban and rural that could be relevant to psychosis risk. For instance, Roberts et al.,¹⁹ in 3 population-based studies using the same protocol (controls matched to incident help-seeking cases) and definition for urbanicity (degree of urbanization based on both population density and density of built-up areas) found that urban vs rural categories were locally meaningful in Trinidad and Tobago, but it was difficult to differentiate urban and rural as opposed to semirural in India and Nigeria.¹⁹ Also, in countries where much larger numbers of people live in remote rural areas, their living conditions could be quite different from those living in semirural areas in Northern Europe. For example, in several areas in Latin America, the migration to wealthier urban areas left a residual population in rural areas that have different age structures and social life.²⁰

Furthermore, some argue that relying solely on general indicators of urbanicity, such as population density, does not adequately consider the multifaceted aspects of urban living. Factors such as air quality, access to services, social interactions, and violence and crime collectively shape individuals’ lived experiences. The complexity of the psychosis-urbanicity association has been well-documented in the literature, with numerous studies emphasizing the need for interdisciplinary approaches that integrate insights from epidemiology, sociology, geography, and urban studies.^21–26 This body of knowledge provides a comprehensive understanding of how urban environments interact with biological and social factors to influence the onset and course of psychosis. Using immersive methodologies, such as video ethnography, allows for a deeper exploration of how personal trajectories shape individuals’ interactions with urban environments.^27–30 This approach captures the nuances of daily urban life and provides valuable insights into how these experiences contribute to the development of psychosis. For example, Winz³¹ proposes an approach that focuses on ambience and affective atmosphere, which captures the intricate and nuanced perceptual, sensory, and conscious experiences of individuals residing in diverse urban and semiurban environments. Similarly, Söderström et al.³⁰ suggest an experience-based approach that employs video elicitation and qualitative methods to attain a more detailed understanding of the relationship between the city and psychosis. Drawing evidence from these fields holds great potential to enhance future research in the epidemiology of psychosis.

This review undertook a comprehensive analysis of different concepts and measures of “urban” living and related community-level factors, such as social deprivation and social fragmentation, in the context of psychosis research, including evidence from the Global North and the Global South. We conducted a qualitative systematic literature review of studies published between 2000 and 2024.³² In contrast to narrative reviews, QSLRs use rigorous and transparent procedures to identify, evaluate, and interpret available research while reducing potential biases. QSLRs are particularly useful for synthesizing, contextualizing, and interpreting available evidence. To our understanding, no QSLRs focusing on extant definitions of urbanicity have been conducted to date, which may explain the lack of an accepted range of constructs that can be used to define urbanicity. Moreover, different definitions for the exposure may affect the nature, direction, and strength of the urbanicity-psychosis association, especially in contexts where this line of research is emerging (eg, Global South).¹⁸ The present review fills this gap by examining the thematic components of extant definitions and contributing to the ongoing discussion of how urbanicity should be understood and measured in psychosis research.

Accordingly, our goals were to (1) examine how urbanicity (in contrast to rural as well as to semirural areas) and related factors have been defined and used in the psychosis literature between 2000 and 2024 and (2) identify and characterize core thematic constructs underlying the definitions of urbanicity and related factors.

Methods

We used the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) to develop this review.³³ First, we conducted a search of scientific reports using the following databases: PubMed, Google Scholar, EBSCO, Ovid, Embase, and SciELO. We searched for grey literature using databases such as OpenGrey (http://www.opengrey.eu/). We identified and distinguished countries from the Global South vs the Global North following the United Nations taxonomy.³⁴ We also checked for cities and/or locales within the Global South (eg, Sao Paulo) in case these were mentioned in titles/abstracts instead of countries. Second, we used a set of keywords in English, Spanish, and Portuguese. We kept search terms broad to find relevant studies even when the specific keywords “urbanicity” (or “semirural” or “rural”) or “psychosis” were not mentioned in the title or abstract. For instance, we used several synonyms for the exposure (eg, “cities,” “urban areas,” “neighborhoods”), the related social factors (eg, “social cohesion,” “social harmony,” “social unity,” “social agreement”), and the outcome (eg, “schizophrenia,” “psychotic disorders,” “severe mental illness”) of interest. We combined the proposed keywords for a more precise search and used to identify titles, abstracts, and full texts in the aforementioned databases (see table 1). We searched for articles published from January 2000 to March 2024.

Table 1.

Examples of Searched Keywords per Construct

Urbanicity	“urbanicity” OR “city” OR “urban area” OR “neighborhood” OR “neighborhood” OR “population density” OR “urban/rural” OR “central city” OR “ town” OR “ghetto” OR “semirural” OR “rural”
Psychosis	“psychosis” OR “schizophrenia” OR “psychotic disorder” OR “severe mental illness” OR “first episode psychosis” OR “FEP” OR “affective psychosis” OR “psychotic symptoms”
Community-level social factors	“social factor” OR “poverty” OR “material deprivation” OR “material constraints” OR “material hardship” “material deficiencies” OR “material problems” OR “marginalization” OR “social disorganization” OR “social dislocation” OR “social disintegration” OR “social disruption” OR “social fragmentation” OR “social ills” OR “social unrest” OR “social upheaval” OR “social inequity” OR “violence” OR “social capital” OR “social cohesion” OR “social harmony” OR “social unity” OR “social agreement” OR “harmonious society” OR “social coherence”

We imported identified papers into Covidence (https://www.covidence.org) and removed duplicates. We included articles published in international peer-reviewed journals, including conference papers, book chapters, and editorial materials, if they (1) examined the urbanicity-psychosis association in any country; (2) offered specific, conceptual, and operationalized definitions for urbanicity and related factors; and (3) were written in English, Spanish, and/or Portuguese. We considered a definition to be conceptual if it specified what needs to be assessed in empirical evidence. We identified additional articles by reviewing the reference sections of the articles found in the literature search.

Following the PRISMA guidelines, we reviewed each paper sequentially (searching, refining search strategy, examining titles and abstracts, and reviewing full articles). First, we screened titles and abstracts to exclude records that were not relevant. Furthermore, we excluded records that were only short commentaries, conference abstracts, book reviews, and letters to editors. For all records selected during the initial screening, we retrieved full-text articles and assessed to check whether they fulfilled the inclusion criteria. We conducted no formal assessment of the validity or quality of the full texts as the aim was to identify a broad range of conceptual definitions used in the literature on urbanicity in psychosis research. Considering the lack of a consensual definition of urbanicity described earlier, we felt it would be arbitrary and possibly counter-productive to rate a priori the quality of some definitions higher than others. However, the quantitative design quality of the studies included in this review was carefully considered, particularly those representing unique circumstances, such as natural experiments in rapid urbanization and internal migration settings. These studies offer unique insights into the effects of urbanization on psychosis and are weighted accordingly in our analysis.

Data Analysis and Synthesis

We used a thematic analytic approach to organize and synthesize definitions of urbanicity and related factors in psychosis research. FM, JL, and XY were the main coders for this review. Our approach was informed and validated by the principles of the grounded theory for qualitative research, as well as previous QSLRs.^32,35–37 To remain close to the data in the development of preliminary codes, 4 articles were randomly selected from the initial pool of studies as a starting point for our inductive coding approach. This allows researchers to avoid applying an external framework to the data and the flexibility for the initial codes to emerge directly from the studies. This step facilitates the subsequent focused and theoretical coding, which iteratively synthesize data into themes and, finally, broader constructs and analysis.^38,39 We extracted the following data from each of the full texts: (1) author and publication year; (2) country or countries where the study occurred; (3) study design, study goals, and sample sizes; (4) which types of psychosis (eg, nonaffective psychosis) the study focused on; (5) definition(s) for urbanicity (and semirural and rural) and related factors; (6) data collection method; and (7) main findings. Moreover, for articles that contained more than one definition or description of urbanicity, we included and organized all the definitions line-by-line under the author.

Our approach for coding data entailed the following steps: (1) Initial coding (ie, create categories based on reading definitions from 4 randomly selected publications); (2) Focused coding (ie, use categories to organize definitions inductively based on thematic similarity and shared principles); and (3) Theoretical coding (ie, integrate categories into broader constructs and level of analysis). For each paper, we identified each definition and divided it into meaningful units that we subsequently coded. Then, we coded the definitions in the remaining full texts while continuously extending the coding sheet if new codes emerged while analyzing new full texts and then going back and checking all prior publications with the final coding sheet. Finally, we grouped them into meaningful clusters based on prior research (eg, aggregated them into different dimensions of urbanicity). The 3 main raters initially worked independently to identify emergent themes within the selected articles. This independent analysis was crucial for capturing a diverse range of insights. Following this, the raters collaborated in a series of discussions to integrate their findings, led by the first author, ensuring that no emergent themes were ignored. This approach allowed us to balance individual perspectives with collective agreement, enhancing the robustness of our thematic analysis and maximizing rigor.⁴⁰

Results

A systematic search yielded 2872 records; we excluded 660 duplicates. We screened 2212 abstracts, of which 2006 were excluded, yielding 206 full articles to be included in this review. Of these articles, we excluded 145 papers, mainly because they did not (1) consider urbanicity as the primary exposure or one of the primary exposures (ie, urbanicity was included as a potential confounder in the analysis), (2) examine the onset of psychosis as a primary outcome but rather outcomes that occur after psychosis onset (eg, treatment adherence, mortality, comorbidity), and (3) report operationalized definitions for urbanicity (eg, literature reviews and viewpoints). We included 2 additional papers identified by reference searching. Finally, 61 articles met the selection criteria. A PRISMA flow diagram of the selection process can be found below (See Figure 1).

Fig. 1.

PRISMA flowchart for QSLR.

Characteristics of the Studies

In this review, we analyzed a variety of reports, with the majority being extensive studies conducted in the Global North (which refers mostly to Australia, North America, and Western Europe),^41–51 including the United Kingdom (n = 11),^41–51 Denmark (n = 11),^52–62 Germany (n = 5),^63–67 Sweden (n = 4),^68–71 the Netherlands (n = 4),^72–75 Australia (n = 3),^76–78 Greece (n = 2),^79,80 the United States (n = 2),^81,82 and France (n = 1).⁸³ More recent studies have reported data from regions in the Global South, such as China (n = 3),^84–86 Brazil (n = 1),¹⁵ Chile (n = 1),¹⁷ India (n = 1),⁸⁷ Indonesia (n = 1),⁸⁸ Romania (n = 1),⁸⁹ Taiwan (n = 1),⁹⁰ Turkey (n = 2),^91,92 and Uganda (n = 1).⁹³ We found a few initiatives collecting and/or analyzing data from multiple countries,^{16,19,94–97} 3 of which included data from countries in the Global South.^16,19,94 Most studies in the Global North were based on national and/or regional health registries with clear definitions for urbanicity and well-defined samples. In contrast, several studies in the Global South did not use representative samples, only had cross-sectional assessments, and did not comprehensively characterize the exposure (eg, only measuring urbanicity at admission) and/or the outcome (eg, relying on simple self-report questions about having or not having a “psychotic disorder”). However, recent, well-designed cohort⁹² and case-control¹⁹ studies have reported preliminary results on the association between urbanicity and psychosis in the Global South. We summarized the geographical and methodological characteristics of the included studies in table 2.

Table 2.

Summary of Characteristics of Retrieved Publications

Author and Year	Country/Region	Study Design (1 = Cohort, 2 = Case-Control, 3 = Cross-sectional)	Data Collection (1 = Registry; 2 = Survey, 3 = Other)	Data Collection (Other)
Allardyce 2005	UK	3	1
Bartlett 2007	UK	3	3	Medical record
Binbay 2012	Turkey	3	2
Bosqui 2022	UK	3	2
Budisteanu 2020	Romania	3	3	Medical records
Cantor-Graae 2007	Denmark	1	1
Chang 2019	Taiwan	1	1
Colodro-Conde 2018	Australia	3	3	Data collected from the population genotype study
Del-Ben 2019	Brazil	2	2
DeVylder 2018	World Health Survey (42 countries)	3	2
Dragt 2011	Netherlands	1	2
Eaton 2019	Australia	1	3	Medical records/clinic files
Engemann 2020	Denmark	1	1
Ergül 2022	Turkey	1	2
Gayer-Anderson 2020	6 countries (UK, Netherlands, France, Spain, Italy, Brazil)	2	2
Gonzalez-Valderrama 2022	Chile	3	1
Guloksuz 2015	Germany	1	1
Haddad 2015	Germany	3	3	MRI imaging data
Harrison 2003	Sweden	1	1
Jaya 2018	Indonesia	3	2
Karcher 2021	United States	3	2
Kaymaz 2006	Netherlands	1	2
Khare 2020	India	3	2
Kirkbride 2007	UK	1	2
Kirkbride 2014	UK	3	3	Health service record data
Kirkbride 2017	UK	3	3	Medical records
Lee 2020	UK	1	3	Medical records at hospitals
Lundberg 2009	Uganda	3	2
Luo 2019	China	3	2
McGrath 2001	Australia	2	2
Mimarakis 2018	Greece	3	2
Newbury 2016	UK	1	2
Newbury 2018	UK	1	2
Newbury 2019	UK	1	2
Newbury 2022	UK	1	2
Paksarian 2018	Denmark	2	1
Pedersen 2003	Denmark	1	1
Pedersen 2006a	Denmark	1	1
Pedersen 2006b	Denmark	1	1
Plana-Ripoll 2021	Denmark, Australia	1	1
Radhakrishnan 2019	Netherlands	3	2
Roberts 2023	India, Nigeria, Trinidad	2	2
Sariasian 2015	Sweden	1	1
Saxena 2022	United States	1	2
Schofield 2017	Denmark	1	1
Smeets 2015	Netherlands, Belgium	3	2
Sorensen 2014	Denmark	1	1
Spauwen 2004	Germany	3	2
Spauwen 2006	Germany	1	2
Stefanis 2004	Greece	3	2
Stepniak 2014	Germany	3	3	Medical records, comprehensive examinations;
Sundquist 2004	Sweden	1	1
Szoke 2014	France	1	3	Data collected by psychiatrists treating the new cases
Torrey 2001	Denmark	3	1
van der Leeuw 2020	Netherlands, Belgium	2	2
van Os 2002	Netherlands	3	2
Van Os 2004	Denmark	3	1
Vassos 2016	Denmark	1	1
Wang 2019	China	3	2
Yang 2015	China	3	2
Zammit 2010	Sweden	1	1

Thematic Analysis

The coding process was developed based on a data-driven approach. Extracting information based on each article’s definition, measurements, and categorizations of urbanicity, we were able to reduce it to a set of themes and categories (see table 3). Of note, more than one of these categories could be used in a single study.

Table 3.

Definitions and Examples of Urbanicity and Community-Level Social Factors

	Category/Code	Definition	Example
A	Interconnected constructs
A1	Population density	Population density is the number of individuals per unit geographic area (eg, number per square meter, per hectare, or per square kilometer).	Gonzalez-Valderrama et al. (2020): Urbanicity was based on population density. Expressed in number of residents per squared kilometer and number of residents per squared hectare.
A2	Total population	Total number of individuals in the census or national registry.	McGrath et al. (2001): degree of urbanization was divided into 3 categories according to population at time of birth (“city” ≥ 100 000, “town” > 10 000 and < 100 000, “rural” ≤ 10 000).
A3	Granularity/categorization of urbanicity	Level of detail in the categories used for urbanicity classifications and data analysis (eg, binary measure, 5-level categories)	Bosqui et al. (2020): Urban–rural classifications, referred to herein as “urbanicity,” were calculated based on 8 settlement bandings, ranging from the most urban area (with a population of 75 000 or over) to the most rural (less than 1000 people)
A4	Administrative divisions	Defined urbanicity directly by predetermined administrative division of areas (eg, neighborhoods, provinces, regions)	Vassos et al. (2016): Municipalities in Denmark were classified according to degree of urbanization as follows: capital, capital suburb, provincial cities, provincial towns or rural areas.
B	Urbanicity Measurement Global North vs Global South
B1	Social and Educational indicators	Composite measures composed of several indicators on urbanicity included but not limited to demographic, social, and economic indicators	Chang et al. (2019): Several indicators, including population density (people/km2), population ratio of people with college or above educational levels, population ratio of elderly people over 65 years old, population ratio of people of agriculture workers and the number of physicians per 100 000 people and used the cluster analysis with squared Euclidean distance and Wald’s minimum variance method, to study the urbanization stratification of varied township in Taiwan. Tessa et al. (2023): Urbanicity in India was defined as a minimum population of 5000; at least 75% of the male main workers engaged in nonagricultural pursuits; and a population density of at least 400 per sq. km’
B2	Geographical unit	Urbanicity measured for unit of area (eg, districts, cities, provinces)	Allardyce et al. (2005): The degree of urbanicity is calculated for each postcode sector by adding to the population total the population of each directly adjacent neighborhood: category 1 is most urban, and categories 5 and 6 are the most rural.
B3	Timing	When urbanicity is measured (eg, at birth, upbringing, at first contact)	Newbury et al. (2020): Urbanicity was derived from classifications from 2011 census data (Office for National Statistics, 2013) and linked to participants’ home postcodes at ages 5, 12, and 18.
C	Community-level social mechanisms
C1	Ethnic density	Suggests that the risk of psychosis increases for ethnic minority groups as they live in communities with fewer members of their own ethnic group	Kirkbride et al. (2014): We used the 2001 census to estimate population density (people per hectare), own-group ethnic density, own-group ethnic separation, and social fragmentation. We defined own-group ethnic density in each statistical ward as the size of one’s own ethnic group as a proportion of total neighborhood population.
C2	Social fragmentation	Absence or underdevelopment of connections between individuals and society in a particular geographic area	Zammit et al. (2010): Municipality-level data included measures of urbanicity (city [Stockholm, Gothenburg, and Malmo], town [20 000 inhabitants in 1980], rural [20 000 inhabitants]), population density, and markers of deprivation (derived by summing z scores for mean income, proportion unemployed, and proportion receiving welfare benefits) and social fragmentation (derived by summing z scores for proportion of people migrating in/out of the municipality, voting in municipality elections, individuals married, and single-person households).
C3	Social Deprivation	Limited access to society’s resources due to poverty, discrimination, or other disadvantage	Lee et al. (2020): We adopted the Welsh Index of Multiple Deprivation (WIMD) 2011 as a measure of area deprivation at lower-layer super-output area (LSOA) level, the geographic units used in the calculation of WIMD and the reporting of small area statistics comprised of approximately 1500 individuals (Welsh Government, 2017). Eight different domains of deprivation were assessed, namely, income, housing, employment, geographical access to services, education, health, community safety, and physical environment.
C4	Social capital	How social relations and networks influence collective action for mutual benefit	Ergül et al. (2022): Four dimensions of neighborhood-level social capital were assessed: informal social control, social cohesion and trust, social disorganization, and cognitive social capital. The informal social control scale included 8 questions measuring the willingness to intervene in hypothetical neighborhood-threatening situations such as children misbehaving, using a 5-point Likert scale ranging from “strongly disagree” to “strongly agree.” The social cohesion and trust scale measured bonds and trust among neighborhood residents. The social disorganization scale consisted of 8 questions rating the frequency of certain scenarios occurring in the participant’s neighborhood, such as the presence of graffiti, vandalism, burglary, and racist attacks. The cognitive social capital scale included 3 questions measuring perceptions of support, reciprocity and sharing among the residents of the neighborhood. For each social capital dimension, sum scores were (negative items were reversed) divided by the number of items and aggregated to the neighborhood level.
C5	Walkability and Perception of neighborhood safety	A walkable place is a place easy to walk around based on 3 indicators: design, distance, and diversity. Parents’ views of neighborhood safety based on exposure to crime and violence.	Karcher et al. (2021): A neighborhood walkability index was created based on data obtained from EPA. Karcher et al. (2021): Perception of Neighborhood safety was calculated as a summation of 3 parent-rated questions (ie, “I feel safe walking in my neighborhood, day or night”; “Violence is not a problem in my neighborhood”; “My neighborhood is safe from crime”); each was rated on a scale from 1 to 5.

In summary, the most common category used to define urbanicity was “population density,” accounting for 67.74% of the papers included (n = 42), followed by the codes “granularity/categorization of urbanicity” (n = 34) and “administrative divisions” (n = 30), which were utilized in 54.84% and 48.39% of the papers, respectively. Furthermore, codes pertaining to themes related to neighborhood-level mechanisms were used the least often. We found that “walkability” was used in 5 of the papers, followed by “social deprivation” (n = 4), “social fragmentation” (n = 3), and “social capital” (n = 1).

Through our data analytic approach, we identified a series of codes, presented in table 3. These codes were aggregated into 3 thematic categories that reflect the contents and scope of the definitions of urbanicity and how they have been used when studying the association between urbanicity and psychosis. These categories include: (1) Urbanicity involves several interconnected constructs, (2) Urbanicity measurement varies across studies and regions, and (3) Urbanicity operates through some key neighborhood-level mechanisms.

Urbanicity Involves Several and Interrelated Constructs

As expected, we did not find a single, coherent definition of urban vs rural in the psychosis research literature. Most studies used “population density” as a proxy for defining and assessing urbanicity. However, there was substantial variation in how this construct was operationalized in terms of what constituted an “individual” (eg, residents at a time point, incident vs prevalent, current urbanicity vs birth/childhood urbanicity), a “unit” (eg, per hectare), or a “geographical area” (eg, region).

Moreover, researchers have used other constructs such as total urban and rural population (ie, the total number of individuals in the census or national registry),⁷⁴ administrative divisions (ie, previously determined codes of urban/rural, that is, administrative areas that have been previously defined as urban/rural),⁶² and multidimensional criteria to distinguish urban from rural or semirural areas (eg, dimensions such as hectare grid squared postcodes, and settlement polygons).⁴⁷ Others have noted that urban areas usually entail a series of conditions such as overcrowding, environmental pollution, and violence compared to rural areas.⁹⁰ Other characteristics of urbanicity can include people involved in nonagricultural activities (“at least 75% of the main male workers engaged in nonagricultural pursuits”)¹⁹ and remoteness (“distance from main cities or difficult access”).¹⁹

These measures have different meanings, depending upon what constitutes the geographical region of interest. As noted by Kirkbride and colleagues,¹³ a broad geographical area often contains a mixture of many different social conditions, making it difficult to characterize it as entirely urban or rural under any concept. Small areas may be more socially uniform, but small area characteristics also may not be the appropriate level to measure the concept, for example, if the concept is living within a sizeable dense metropolis. These constructs are often interrelated (eg, population density is used to construct geographical categories or units).

Urbanicity Measurements Vary Between Countries From the Global North and the Global South

Urbanicity measurement varies significantly across studies and regions in psychosis research. Most studies in the Global North, especially in Northern Europe,^50,56,59 have used population density as the main proxy for urbanicity based on national census or registries and have detected a strong association between urbanicity (at birth or upbringing) and psychosis onset.^55,56 In contrast, definitions from some studies in the Global South were somewhat more complex and comprehensive involving level of urbanization (eg, urban areas vs semirural and rural areas)⁸⁹ and local geographical classifications (eg, village, district, province, metropolitan area),⁹² as well as social (eg, agriculture vs nonagricultural jobs)⁹² and educational (eg, population ratio of people with college or above educational levels)⁹⁰ indicators. More recently, however, several studies from the Global South have also used population density as a proxy for urbanicity.^15,17  Table 3 provides further details on the different operationalizations for population density and other indicators for urbanicity.

This variation is partly due to the complex and multifaceted nature of urban environments across countries and regions, which makes it challenging to capture all relevant aspects of urbanicity in a single, standard measure. Some measures may be more relevant in different regions or populations. As shown in the selected studies, measures based on population density may be more relevant in highly urbanized countries with advanced infrastructure (eg, in transportation) (eg, Global North), while measures based on administrative divisions may be more relevant in regions with many people living in urban and even mega-urban areas as well as many people living in dispersed villages and in “mixed” regions. In the Global South, there are often no data available except for entities such as administrative regions, and these are not uniform with respect to urbanicity. As noted in table 3, most predefined administrative divisions are derived from population density and size measures.

Urbanicity Operates Through Community-Level Social Mechanisms

Several studies included in this review suggest that community-level social factors may be potential causal mechanisms for the association between urbanicity and psychosis, such as ethnic density,⁵⁸ social fragmentation,⁷¹ social capital,⁹² and walkability and neighborhood safety.⁸¹ However, these data are still emerging and have not been consistent across countries (even in Northern Europe). There is also variation in terms of how these concepts have been measured. For example, Allardyce and colleagues⁴¹ reported that individuals living in the most socially deprived areas were at higher risk of developing psychosis compared to those in less affluent areas. In this study, social deprivation was assessed using census data on overcrowding, male unemployment, low social class, and not having a car. However, a large Swedish family- and register-based study could not detect an association between population density at first diagnosis (eg, natural log of the absolute population size per square kilometer) and neighborhood deprivation (ie, proportion of individuals with less than secondary school qualifications, proportion not married, proportion not born in the Nordic countries, and neighborhood crime rate) with schizophrenia.⁶⁹

In terms of social cohesion/fragmentation, Zammit and colleagues,⁷¹ in a register-based study in Sweden (n = 203 829), found that the association between living in a city and psychosis was primarily explained by school-level measures as a proxy for neighborhood social fragmentation (ie, proportion of children who migrated into Sweden, moved into a different municipality between ages 8 and 16 years, or were raised in single-parent households). The effect of social fragmentation persisted even after controlling for individual-level confounding factors, such as age and foreign-born status. Furthermore, Kirkbride and colleagues⁴⁵ identified a nonlinear association between levels of social cohesion/fragmentation in neighborhoods (eg, indicators on the presence of graffiti, teenagers, vandalism, attacks due to race or skin color, other attacks, burglary, and theft of or from vehicles in a particular neighborhood) and schizophrenia incidence.

In the Global South, these factors are yet to be fully characterized and examined. Urbanicity in some countries, in both Global North and South, may have positive effects. For example, the availability of health and social resources in cities vs rural areas may moderate (ie, mitigate) the negative urbanicity effects.¹³ Some studies have used complex yet locally sensitive operationalizations. For instance, González-Valderrama et al.¹⁷ measured poverty using a multidimensional index from the Chilean Biennial National Socioeconomic Characterization Survey, which has been administered for over 2 decades, including indicators such as childhood malnutrition, lack of health insurance, deficit in healthcare, school attendance, low level of education, lack of employment, lack of social security, overcrowding, poor structural housing quality, and deficit in essential services.

Discussion

In this review, we examined how urbanicity, a major social determinant of mental health, has been defined in the psychosis research literature. There was not a single, coherent definition of urban vs rural/semirural areas across studies and regions. We discuss the main findings of the review below.

Urbanicity is a multifaceted construct, including not just population density and geographical units, but also encompassing various capabilities such as education opportunities, access to cultural resources, and social engagements. Instead of a monolithic definition based on numbers of people in a particular geographical area, urbanicity should be understood as a dynamic space where individuals’ capabilities can be nurtured or restrained. To truly capture this dynamic, researchers should consider a multidimensional approach to operationalize urbanicity, one that is informed by local conceptualizations of valued capabilities and resources rather than relying on a singular, standardized measure such as population density. As noted by Susser,⁹⁸ the study of urbanicity requires an analysis of historical and contextual factors, such as economy, migration, and politics, rather than focusing exclusively on the number of individuals in a particular area (eg, literature on precarity).⁹⁸ For instance, Ida Susser’s work on urban precarity explores the pressures and rapid maturation demanded by urban life, particularly in contexts of socioeconomic inequality.⁹⁹ Her analysis provides a framework for understanding how these pressures contribute to mental health challenges, such as psychosis, by forcing individuals into accelerated developmental trajectories that can affect psychological well-being.

Moreover, different operationalizations of urbanicity may explain why the consistent results from Northern Europe have not been replicated in other countries in the Global North (eg, Italy, Spain)¹⁹ and in several countries in the Global South.⁹⁴ Kirkbride and colleagues¹³ have noted that these results may represent true differences in the meanings and effects of urban and contextual factors (eg, although “population density” is a standard measure for urbanicity, the nature of living in urban areas may have different meanings for psychosis research). More research in the Global South (and other uncharacterized regions and populations in the Global North) is needed to understand further the particularities and implications of urbanicity compared to semirural and rural areas. Future research should consider local understandings and administrative data. Furthermore, the differences between the Global South and Global North provide an opportunity to reexamine the operationalization of urbanization concepts. While countries in the Global North may share certain similarities, the contextual differences between regions such as North America and Northern Europe are important. Acknowledging these differences allows for a more nuanced comparison of study sites in the Global North, which is crucial for interpreting the effects of urbanicity on psychosis across diverse contexts.

Several studies in the Global South used comprehensive, locally sensitive measures in addition to more standard constructs such as population density (eg, Chang et al.,⁹⁰ Roberts et al.¹⁹). This finding is promising as in some instances rural areas may have higher levels of population density than parts of urban areas in the Global South, which can complicate the use of population density as a measure of urbanicity. However, as noted previously, this finding is subjected to the way urbanicity is defined and operationalized. Similarly, in some regions, administrative divisions may not accurately capture urban environments’ social and cultural dynamics. Therefore, it is important for researchers to consider the operationalization of urbanicity in their studies carefully and to select appropriate measures for the population and region being studied as well as the questions being examined. This may involve developing context-specific measures of urbanicity that consider the unique characteristics of different urban and rural/semirural environments (including potential moderators such as access to healthcare and other social services). The INTREPID II study was full of challenges, but it serves as a good example of using meaningfully informed definitions for urbanicity.¹⁹

Moreover, most studies in this review did not thoroughly characterize the comparison group/area (ie, rural or semirural areas). There has been a great interest in characterizing urban living, but much less is known about the effects of being born and growing up in rural or semirural areas, ignoring the capabilities these areas might offer or lack. There is a dearth of psychosis research in rural areas in both the Global North and the Global South.¹⁰⁰ Rurality, often used as a baseline for studies of urbanicity included in this review, encompasses a range of socioeconomic and environmental factors that differ significantly from those in urban areas. Socioeconomic deprivation, for example, may have different implications in rural contexts where access to services and resources is more limited. These inequities can influence psychosis outcomes just as much as the urban environment. In fact, “truly rural” studies are quite rare, and therefore “semirural” studies are often lumped with rural. This is problematic for psychosis research in both regions but for different reasons. In the Global North, rural areas have been transformed into semirural areas in most countries recently, which has changed the living conditions of residents at several levels. The effects of such changes are yet to be understood, but they may modify the direction and strength of the urbanicity/psychosis association. In the Global South, although the level of urbanization has increased dramatically over the years, there are still many people who live in rural areas and who experience disadvantageous conditions, such as social deprivation and social fragmentation, which are typically found in urban areas in the Global North. Therefore, the varied definitions of urban vs rural across different regions and the transformation of many rural areas into semirural spaces suggest the need to delve deeper into how these environments nurture or diminish capabilities, potentially increasing psychosis risk.

As noted previously, different studies operationalize urbanicity in various ways, reflecting the diverse contexts in which urban environments are studied. These operationalizations range from population density to socioeconomic indicators (eg, employment and education indicators). The variability in these measures suggests that urbanicity may influence mental health through multiple, context-specific pathways, particularly at the community and neighborhood levels. For instance, neighborhood-level factors such as social cohesion, violence, green space, and access to resources may mediate the association between urban living and psychosis. These factors could moderate or mediate the association between urbanicity and psychosis. For instance, access to nature and green spaces (or the lack of) is increasingly recognized as a vital component of urban living, particularly in its role in increasing the risk of nonaffective psychosis.¹⁰¹ Green spaces offer several potential benefits such as stress reduction, improved mood, and enhanced cognitive functioning.^22,102 Although the current paper focuses on urbanicity as the main exposure, we acknowledge that integrating green space into urban health research is essential for a comprehensive understanding of urbanicity’s impact on psychosis. As with urbanicity, the definitions for these factors also vary substantially across studies. There is not a comprehensive, multidimensional approach for measuring and testing the mediational effects and the association between factors, as many could be both mediators and moderators of the urbanicity-psychosis association depending on the research question. Additionally, the characterization of these factors in the Global South is very limited.

Moreover, gender differences in the context of the psychosis-urbanicity association is practically unexplored. Gender plays a crucial role in shaping experiences of city living, influencing factors such as safety, accessibility, and social interactions.² Women, for example, often face unique challenges in urban environments, including higher risks of harassment and violence, which can exacerbate stress-related experiences.¹⁰³ Incorporating gender perspectives into the study of urbanicity allows for a deeper understanding of how different social identities interact with urban spaces and influence mental health outcomes.

Future research on psychosis and urbanicity should aim to address some of the current limitations of the existing literature. We focus on the Global South as many of these gaps have been reported there, although some often apply to the Global North as well. Potential avenues of research include:

• Greater attention to historic and contextual factors: Given the vast cultural, economic, and social differences between the Global South and Global North, and within the Global South, future studies should address contextual factors that may influence the relationship between urbanicity and psychosis. This could include factors such as social norms; local and cultural practices; social, health, and economic inequities; new waves of migration from rural to urban areas within and between countries (eg, South to South migration)¹⁰⁴; and rapid environment changes due to climate conditions, armed conflict, and social and civil unrest. Moreover, we should consider the colonial legacy of drawing arbitrary boundaries between countries that comprise populations without a shared history and culture or creating divisions along these lines that did not exist before. These factors are often the roots of civil unrest now, and they occur in both urban and rural areas but in different ways.

• Use of locally derived measures: To facilitate relevant research on urbanicity and psychosis onset in a particular context, as well as meaningful comparisons across different regions and populations, future studies should use locally informed measures to define and operationalize the exposure and document how these measures align or differ from standard definitions. This could involve the development of region-specific measures that consider local variations in urbanization patterns, change in living conditions over time, and transformation of the healthcare system (and other services). Moreover, “rurality” and its components should be subjected to the same type of key component analysis that “urbanicity” in future work, as this area is largely understudied.

• Focus on the multifaceted aspects of urbanicity: As it has become clear, urbanicity involves a series of contextual, social, and political factors shaping individuals’ experiences and limiting or enhancing human capabilities. Future studies should examine these factors in detail, in particular, their association with mental health well-being and psychosis risk in urban areas. Factors of special relevance include social cohesion, social fragmentation, neighborhood disorder, and religious and cultural traditions. This could involve the use of innovative methods such as spatial analysis and social network analysis.

• Individual-level factors, including personal trajectories and place attachment, influence how individuals interact with and experience urban environments.¹⁰⁵ These factors contribute to the formation of personal niches within urban places, shaping the ways in which individuals engage with their surroundings and the social relationships they develop.¹⁰⁶ The literature in human geography and community psychology on place-making and meaning provides valuable insights into these processes, highlighting how the subjective experience of place can affect mental health outcomes and well-being.¹⁰⁷

• Mixed-methods approaches: Given the complex and multidimensional nature of urbanicity and psychosis, future research should employ mixed-methods approaches that combine quantitative and qualitative data collection methods. This could involve the use of surveys, interviews, and observational methods to provide a more comprehensive understanding of the relationship between urbanicity and psychosis. Research on architecture and urbanism³¹ and urban anthropology,⁹⁸ using ethnographic techniques, can shed light on what it means to reside, interact, and navigate current urban vs rural/semirural areas.

• Multilingual approach: Our literature search was primarily conducted in English, Spanish, and Portuguese due to limitations in our team’s capacity, which may have excluded relevant studies published in other languages, particularly from French-speaking European and African countries. Future research should consider a more multilingual approach to capture a broader spectrum of perspectives, especially in comparative studies between the Global North and Global South.

The urbanity construct, as a composite and ecological framework, has been instrumental in guiding research on the relationship between urban living and psychosis. However, its broad scope may limit our understanding of the specific moderators and mediators that drive these relationships. While there is a case for focusing on individual component factors at the individual or community levels, we argue that refining the urbanity construct—by clearly defining its elements and contextual implications—can enhance its utility in future research. Rather than abandoning the construct, we suggest a more nuanced approach that retains its ecological perspective while allowing for greater specificity.

Conflicts of Interest

The authors have declared that there are no conflicts of interest in relation to the subject of this study.

Funding

None declared.