Prevalence of mental disorders among young people living in urban slums of low- and middle- income countries: A systematic review and Meta-analysis Authors: Tolulope Bella-Awusah, Oluwabunmi Fola – Bolumole, Sagar Jilka, Ursula Read, Olayinka Omigbodun, Swaran Singh, Domenico Giacco Published: 2026 ABSTRACT AbstractPurposeYoung people living in slums are at increased risk of poor mental health. We examined the prevalence and correlates of mental disorders among youth living in slums in low-and middle-income countries (LMICs).MethodsA search of seven databases (MEDLINE, Embase, PsycINFO, Scopus, Web of Science, CINAHL, and African Journals Online) was conducted from inception to September 2022 and updated to March 2026. Studies were screened by two independent reviewers who also extracted data and assessed the study quality and risk of bias using the Joanna Briggs Institute critical appraisal tool (JBI). Meta–analyses were carried out using the random effects model (REML) to estimate prevalence and 95% CI.ResultsWe identified 24 studies out of 3,946 citations, reporting data on depression, post- traumatic stress disorder (PTSD), anxiety, and alcohol use disorders (AUDs), involving 26,794 participants. The pooled prevalences were depression 31% (95% CI: 0.23–0.40; I2 99·4%; p < 0·01), PTSD 24% (95% CI: 0.14–0.35; I2 99·08%; p < 0·01), anxiety disorders 35% (95% CI: 0.20–0.53, I2 98·91%; p < 0·01), AUDs 23% (95% CI:0.10–0.38, I2 99·08%; p < 0·01). Food insecurity and violence exposure were the most reported risk factors, whereas adult support and neighbourhood connectedness were the most reported protective factors.ConclusionYoung people living in slums carry a disproportionate burden of mental disorders. There is an urgent need for contextually grounded policies and interventions to mitigate identified risk factors, strengthen protective factors, and break down barriers to treatment for those in need.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s00127-026-03154-9.Keywords: Young people, Mental Health, Urban Slums, Prevalence, Systematic Review, Low- and- middle income countries Introduction Mental disorders are one of the leading causes of disability among young people globally, and nearly half of all lifetime mental disorders emerge by the age of 14 [1, 2]. Despite this, most affected youth particularly those who live in low- and middle-income countries (LMICs) often do not receive formal diagnoses or treatments [3, 4]. This treatment gap is often greatest among those already facing greater structural disadvantages such as those living in informal urban settlements known as slums [5, 6]. Globally, an estimated one billion people live in slum settlements, and it is projected that this figure will continue to rise as urbanisation accelerates across sub-Saharan Africa and South Asia [7–9]. Young people make up a large proportion of slum dwellers, yet they remain largely invisible in the mental health literature. Life in slums is often characterised by a concentration of chronic adversities which have well-established associations with poor mental health [10]. Extreme poverty, overcrowding, inadequate sanitation, food insecurity, exposure to community violence, disrupted schooling, and limited access to health services all co-occur in slum environments [8, 11, 12] where they interact synergistically, leading to compounded risks across biological, psychological, and social spheres of life [13, 14]. The resulting ecology of adversity in slums therefore appears distinct from non-slum urban or rural ecologies, and should not be treated as a subcategory of either [13]. A growing body of research has begun to examine mental health outcomes among youth living in slum settings (YLS), with reports of elevated rates of depression, anxiety, post-traumatic stress disorder (PTSD), and alcohol and substance use disorders [15–17]. However, the available evidence remains largely fragmentated making it difficult to draw reliable conclusions about the true prevalence of mental disorders in this population, and to identify risk and protective factors which could be targets of interventions. Therefore, this systematic review aims to address a critical gap in the global mental health literature by synthesising evidence on the prevalence and correlates of mental disorders among YLS. Methods Protocol and registration This systematic review and meta-analysis was conducted according the PRISMA guidelines [18]. The review protocol was registered with PROSPERO (ID number CRD42022359756). Eligibility criteria The inclusion criteria were as follows: studies reporting point prevalence or period prevalence of up to 12 months on disorders from F1 to F4 on the ICD-10 classification of mental disorders; studies that used validated diagnostic and/or self-report instruments; study population had to reside in an urban slum area as identified by the article authors; or the study provided sufficient information to determine whether households lacked one or more of the five indicators used by UN-HABITAT [19] to identify slum households. These five indicators are (1) insecure residential status, (2) poor structural quality of housing, (3) overcrowding, (4) inadequate water access, and (5) inadequate sanitation access. Eligible studies could be observational, cross-sectional, case-control, cohort studies, randomized controlled trials (RCT), and nonrandomized controlled trials (nRCT), if they reported the prevalence of the selected disorders in the selected population. Studies using secondary data or based in clinical settings of LMICs were excluded. LMICs were identified based on World Bank classification [20]. Search strategy The following seven databases were searched: MEDLINE, Embase, PsycINFO, SCOPUS, Web of Science, CINAHL, and African Journals Online. OpenGrey was listed as a planned grey literature source in the registered protocol but was not searchable due to discontinuation of the source. The original search was conducted on 1 September 2022. Search terms relating to youth, mental health, poor urban environments, and LMICs were developed in collaboration with an information specialist. The reference lists of included papers were manually searched for additional studies. There were no restrictions on search dates or language. During the peer review period, the search was updated to 12 March 2026 to ensure that the evidence base was up to date. The search strategy was expanded to include additional terms for ‘prevalence’ and ‘poor urban environments’ to improve comprehensiveness. (Supplement 1). Study selection After removing duplicates, the titles and abstracts were screened by two independent reviewers (TB-A and OF-B). If the title and abstract did not provide sufficient information to determine eligibility, the article was retained for a full-text screening. The full texts of the articles that were found eligible or unclear at this stage were retrieved and screened again by two independent reviewers. Disagreements between reviewers were resolved through discussion or in cases of persistent disagreement, by a third adjudicating reviewer (DG). Data extraction and preparation Two independent reviewers, TB-A and OF-B A used a standardised Excel spreadsheet to extract data from all included studies. The following data were extracted from each study: author and year, country, study design, sampling method, age range, mean age, sex ratio, mental disorder(s), assessment tools, diagnostic criteria, total sample size, risk factors, and protective factors. Risk of bias assessment The risk of bias assessment of the included studies was performed by two independent reviewers, TB-A and OF-B, using the Joanna Briggs critical appraisal tool (JBI) for prevalence studies [21]. The checklist rates studies using nine criteria, such as appropriateness of the sampling frame, sampling method, sample size, rigour of data analysis, clear target population definition, appropriate measurement, valid/reliable tools, appropriate statistical analysis, and response rate. Each question is rated as Yes, No, Unclear or Not Applicable. Each Yes response was assigned a score of one point. Thus, the studies were scored on a scale of 0–9 points (Supplement 2). Data analysis The extracted data were imported into Stata software version 18 (StataCorp LP, College Station, TX, USA) for meta-analysis. Meta-analysis results were expressed as prevalence estimates of mental disorders calculated with 95% CIs in the pooled data. The Random effects model was used, assuming heterogeneity among the studies (using the REML model). This statistical model assumes that the samples of the included studies were drawn from different populations. Heterogeneity was assessed using the I2 statistic. The I2 value provides a measure of the variation explained by the differences between the included studies rather than by chance [22]. Possible sources of heterogeneity between studies were investigated, where reported data allowed subgroup analysis using the Meta suite in Stata 18.0. These included age, region, sampling method, risk of bias, study instruments, and mode of administration of the assessment tool. A narrative synthesis was also conducted on the individual-level correlates of the mental disorders identified in the review. Results Search results Figure. 1 A search of the seven databases retrieved 3,946 citations for review. These were imported into the Endnote 25 reference manager, and 116 duplicates were removed, leaving 3,830 articles for screening. These were then imported into the Rayyan Artificial Intelligence software [23] for title and abstract screening. A total of 308 articles were included for full-text screening, and 21 articles were finally included in the review (Fig. 2). One article reported a multisite study with four independent samples, each treated as a separate study in the review resulting in 24 independent studies. PRISMA 2020 flow diagram Forest plots of pooled prevalence for Depression and PTSD Characteristics of the included studies Twenty-one articles (comprising 24 independent studies) were included in the review [15, 24–43]. All the articles were published between 2012 and 2026. The studies were from three regions: thirteen from Africa (Nigeria (n = 1), South Africa (n = 7), Kenya (n = 4), Uganda (n = 1), nine from Asia (India (n = 5), China (n = 1), Bangladesh (n = 2), Thailand (n = 1), and two from South America (Brazil and Paraguay). A total, 26,794 participants were included in this review. Of these, 11, 748 were males and 14,769 were females and 277 were from one study which did not report the sex breakdown. The studies reporting sex breakdown resulted in a female proportion of 0·55. All studies were cross-sectional in design, and 13 used non-probability sampling methods (Table 1). The ages of the participants ranged from 12 to 30 years, but only six studies specified the mean age and SD of the sample population. Five studies had 100% female participants, and three studies had 100% male participants. Table 1Characteristics of included studiesS/NAuthorYearRegionCountryStudydesignSamplingmethodAgerangeGenderRatio (female)Sample sizeMental disordersAssessment ToolMode of administration1. Avanci 20212021South AmericaBrazilCross sectionalMultistage cluster sampling13 to 1965862PTSDUCLA PTSD reaction IndexSelf2. Cheng et al. 2014 a.2014AfricaNigeriaCross sectionalRespondent driven sampling15 to 1951449PTSDDepressionCESD − 10; (cut off 11) PTSD checklist civilian versionSelf3 Cheng et al. 2014 b.2014AsiaIndiaCross sectionalRespondent driven sampling15 to1950500PTSDDepressionCESD − 10; PTSD checklist civilian versionSelf4 Cheng et al. 2014 c.2014AsiaChinaCross sectionalRespondent driven sampling15 to1949.3438PTSDDepressionCESD − 10; PTSD checklist civilian versionSelf5 Cheng et al. 2014 d.2014AfricaSouth AfricaCross sectionalRespondent driven sampling15 to 1945.2497PTSD DepressionCESD − 10; PTSD checklist civilian versionSelf6 Gibbs et al. 20182018AfricaSouth AfricaCross sectionalSnowball sampling18 to 3049.8213DepressionAUDCES D −20 (cut off 16)AUDITSelf7 Harder et al. 20122012AfricaKenyaCross sectionalStratified random sampling12 to 1853.5275PTSDAUDUCLA PTSD reaction index; Clinician PTSD scale; AUDITClinician rated(PTSD), self8 Nasreen et al. 20162016AsiaBangladeshCross sectionalMultiple random sampling13 to 19501220DepressionBeck’s depression InventoryInterviewer9 Ndungu et al. 20202020AfricaSouth AfricaCross sectionalPurposive and snowball sampling18 to 30100680PTSD, Depression AUDAUDIT (cut off 8); CESD: Harvard Trauma QuestionnaireInterviewer (DEP), self10 Bantjes et al. 20182018AfricaSouth AfricaCross sectionalSystematic random sampling18 to 290647DepressionCES-D 20 (cut off 16)Interviewer11 Barman et al. 20152015AsiaIndiaCross sectionalSystematic random sampling15 to 19100300DepressionAnxietyKUTCHER Depression scale;GAD − 7Interviewer (DEP), self12 Hatcher et al. 20192019AfricaSouth AfricaCross sectionalCluster and convenience sampling18 to 3002427DepressionCES-D −20 (cut off 21)Interviewer13 Mridha et al. 20212021AsiaBangladeshCross sectionalCluster random sampling10 to 1950.61206DepressionPHQ 9Interviewer14 Rani et al. 20182018AsiaIndiaCross sectionalStratified two stage random sampling13 to 19100418DepressionAnxietyMental Health Inventory MHIInterviewer15 Somrongthong et al. 20132013AsiaThailandCross sectionalSystematic random sampling12 to 2257.2871DepressionCES- D −20 (cut off 22)Self16 Wado et al. 20222022AfricaKenyaCross sectionalRandom sampling12 to 191002106PHQ − 9Self17 Daniel et al. 20222022AsiaIndiaCross sectionalCensus based purposive sampling15 to 1945.23490DepressionPHQ − 9Interviewer18 Friedberg et al. 20232023AfricaKenyaCross sectionalCluster randomised trial10 to 1479.84095PTSDDepressionAnxietyCPSS; CDI-2; BAISelf19 Kaufman et al. 20142014AfricaSouth AfricaCross sectionalPurposive cluster sampling12 to 2055.64484AUDAUDITSelf20 Swahn et al. 20252025AfricaUgandaCross sectionalConvenience sampling18 to 24100300DepressionAnxietyPHQ- 9; GAD-7Interviewer21 Mahanta et al. 20262026AsiaIndiaCross sectionalRandom sampling12 to 1949.2276AUD Tobacco use disorder Cannabis use disorderWHO ASSISTSelf22 Musyoka et al. 20252025AfricaKenyaCross sectionalSnowball sampling15 to 2445.794DepressionAnxietyPHQ − 9;GAD − 7Interviewer23 Oyekunle et al. 20232023AfricaSouth AfricaCross sectionalSnowball sampling18 to 300669PTSD, DepressionAUDCES-D 20; Harvard Trauma Questionnaire; AUDITInterviewer24Vujk et al. 20252025South AmericaParaguayCross sectionalPurposive sampling12 to 18unknown276DepressionAnxietyDASS- 21Self Characteristics of included studies PTSD Depression PTSD Depression PTSD Depression Depression AUD CES D −20 (cut off 16) AUDIT PTSD AUD Clinician rated (PTSD), self Depression Anxiety KUTCHER Depression scale; GAD − 7 Depression Anxiety PTSD Depression Anxiety Depression Anxiety Depression Anxiety PHQ − 9; GAD − 7 PTSD, Depression AUD Depression Anxiety Nine studies examined PTSD using data from 8,465 participants [15, 24, 27, 29, 35, 36]. Twenty studies examined depression using data from 20, 896 participants [25–30, 33–43]. Six studies examined anxiety disorders with data from 5,483 participants [26, 29, 33, 37, 39, 41], and five studies examined alcohol use disorders with data from 6,323 participants [30–32, 35, 36]. One study examined tobacco and cannabis use disorders [32]. Only one study used a clinician’s rating scale in addition to a screening scale [15]. All other studies used self-rated or interviewer-rated, validated screening scales. For depression, 9 studies [25–27, 30, 36, 38] used the Centre for Epidemiological Studies Depression scale CES-D 10 item and 20 item versions [44]. One study [34] used the Beck Depression Inventory (BDI) [45], one study used the Beck child depression inventory (CDI) [29], five studies [28, 33, 40, 41, 43] used the Patient Health Questionnaire (PHQ- 9) [46], one study [37] used the Mental Health Inventory (MHI) [47], one study [26] used the Kutcher Adolescent Depression Scale (KADS) [48], and one study [39] used the Depression and Anxiety Stress Scale [49]. For PTSD, four studies [27] used the PTSD checklist civilian version – the 6-item short form [50], and two studies [15, 24] used the University of California at Los Angeles (UCLA) PTSD reaction index [51]. One study [15] used both the UCLA PTSD reaction index and the clinician-administered PTSD scale [52]. Two studies [35, 36] used the Harvard Trauma Checklist [53] and one study [29] used the Child PTSD symptom scale [54]. For anxiety, one study [37] used the Mental Health Inventory (MHI), one study [39] used the Depression and Anxiety Stress Scale [49], three studies [26, 33, 41] used the Generalised Anxiety Disorder Scale (GAD-7) [55], and one study [29] used the Beck’s Anxiety Inventory [56]. For Alcohol Use Disorders (AUD), four studies [30, 31, 35, 36] used the Alcohol Use Disorders Identification Test (AUDIT) [57] and one study [32] used the WHO Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) [58]. A total of 13 studies [15, 27–29, 31, 33, 36, 38, 41, 43] mentioned cultural adaptations of the screening instruments used, while 15 studies [15, 27, 28, 31, 33, 35–38, 40–42] mentioned administration in the local language in addition to or alternatively to the English language. Studies with JBI scores of less than 6 were rated as moderate to high levels of bias, and nine out of the 24 studies met this criterion. All studies were included in the meta-analysis regardless of quality. Prevalence of mental disorders The pooled prevalence for depression was 31% (95% CI: 0.23–0.40, N = 20,896), with a range of 7·0–73.0%, and heterogeneity I2 of 99·4% (p < 0·01), (Fig. 2).The pooled prevalence for PTSD was 24% (95% CI: 0.14–0.35, N = 8,465), with a range of 8.0–60.0%, and heterogeneity I2 of 99·08% (p < 0·01) (Fig. 2). The pooled prevalence of anxiety disorders was 35% (95% CI: 0.20–0.53, N = 5,483), with a range of 13.0–64.0%, and heterogeneity I2 of 98·91% (p < 0·01) (Fig. 3). The pooled prevalence for alcohol use disorders was 23% (95% CI:0.10–0.38, N = 6,323), with a range of 4.0–44.0%, and heterogeneity I2 of 99·08% (p < 0·01) (Fig. 3). One study [32] reported prevalence of tobacco and cannabis disorders as 13% and 0.7% respectively. Forest plots of pooled prevalence for Anxiety and Alcohol use disorders Sub-group analysis For depression, prevalence differed significantly by region (p < 0.001), sampling method (p = 0.03), and measurement tool (p < 0.001), but not by mode of tool administration, risk of bias, or age group. Regionally, South America had the highest prevalence (54%, 95% CI: 0.48 − 0.60), followed by Africa (36%, 95% CI: 0.25–0.48) and Asia (22%, 95% CI: 0.13–0.33). Non-probability sampling yielded higher estimates than probability sampling (38%, 95% CI: 0.28–0.49 vs. 21%, 95% CI: 0.11–0.33). Estimates varied widely by tool, ranging from 7%, 95% CI: 0.04–0.10 (Kutcher Adolescent Depression Scale) to 58%, 95% CI: 0.53–0.62 (Mental Health Inventory). For PTSD, all subgroup variables were significant moderators, including region, sampling method, measurement tool, mode of administration, risk of bias, and age group (all p ≤ 0.04). Asia had the highest regional prevalence (35%, 95% CI: 0.29 − 0.40), followed by Africa (24%, 95% CI: 0.12–0.39) and South America (8%, 95% CI: 0.06 − 0.10). Non-probability sampling also yielded higher estimates (32%, 95% CI: 0.20–0.45 vs. 11%, 95% CI: 0.08–0.14). Self-administered tools reported higher prevalence than clinician-administered tools (25%, 95% CI: 0.15–0.38 vs. 13%, 95% CI: 0.10–0.18), as did studies with moderate-to-high vs. low risk of bias (40%, 95% CI: 0.27–0.53 vs. 13%, 95% CI: 0.10–0.18). Older youth had higher PTSD prevalence than younger youth (32%, 95% CI: 0.20 − 0.45 vs. 11%, 95% CI: 0.08–0.14). See Fig. 4. Forest plots showing subgroup analysis for Depression and PTSD Individual level correlates of mental disorders Individual correlates of mental disorders were explored and extracted from 21 studies. These were calculated by the study authors using multivariate regression models at a p value of < 0·05. Depression The risk factors identified for depression in both genders included food insecurity, poor relationships with parents, older age (> 15 years), female gender, high slum adversity score, and positive Covid 19 infection [28, 34, 59]. Stealing food, involvement in criminal activities, a family history of depression, being fired from work, dissatisfaction with income, financially supporting two or more dependents, problematic alcohol use, and childhood physical, psychological, and sexual abuse were also identified as risk factors for depression in males [25, 36, 42]. For females, physical, psychological, or sexual abuse, poor maternal connection, a history of adverse events in the family, perceiving the neighborhood as unsafe, being unmarried, being a homemaker, having a reproductive illness, being a victim of attempted rape or victim of ‘eve teasing’ (an Indian term for sexual harassment in a public place) all increased risk [29, 34, 40]. The protective factors for depression in males was a caring adult male in the home. A caring adult was defined as an adult who had high expectations of the young person such as following rules and being successful in their activities. These adults also listened to and helped youth as needed. Caring adult males and females in the home was protective for females [27]. Feeling connected to the neighborhood (defined as knowing most people and feeling part of the neighborhood) was protective for both genders [27]. Higher education and positive peer support (a friend who could be trusted and provided practical assistance, with unconditional acceptance) was protective for both genders [25, 28]. More positive feelings about work and less control of partners was also protective for males, whereas less stress at work was protective for females [27, 59]. PTSD The risk factors for PTSD across both sexes were older age, impaired social functioning, severe maternal physical violence, psychological violence by a significant other, witnessing one to two or more traumatic events, and being a victim of rape [24, 29],. Food insecurity and problematic alcohol use were also risk factors for males [36]. A protective factor for both sexes was feeling connected to their neighbourhood and completing secondary school education. Having a caring male or female adult in the home was protective only in males [27, 36]. Anxiety Disorders The risk factors for anxiety disorders in females included a history of rape, lower levels of education, lower social class, and food insecurity [26, 29, 37]. Protective factors were not reported in the included studies. Alcohol use disorders (AUDs) Risk factors for alcohol use disorders included older age, male sex, drug use in the past year, non-partner sexual violence in the past year, transactional sex with a partner, depressive symptoms, being from a joint family, no parental education, being sexually active, two or more sexual partners in the last 12 months, and daily social media use [31, 32, 35]. Protective factors were not reported in the included studies. Publication Bias Publication bias assessment was limited to depression, as it was the only disorder with a sufficient number of studies (n > 10) to allow for reliable testing. A funnel plot showed an asymmetric distribution, suggesting the presence of publication bias (Supplement 3). Egger’s test [60] did not reach statistical significance (β₁ = 7.77, 95% CI: −1.00 to 16.54, p = 0.08). However, given the borderline p-value and visual asymmetry of the funnel plot, publication bias cannot be fully excluded. Discussion This review found data on four disorders: depression, PTSD, anxiety, and substance use disorder. Depression with a pooled prevalence of 31% with was comparable to the 0–28% among youth in LMICs [61], and the 34% among youth globally. Notably, LMIC and global figures include youth living in very difficult circumstances, such as those living with HIV, in areas of prolonged conflict, and orphaned youth [62]. While the rates of depression in this review were comparable with LMIC and global estimates, depression often occurs comorbidly with PTSD, Anxiety, and AUDs [63, 64] which were all elevated in this review. However, data on comorbid disorders was not available for extraction from the included studies. PTSD with a pooled prevalence of 24% was approximately 1.5x higher than the 15.9% reported among trauma-exposed youth globally [65]. This pooled prevalence was also much higher than the 6.2–15.3% reported among trauma exposed youth in ten LMICs [66]. Older age was a risk factor for PTSD in this review, although a previous meta-analysis did not find age-related PTSD associations [64]. This association may reflect the increased vulnerability to mental health problems during puberty [2, 67], which when coupled with growing social responsibilities, academic, and economic pressures, may intensify stress responses increasing YLS vulnerability to PTSD [68]. Similar to previous studies, other risk factors for PTSD identified in this review were violence from significant others, and multiple traumatic exposures [61, 65]. These may increase PTSD risk by overwhelming YLS coping resources and weakening their sense of safety and trust [69, 70]. Anxiety with a pooled prevalence of 35% was higher than the 8–27% prevalence reported in other LMIC countries [61]. Social and economic factors common in slums such as chronic uncertainty, violence, and economic precarity [29, 37] are likely to increase anxiety risk among YLS beyond estimates reported in broader LMIC surveys, which typically combine urban and rural youth which may mask the concentrated adversity typically found in slum settlements. The pooled prevalence of AUD (23%) was also three times higher than the 7.7% reported in a previous sub-Saharan Africa meta-analysis [71]. Structural and economic factors specific to slum communities such as high youth unemployment, limited structured recreational activities, and the concentration of informal alcohol shops may create permissive environments for harmful use [72, 73]. Social norms around alcohol may also differ within slum communities, where drinking may function as a coping mechanism for chronic adversity and social stress in ways that sustain higher rates of its disordered use [74]. The most frequent risk factors in this review were food insecurity and violence exposure. Food insecurity is often used as a proxy for poverty [75], and an increasing body of research has explored the association between food insecurity and youth mental health among youth in LMICs [40, 76–78]. However, interventions addressing food insecurity have largely focused on nutritional and cognitive outcomes during early life [79]. Our findings suggest the need to consider mental health as an important outcome for interventions targeting food insecurity among youth in LMICs [80]. Strengths and limitations A key strength of this review was its systematic synthesis of data across diverse geographical, cultural, and socioeconomic contexts, highlighting global patterns and context-specific variations. To the best of our knowledge, this is the first review of its kind. However, it also has several limitations. Although no restrictions were applied to publication dates or language during the search, all studies that met the inclusion criteria were published in English. Thus, the findings from non-English-speaking slum settings are likely underrepresented. Representation across the globe was also not as diverse as expected with only two studies from South America and seven from South Africa. There were wide variations in the study assessment tools which were all developed in high-income countries, though validated and adapted for LMIC settings. All studies except one, exclusively utilised screening instruments which may under or overestimate the true burden of mental disorders in LMIC slum settings due to differences in cultural idioms of distress, and stigma [81, 82]. Screening instruments may however detect subclinical symptoms which may also cause substantial distress and functional impairment. Capturing these symptoms is therefore valuable as they may represent important targets for early intervention. Overall, considering these limitations and the substantial heterogeneity among studies, the review findings should be interpreted with caution. Implications for policy and research The high burden of mental disorders and AUDs among YLS makes a strong case for including mental health in urban development and slum upgrading policies. Social protection policies addressing poverty, food insecurity, and violence should recognise their mental health implications. National mental health policies should address the heightened vulnerability of YLS by ensuring their visibility in resource allocation and service planning. Mental health services for YLS should be community-based, culturally acceptable, and delivered by trusted actors. Interventions should focus on reducing risk exposure while building protective resources. Future research should prioritise longitudinal designs for better understanding of the associations between slum-related adversities and mental health outcomes. The use of standardised measurement approaches, as well as validated diagnostic tools, could help to improve comparability across studies. Engaging young people as active participants in research design and implementation is also a priority. Conclusion Young people living in slum settlements face a disproportionate burden of mental health disorders. There is an urgent need for culturally responsive, contextually grounded policies and interventions which address identified risk factors such as violence exposure, and food insecurity, while strengthening the protective factors that buffer against adversity. Translating this evidence into action will require coordinated investments across health, education, social protection, and urban planning sectors. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements We thank the Warwick Medical School Information Specialist (Dr. Samantha Johnson), for her support in developing the database search strategy. Author contributions T.B.A., D.G., U.R., S.J. and O.O. contributed to conceptualising and developing the review protocol. T.B.A. performed the data search. T.B.A. and O.F.B. performed the title, abstract, full-text screening, data extraction, and quality assessment. T.B.-A. wrote the first draft of the manuscript. D.G., U.R., and S.J. contributed substantially to interpreting the results. All authors read and approved the final manuscript. Funding This study/project is funded by the UK’s National Institute for Health and Care Research (NIHR) (Award number: NIHR200846) for the “Transforming Access to Care for Serious Mental Disorders in Sums: the TRANSFORM Project”. The funders of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the manuscript. Data availability Data is provided within the manuscript and supplementary information files at the University of Warwick institutional repository: http://go.warwick.ac.uk/wrap. Declarations Conflict of Interest The authors have no competing interests to declare relevant to the article’s contents. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement Data is provided within the manuscript and supplementary information files at the University of Warwick institutional repository: http://go.warwick.ac.uk/wrap.