A Systematic Review of the Effects of Urban Living on Suicidality and Self-Harm in the UK and Ireland

We conducted a systematic review to answer the following: (a) Is there any evidence to support increased prevalence of suicidality and self-harm (i.e. self-harm or suicidality) in urban versus rural environments? (b) What aspects of the urban environment pose risk for suicidality and self-harm? Thirty-five studies met our criteria. Our findings reflect a mixed picture, but with a tendency for urban living to be associated with an increased risk of suicidality and self-harm over rural living, particularly for those living in deprived areas. Further research should focus on the clustering and additive effects of risk and protective factors for suicidality and self-harm in urban environments.

. Despite research increasingly addressing the relationship between the urban environment and mental health, we know comparatively less about the role of the urban environment in suicidality and self-harm [10][11][12].
One review in this area identified a stark urban-rural difference in suicide rates, whereby worldwide rates of suicide were highest in urban areas [13]. However, these figures reflect a one-dimensional conceptualisation of suicide that also masks potential between-country differences. The narrow focus of analysing suicide deaths has issues around accuracy and ignores substantial complexity in this area. Rates of completed suicide are dependent on coroner reports which have inherent flaws and are likely result in an underestimation [14]. Key theoretical models in this area acknowledge the importance of suicidal ideation and intentions, and self-harm in understanding suicide risk [15,16]. Moreover, suicide rates are likely to vary between countries, given the significant heterogeneity in both global urban and rural living standards [17]. In the UK in particular, urbanicity is a pertinent issue. In 2019, 82.9% of England's population were living in urban areas, with future projections predicting further increases [18].
We assert a more localized and nuanced analysis of rural-urban differences in suicidality is needed. Given the growth of urbanicity in the UK and Ireland, we have focussed our review on the literature conducted in these specific localities. We have also widened the definitions of suicide used in previous reviews to include completed suicides as well as suicidal ideation and intentions, and all forms of self-harm. This systematic review aims to answer the following questions: (a) Is there any evidence to support increased prevalence of suicidality and self-harm in urban versus rural environments? (b) What specific aspects of the urban environment pose particular risk in terms of suicidality and self-harm?

Registration
The systematic was registered prospectively on PROSPERO, and no amendments to this protocol were made (reference: CRD42020165785).

Search Strategy and Selection Criteria
The following databases, AMED, BNI, CINAHL, EMBASE, HBE, HMIC, MEDLINE, PsycINFO, and PubMed, were searched from inception to the end of December 2019. Searches were not updated due to the imminent coronavirus pandemic. It was thought that data from the pandemic would be atypical and therefore not representative of typical times, given the established increase in suicidality and selfharm throughout the pandemic [19][20][21]. Additional searches of the reference lists of articles eligible for inclusion were also conducted.
Full search terms are provided in Appendix I. These search terms were derived by reviewing the search terms used in previous systematic reviews on urban living [22] or suicide and self-harm [23].

Inclusion and Exclusion Criteria
Suicidality and self-harm were operationalized as any incidence of self-harm and/or suicidality, including completed or attempted suicide rates and suicide ideation.
To be eligible for inclusion, studies had to (1) present empirical data, (2) be available in English, (3) include a measure of suicide (either actual rates or ideation) and/or self-harm, (4) include a measurement and/or condition explicitly related to an aspect of urban living, (5) report data from the UK and Ireland, and (6) be published in a peer-reviewed journal. We classified aspects of urban living as features of the built environment (e.g. buildings, transport), and the environmental elements of urban living (e.g. access to green space, noise, pollution). Social aspects of the urban environment, namely socioeconomic status, social fragmentation, and characteristics of local communities, were also considered important features of the urban environment.
The following were excluded: (1) protocol, theoretical, or position papers; (2) studies focusing on assisted suicide; (3) studies measuring suicide or selfharm stigma only; (4) studies focussing on rural living only (without an urban comparison); (5) those focussing on murder/homicide preceding suicide; (6) those focussing on substance abuse or overdose (where intent cannot be established); (7) those including a measurement/condition related to geography and/or living but this is not explicitly tied to urbanicity; (8) reports data from the UK that could not be separated from multi-national data; or (9) reports suicide/selfharm rates in the context of persons with a specific condition (e.g. people with a particular physical or psychological difficulty). Studies were not excluded based on the methodology used.
Screening Procedure First, one author screened studies by their title and removed duplicates, and after which, two authors separately screened papers by their abstract, and finally by the full paper. The screening procedure and papers removed at each stage of the screening process are outlined (Fig. 1).

Data Extraction and Quality Assessment
Studies were independently reviewed by the research team, who extracted information relating to (1) author and year of publication; (2) study population, including sociodemographic details or data resource where relevant; (3) definition of urban-rural categorisation; and (4) measures relating to urbanicity and suicidality and self-harm.
Quality assessment was completed independently by three authors, using the National Institute of Health's tool for cross-sectional research, which requires reviewers to rate 14 statements about research quality [24]. Each statement is rated on a three-point scale, good, fair, or poor. The overall quality rating of each study focusses on 14 key concepts (e.g. study population, sample size, confounding variables) and how this impacts study validity. Statements are not used to develop an overall quality score but used to inform an interpretation of overall study quality. One-third of included studies were quality assessed by more than one author; an interrater reliability analysis using the Kappa statistic was performed to determine consistency among the research team.

Data Synthesis
Extracted data was entered into evidence tables showing study characteristics and results. Variation in definitions for urbanicity across studies made the data  unsuitable for meta-analysis. Analysis across studies was completed, comparing characteristics, methods, and findings, and a narrative synthesis of findings was applied to summarize the strength of evidence. Given the financial, social, and environmental changes in urban environments, the time period in which data was collected is highlighted. As there is no established time period in which these changes to urban environments took place, we used a descriptive assessment of time periods where appropriate.
For the first research question, evidence to support increased prevalence of suicidality and self-harm in urban versus rural environments, unadjusted and adjusted comparisons were extracted from those studies that compared rates of suicidality and self-harm across rural and urban environments. For the second research question, aspects of the urban environment that pose particular risk in terms of suicidality and self-harm, only those studies that specifically reported correlates of suicidality and self-harm in urban environments were included. Where risk factors for suicidality and self-harm were associated with the living environment in general, but not urban living specifically, data was not extracted.

Results
We identified 35 relevant studies that met the inclusion criteria in the search (Fig. 1). An overview of the results from each study is provided (Table 1).

Study Characteristics
Sample sizes ranged between 95 and 366,348 persons; these were individual participants, as opposed to events of suicidality and self-harm (one participant may experience more than one event of self-harm or suicide attempt). Eight studies did not record the total number of persons [29,31,34,35,37,39,41,50]. Most studies were conducted in England only (n = 13), with eight in the Republic of Ireland, six in England and Wales, five in Scotland, two in Northern Ireland, and one in Wales only.

Definitions of Urban-Rural Environments
Definition and categorisation of urban areas varied greatly across included studies. Population density was used as a proxy indicator for levels of urbanicity across most studies (40%), three inferred population density by participants home residence, two used the proportion of green space, and two used a combination of population density, proximity to the city center, and the proportion of green space. Four studies did not provide a definition and two used "statedefined" classifications but did not expand further. Five studies were completed across urban cities (Edinburgh, London, Wolverhampton, Bristol, with one study focussing on three cities, Oxford, Derby, and Manchester) [26,33,42,47,55], and one divided the city of London into inner and outer city areas [29]. One based the definition of urban-rural environment on a combination or variables derived from factor analysis [34].

Measurement of Suicidality and Self-Harm
Suicidality and self-harm indicators varied, with 12 (34%) studies measuring this as intentional self-harm, 20 (57%) as suicide, and 1 as suicide attempt. One study used a combination of intentional self-harm and death recorded as suicide [50], and another as a combination of suicide attempt and death recorded as suicide [29].
Instances of suicide were assessed via publicly held records (e.g. coroner's reports and Census data); one assessed suicide attempt through referrals for psychiatric assessment post suicide attempt [28]. Measurement of self-harm was assessed via local self-harm registries based on hospital presentations for self-harm across all studies except Gunnell et al. [33] who initially assessed self-harm through a survey across three emergency departments in the 1970s, and then followed this up in the 1990s with the local self-harm registry.

Quality Assessment
Inter-rater agreement for quality assessment for included studies across reviewers was moderate (Kappa statistic = 0.74-0.79); nineteen of the included studies were determined to be of good quality (54%). Notably, ten studies did not report the age or gender of their population. Fourteen studies did not include any key confounding variables to adjust for the impact of relationships between urban-rural environment on suicidality and self-harm. Confounding variables that were adjusted for included gender or age, as well as area-level socioeconomic deprivation or social fragmentation, both of which were defined by established measures obtained from Census data. Only five studies examined different levels of urbanicity in relation to suicidality and self-harm [43,44,49,52,57].

Evidence to Support Increased Prevalence of Suicidality and Self-Harm in Urban Versus Rural Environments
Twenty-eight studies provided unadjusted rural-urban comparisons for suicidality and self-harm. The majority (54%; 6 assessing self-harm, 9 assessing suicide) reported significant associations between greater urbanicity and increased rates of suicidality and selfharm. Rates of suicide ranged from 6.16-28.35 per 100,000 in urban areas, and 1.78-10.5 per 100,000 in rural areas [25,27,32]; one study reported rates of self-harm to be more than three times higher in more densely populated areas (IRR = 3.47, 95% CI = 3.08-3.92) [59].
Although reporting associations with suicidality and self-harm across urban-rural environments, several authors noted challenges with these interpretations, given inconsistencies within rural and urban environments [25,31,41,43]. For example, overall, Corcoran, Arensman, and Perry reported higher rates of self-harm in urban (vs rural) environments, but reported lower rates of self-harm in Dublin, the capital city of Ireland, than other Irish cities, despite being the most urbanized city [43].
An additional four studies assessed changes in rates of suicidality and self-harm across urban-rural environments between 1972 and 1996; again, results were inconclusive [33][34][35][36]. Over time, two studies reported an increase of suicidality and self-harm rates in urban environments [24,33,34]. However, two further studies reported a narrowing of urban-rural rates of suicidality and self-harm over time, resulting from an increase in rates of suicidality and self-harm in rural environments over more recent years [35,36].

Adjusted Associations
Of the fifteen studies that reported greater rates of suicidality and self-harm in urban (vs rural) areas, six (40%) adjusted for potentially confounding variables within the urban environment. Two adjusted for area level social fragmentation and socioeconomic deprivation [43,59]; small but significant relationships remained between urbanicity and self-harm after adjustment in both these studies. Five studies adjusted for area level socioeconomic deprivation only [32,37,40,43,49]; this socioeconomic deprivation largely explained higher rates of suicide in urban areas (3/5 studies). Corcoran et al. [43] reported instances of self-harm were still greater in urban environments after accounting for area level socioeconomic deprivation. Gartner et al. [49] found that adjustment for deprivation changed the direction of relationships; although deprivation explained urban-rural differences in female suicide, after adjustment, suicide appeared greater for men in rural areas. The authors concluded that not adjusting for deprivation appeared to mask the increase in male suicides in rural areas. One additional study adjusted for individual living circumstances and clinical factors [51], after adjustment, an independent positive association remained between self-harm and urban residence.
For those that reported greater rates of suicidality and self-harm in rural (vs urban) environments, after adjustment for area level, socioeconomic deprivation and/or fragmentation did not change the strength of risk of suicide associated with rural environments in 3/5 studies [36,54,56]. In contrast, two studies reported that population density was not associated with suicide [39] or self-harm [57] when adjusting for area level socioeconomic deprivation.
Aspects of the Urban Environment that Pose Particular Risk in Terms of Suicidality and Self-Harm Twenty (57%) of the included studies assessed at least one aspect of the urban environment in relation to the risk of suicidality and self-harm. Aspects of the urban environment assessed included ethnic diversity of the area, area level socioeconomic deprivation, social fragmentation, crime, and features of the built environment.

Features of the Built Environment
Surprisingly, only two studies assessed the role of environmental features. Bixby et al. [53] reported no association between the presence of urban green space and suicide rates across England, whereas McCulloch et al. [26] reported greater suicide rates in Scottish urban areas with greater overcrowding and tenement housing.

Urban Crime
Area-level crime or juvenile delinquency [26,32] was associated with increased rates of suicide in urban areas. In a review of coroner's records within the city of Bristol, areas with high homicide and violence were associated with increased instances of suicide [32].

Ethnic Diversity
Three studies assessed ethnicity as a risk factor for self-harm (n = 2) and for suicide (n = 1). Self-harm in the city of Oxford, was more likely to be completed by "non-white ethnic" individuals, 48 and across the cities of Oxford, Manchester and Derby self-harm was more common in young black females [47]. In comparison, there was no association between ethnicity and suicide in London [32].

Socioeconomic Deprivation
Area-level deprivation was assessed using a variety of measures; the most common was the Townsend Index (n = 4). Twelve studies identified area-level indicators of socioeconomic deprivation as a risk factor for suicidality and self-harm in urban environments, irrespective of the location, date of data collection, or suicidality and self-harm type. However, local variations in suicidality and self-harm were apparent across urban areas; some urban areas had lower suicide rates than expected, given their high socioeconomic deprivation scores [32,36,40].
The effect of urban environment varied by gender; two studies completed by the same author, reported on the gendered effects of socioeconomic deprivation. For males, suicide attempts and completed suicide in urban environments were most strongly influenced by socioeconomic factors [29,50].

Social Fragmentation
All four studies assessing area-level social fragmentation, typically assessed via the Congdon index [29], concluded that great area-level social fragmentation increased the risk of suicidality and self-harm in urban areas [33,43,48,52]. However, on further examination, for young adults in Dublin, areas of higher area-level social fragmentation were associated with lower rates of self-harm [43]. The authors suggest this finding as an artefact of the measures of social fragmentation, which may be limited in inner city areas characterized by a young, unmarried population, who more often live alone. Again, gendered effects were apparent. Females attempting or completing suicide in urban settings appeared most influenced by social factors, in contrast to males who were mist influenced by socioeconomic factors [29,50].

Discussion
This systematic review identified 35 studies reporting on suicidality and self-harm in urban environments across the UK, prior to the coronavirus pandemic. We identified varied and often contradictory outcomes across studies, which were often limited by definitions of urbanicity, and measurement of suicidality and self-harm. Across most included studies, living in an urban environment was associated with an increased risk of suicidality and self-harm, compared to rural living, but findings were inconsistent.
Area-level socioeconomic deprivation and social fragmentation appeared to increase the risk of suicidality and self-harm for those living in urban environments. This result is not surprising, as both have been highlighted as key risk factors for suicidality and selfharm across several academic, clinical, and policy reports [7,60]. However, relationships between deprivation, fragmentation, and suicidality and self-harm were not as consistent as we would perhaps expect, with variation apparent within urban environments and evidence for gendered effects of urban living on suicidality and self-harm; males in urban areas appeared more influenced by socioeconomic factors, whereas females in rural areas appeared more influenced by social fragmentation [29,50]. These variations may result from the way in which arealevel fragmentation and deprivation is assessed. For example, Congdon suggests that indicators used to measure socially fragmented communities may not measure fragmentation, but rather younger communities with young professionals or students, especially within urban areas [29]. Similarly, assessment of socioeconomic deprivation relies on Census data which is collected every 10 years. However, urban environments are susceptible to change within short periods of time, now so, more than ever before, with the coronavirus pandemic, which Census data may fail to capture [61].
Community or environmental factors were rarely assessed in relation to suicidality and self-harm within urban environments. Notably, only one study explored the role of green space, reporting no association with suicide rates across England [53]. This is surprising, given well-established theoretical models, which point to the importance of considering the resources and environmental characteristics of urban communities that might protect against impaired mental wellbeing [62]. Across the UK and Ireland, urban areas have undergone increasing gentrification, with greater investments in housing, and the introduction of resources and services [63]. Aspects of the local environment, including community support, availability of public transport, and green space, can act as protective factors providing individuals with resources to cope with stressors [64,65]. Without knowing more about how multiple factors interact to influence suicidality and self-harm in urban environments, it is impossible to develop interventions that address this real-world complexity.
Those living in urban environments may be disadvantaged on many levels, experiencing increased crime, social fragmentation, socioeconomic deprivation, poorer quality housing, and/or limited access to green space [66]. It is likely the cumulative stress of these factors, in combination with protective factors, which interact to impair wellbeing. To further understand the interactions between suicidality and self-harm within urban environments, greater insight on the interrelationships between social context, environment, and suicidality and self-harm is required. As summarized by Curtis et al. [67], individual factors, familial attributes, characteristics of the local community and the wider national or regional context are all likely to interact to influence wellbeing. The low rates of suicidality and self-harm in urban environments described across two studies included within this review [43,57] highlight the need to consider the context of urban environments, and the ways in which suicidality and self-harm is experienced in different urban settings and across different populations [68].

Future Directions
This review highlights important gaps in the design and evaluation of research that can help answer these questions. We suggest that research, policy, and practice need to go beyond the urban-rural division, focussing on the characteristics of local communities, and how interactions with local environments, spaces, and communities modify risk factors for suicidality and self-harm. The UK government's focus on preventative and integrated approaches to care at the local level, which coordinate health and social services to meet the needs of the local community, provides a promising basis for further work [69]. In line with these frameworks, high-quality, longitudinal analyses of routinely collected data may be beneficial in exploring how these variables cluster and interact, whilst qualitative approaches have the potential to help refine population and exposures for these analyses and help identify key experiences that help individuals overcome adversity within urban environments.
From a theoretical perspective, there is a need to explore the complexity of the urban environment and its relationship with suicidality and self-harm. We are now starting to see intersectional approaches applied to our understanding of urban environments and suicidality and self-harm [70,71]. Intersectional approaches move us beyond considering single social determinants, such as socioeconomic deprivation, instead, considering these in combination with social processes and environmental influences (e.g. social support, employment, green space). An intersectional framework may provide more precise identification and understanding of suicidality and self-harm in urban areas, prioritising the voice of those most affected by these issues.

Strengths and Limitations
This is the first systematic review to assess the implications of urban living on suicidality and self-harm in the UK and Ireland and includes a comprehensive variety of studies conducted over time. We have been able to draw conclusions regarding urban-rural differences in suicide and self-harm risk based on 35 studies. Our conclusions regarding specific urbanrelated risk factors however are based on a smaller sub-set of studies and should be interpreted in light of this limitation.
Despite no restrictions on methodology, most studies included in this review were cross-sectional in nature, meaning no conclusions can be made about the causal effect of urban living on suicidality and self-harm. We acknowledge that the detailed nature of our inclusion criteria may have prevented the inclusion of qualitative studies but highlight that no qualitative or mixed method studies were included in the identified eligible studies.
To be eligible for inclusion, articles were required to assess aspects of the urban environment and suicidality and self-harm, definitions for which varied across included studies. Urban environment was largely defined via population density, although varied widely across studies. In addition, suicidality and self-harm was largely assessed via publicly held records (i.e. coroner reports). As such, the rates of suicide and deliberate self-harm reported here may be an underestimation due to the high burden of proof required to declare a death a suicide [72], and hospitals' poor recording of nonadmitted self-harm cases [73]. The results presented here should be interpreted with a clear understanding of the time in which each study collected data, given the changes in urban living over recent decades, as well as each studies definition of the urban environment. To aid interpretation, we present descriptions of the time periods in which data was collected, where relevant, and include information on this and definitions of the urban environment within our tables.

Conclusions
There is a need for high-quality, theoretically informed research to further understand and inform preventive, local interventions to address suicidality and self-harm in urban environments across the UK and Ireland. Drawing on our findings, we highlight the limitations of urban-rural distinctions in an increasingly complex world, prioritising a focus on the relationships between urban living, protective and risk factors, as well as individual experience.
Free text words ("suicid*" OR "overdos*" OR "self?harm*" OR "self?injur*" OR "self?cut*" OR "self?destruct*" OR "auto?mutilat*" OR "auto?destruct*" OR "self?inflict*" OR "self?poison*" OR "self?mutilat*") AND ("Moderni?ation" OR "urban*" OR "rural*" OR "open space*" OR "park*" OR "green" OR "wood*" OR "forest*" OR "garden*" OR "environment*" OR "communit*" OR "grow*" OR "city" OR "civili?at*" OR "neighbo?rhood" OR "geography" OR "public space" OR "natur*" OR "landscape" OR "tree*") MeSH Field Title; Abstract Limits None Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.