The science of child and adolescent mental health in Brazil: a nationwide systematic review and compendium of evidence-based resources

Abstract

Background

Brazil is home to 50 million children and adolescents, whose mental health needs require context-sensitive research. Although scientific output is growing in the country, publications remain scattered and often inaccessible.

Methods

This systematic review compiles prevalence estimates, assessment instruments and interventions for child and adolescent mental health-related outcomes in Brazil (PROSPERO: CRD42023491393). We searched international (PubMed, Web of Science, PsycINFO, Google Scholar) and regional (Scielo, Lilacs, Brazilian Digital Library of Theses and Dissertations) databases up to July 2024. We consulted reference lists and experts. Extraction followed Consensus-based Standards for the selection of health Measurement Instruments and Cochrane manuals.

Results

We included 734 studies on 2576 prevalence estimates, 908 studies on 912 instruments and 192 studies on 173 intervention trials. The prevalence of any mental disorder ranged from 10.8% (12-year-olds; Pelotas, RS) to 19.9% (ages 7–14; Porto Alegre, RS and São Paulo, SP) although a nationally representative study is lacking. There is an alarming rise in self-harm notifications, reaching 133 in 2019 (per 100 000 aged 10–19). Indigenous youth face suicide rates of 11 (ages 10–14), far exceeding national numbers (0.652). Nationwide surveys reveal severe violence exposure (eg, 21% of adolescents suffer physical violence at home in the previous year), disproportionately impacting Black youth and increasing risk for mental disorders. There are reliable instruments for assessing psychopathology, yet most lack cross-cultural validation. Interventions remain underimplemented. The largest trials adapted substance-use prevention programmes from high-income countries, proving ineffective in Brazil. Public investment is the primary driver of research, which is centralised in wealthier states and misrepresents social minorities.

Discussion

This review provides timely access to appraised evidence-based resources, facilitating uptake into practice. Brazil’s historical sociocultural challenges impact youth mental health, with public health priorities including systemic violence, racism and indigenous suicide.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Scientific resources in Brazil are expanding but remain fragmented, hindering their translation into policy and practice.

WHAT THIS STUDY ADDS

• Mental health condition prevalence among Brazilian youth aligns with global estimates, but adolescents face high exposure to substance use, violence, increasing self-harm and racial inequalities.

• Intervention programmes and assessment instruments are largely based on international models with limited adaptation to Brazil’s sociocultural context.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This compendium provides ready access to scientific evidence that can facilitate uptake into real-world practice, highlighting urgent public health priorities.

• Public funding should direct incentives to develop culturally sensitive resources for mental health, including interventions focused on social minorities and incorporating anti-racist goals.

Introduction

Brazil, the largest country in Latin America, is home to 50 million children and adolescents, yet this population has been historically neglected by the country’s public policies.^{1 2} Despite early 21st-century growth, the country grapples with severe inequality, high poverty and violence rates, and economic downturns.^{3 4} Brazilian youth experience adverse environments, with estimations that approximately 20% of those aged 10–19 suffer from a mental disorder.⁵,⁷ The Psychosocial Care Network (‘Rede de Atenção Psicossocial’, or RAPS) addresses this need, operating as a community-based mental health system within Brazil’s Unified Health System (‘Sistema Único de Saúde’).⁸ Despite notable achievements, RAPS is strained by underfunding, coverage gaps and quality deficits. The educational system is also critical for mental health, with the 2024 National Policy for Psychosocial Care in School Communities established to address shortages in programmes, resources and specialised staff in public schools.⁹

Mental health practices are inconsistent across the country, with insufficient uptake of clinical guidelines nationally established for the public system.⁸ While oriented by international literature and suitably determining pharmacological approaches, these protocols are arguably insufficient to address Brazil’s diverse contexts. According to the national census,¹ people under 20 are distributed across five racial groups (49.47% brown, 41.35% white, 7.95% black, 0.97% Indigenous and 0.25% yellow) that experience distinct social realities: mortality rates for brown, black and Indigenous youth are approximately 1.25, 1.5 and 2.25 times higher than their white counterparts.¹⁰ Despite bearing the greatest burden, Indigenous populations remain historically underrepresented in national health data and research.¹¹ This reveals the need for locally established science and decolonial approaches addressing the unique mental health struggles of Brazilian children and adolescents, including members of Indigenous populations, quilombos (communities with a strong African-Brazilian heritage that were originally composed of settlements formed by self-liberated enslaved people), favela residents and victims of systemic racism.¹⁰¹²,¹⁶

Brazilian science has witnessed significant progress driven by public funding and technological development.¹⁷,¹⁹ Since the 2000s, the country has experienced a surge in scientific publications, growing at more than double the international rate.^{20 21} This has positioned Brazil as the 13th largest global scientific producer, with a notable emphasis on the health field.²²,²⁴ However, impact metrics have not kept pace; citation indexes remain below global and Latin American averages, and many works go uncited.^{22 25} The scattered nature of publications highlights weaknesses in scientific communication, lacking cross-referencing essential for advancing knowledge.^{26 27} In isolation, articles become costly to locate and appraise, creating a ubiquitous barrier to the uptake of science in practice.

This systematic review aims to identify and appraise prevalence surveys, assessment instruments and interventions on child and adolescent mental health-related outcomes in Brazil. Our goal is to create an accessible resource for consulting the evidence base, evaluate the current state of research in the field and take practical insights for policy.

Methods

We initially conducted an umbrella search, screening studies into three areas: prevalence estimates, assessment instruments and interventions. Then, we carried out a three-arm review with specific methods for each topic. Our approach was based on a similar country-wide systematic review in Greece.²⁸ We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement (online supplemental table 1) and the PRISMA-Consensus-based Standards for the Selection of Health Measurement Instruments (COSMIN) manual for reporting systematic reviews of measurement instruments.^{29 30} Protocol was registered with PROSPERO (CRD42023491393).

Search strategy

We employed a comprehensive strategy without restrictions of language from inception to 18 July 2024. Our query associated terms related to mental health, to children and adolescents and to Brazil, being adapted to English and/or Portuguese according to databases (see online supplemental table 2).

The multistep search procedure consisted of (1) international databases (PubMed, PsycINFO and Web of Science); (2) regionally representative databases (SciELO and Lilacs); (3) Google Scholar and (4) a national repository of theses and dissertations (BDTD—Biblioteca Digital Brasileira de Teses e Dissertações).

This was complemented by (5) identifying review articles during screening and inspecting their reference lists; (6) snowballing inclusions based on relevant studies; (7) consulting a panel of eight experts for additional references and (8) conducting a specific search on literature gaps identified during extraction (see online supplemental table 3). The specialist panel was composed of an international epidemiology researcher with 20 years of experience in global mental health surveys (Zeina Mneimeh) and seven psychiatry professors from Brazilian federal universities with extensive research in mental health epidemiology (Arthur Caye, Christian Kieling, Luis Augusto Rohde, Guilherme V. Polanczyk, Jair Mari, Rodrigo A. Bressan and Euripides Constantino Miguel).

References were uploaded to Rayyan.³¹ Studies retrieved in steps 1 and 2 were automatically deduplicated using EPPI-Reviewer V.4.0,³² with a second manually verified deduplication in Rayyan.

Screening

We conducted primary screening based on title/abstract, categorising studies into one or more areas: prevalence, instrument and intervention. The team conducted a pilot calibration of inclusion criteria based on at least 10% of the results and held regular discussion meetings. Pairs of independent researchers (ACC and CZ; ACT and CBC) performed abstract screening for studies retrieved from steps 1 and 2. Discrepancies were discussed and resolved by a third reviewer (LEM) when needed. Data sources described in steps 3 and 4 did not allow for exportation and were manually inspected by a single screener (ACC and CZ). In Google Scholar, results were screened until reaching 100 consecutive studies without novel inclusions. Due to its crawler-based nature, Google Scholar returns overwhelming results and is recommended as a supplementary datasource.³³

Full-text screening and data extraction were performed independently for each area, using a single-reviewer approach with at least 20% of the articles verified by another team member (LEM), following the protocol of the previous nationwide Greek review.²⁸ While Cochrane manual recommends independent screening to minimise the risk of missing relevant studies, it also recognises that single-reviewer screening yielded adequate performance in some field tests.³⁴,³⁶ With an emphasis on sensitivity, we applied independent screening at the abstract stage to avoid early exclusions,³⁷ then proceeding with single-reviewer approaches to enhance efficiency.

Inclusion criteria

We included studies reporting prevalence rates, assessment instruments or interventions related to child and adolescent mental health outcomes and risk factors (eg, includes mental disorders, substance use, suicidality, maltreatment and neglect, bullying, quality of life, wellness, cognitive neurodevelopment, learning difficulties, personality traits, mental health literacy and awareness and other associated constructs). The eligible population was children/adolescents up to 19 years old in Brazil (school-based samples with outliers above 19 years old were also eligible), including assessments/interventions involving caregivers/professionals. Research articles, editorial letters with original data, dissertation theses and book chapters were eligible.

For prevalence studies, we included surveys on community- and school-based samples using structured instruments or epidemiological registers to report prevalence rates or symptom levels.

Instrument studies were included if they developed, translated or validated assessment instruments (or solely applied the instrument if this was the unique report on that tool).

Intervention studies were included when reporting experimental designs (from pre–post uncontrolled studies to randomised clinical trials (RCTs)) or adaptation/translation of interventions from other settings.

Exclusion criteria

We excluded multicountry or general population studies that did not disaggregate results between countries or age groups. We excluded studies on neurodevelopmental disorders reporting only motor or phonetic development, on mental health outcomes of caregivers or professionals that do not directly concern children and adolescents (eg, maternal depression or teacher burnout). Conference abstracts were ineligible.

We excluded prevalence studies on clinical populations (eg, quality of life among leukaemia patients), on populations with very specific determinants (eg, survivors of traumatic events), reporting the absolute numbers of suicide or violence notifications without population rates, and datasets with duplicate information (we only included the most comprehensive and/or updated).

Instrument studies were excluded when reporting guides for qualitative interviews, projective tests or tools assessing quality of life in the context of clinical conditions.

Intervention studies were excluded when reporting case studies or series with fewer than five participants.

Data extraction and synthesis

Prevalence studies

Procedures were based on a widely cited systematic review on the prevalence of child and adolescent mental health disorders.³⁸ Risk of bias was assessed using the validated tool for prevalence studies developed by Hoy et al,³⁹ which evaluates external and internal validity, analysis bias and sample representativeness. Each estimate from each study was a separate entry. After extraction, a synthesis table aggregated all estimates per construct.

Instrument studies

The COSMIN guidelines oriented the extraction, quality assessment and psychometric evaluation.⁴⁰ We employed a three-level procedure encompassing extraction (recording data per each instrument at each study), evaluation summary of its psychometric properties, and synthesis (aggregating data for each instrument) (see details in online supplemental tables 4–6).

Intervention studies

Procedures were based on the Cochrane Manual for Systematic Reviews of Interventions.⁴¹ Methodological quality was assessed with the revised Cochrane tool for randomised trials (RoB 2) and the Joanna Briggs Institute tool for non-randomised designs.^{42 43}

Results

We included 734 studies reporting 2576 prevalence estimates, 904 studies reporting 912 unique assessment instruments, and 192 studies reporting 173 trials (see online supplemental figure 1 for flow chart; see online supplemental table 7 for reasons for exclusions). The full dataset can be navigated in online supplemental file 1.

Overview of scientific production

Figure 1 presents a time trend analysis of publications and topics of inquiry (see heatmap graph in online supplemental figure 2). Since 2010, Brazilian scientific production on youth mental health has grown significantly, but intervention studies represent only a tenth of all articles. Studies concentrate on topics such as substance use, psychosocial symptoms, exposure to adversities and depression. There is significantly less focus on areas specifically relevant to children and adolescents, including autism spectrum disorder (ASD), intellectual disability and learning disorders. A particularly overlooked domain is sexuality and gender. There are only two works on instruments assessing sexual trajectories discrimination,^{44 45} and no reports on gender dysphoria.

Figure 1. Time trend analysis and areas of concentration of scientific production on child and adolescent mental health in Brazil. Note: A single study may be included in more than one area. For example, intervention studies could provide information on the responsiveness of an instrument, while prevalence studies might offer psychometric validation of an instrument. ADHD, attention deficit/hyperactivity disorder; ASD, autism spectrum disorder; OCD, obsessive–compulsive disorder.

Figure 2 shows the regional distribution of prevalence surveys (see heatmap graph in online supplemental figure 3). There is a concentration of studies in the historically wealthier and whiter southern and southeastern states, particularly São Paulo (SP) and Rio Grande do Sul (RS). This reflects a centralisation of scientific production in specific universities across these regions, with studies usually representing urban populations near these academic centres.

Figure 2. Regional analysis of represented population and conditions according to the number of studies. Note: A single study may be included in more than one condition and region. ADHD, attention deficit/hyperactivity disorder.

Prevalence estimates

Appraisal of the literature

Regarding quality parameters, 517 out of all the 734 studies (70.44%) used probabilistic samples. Most studies on the prevalence of mental disorders relied on screening questionnaires using cut-off points for establishing diagnosis or at-risk population,⁴⁶ often based on international normative references that lack validation for the Brazilian context.⁴⁷ Only 27 studies (3.68%) employed structured clinical interviews based on standardised diagnostic criteria to confirm diagnoses. Only 69 studies (9.40%) included more than 5000 participants and 75 studies (10.22%) recruited participants from all five national regions. Many studies reported small samples from specific locations, with 530 studies (72.21%) recruiting participants from a single state.

There is no nationwide study with clinical interviews on mental disorders prevalence in Brazil, and best evidence is derived from high-quality studies in specific regions. There are nationwide representative surveys on several mental health-related constructs (eg, substance use, bullying), primarily conducted by governmental entities. Official registers available from the public health system register system (DATASUS) provide nationally representative data on self-injury notifications, hospitalisations due to psychoactive substances and rates of violent death.

Findings

Table 1 selects best-available evidence on the prevalence of mental disorders (see online supplemental file 1 for full dataset). Current prevalence of any psychiatric diagnosis ranged from 10.8% (at 12 years old in Pelotas, RS) to 19.9% (age 7–14, in Porto Alegre, RS and São Paulo, SP).^{48 49}

Table 1. Best data on the prevalence of mental disorders: selection of 10 studies (N ≥1000, with structured clinical interview and diagnostic manual criteria).

Study	Instrument / criteria	Description	City (state) region	Year of collection	Sample size	Age groups Girls percentage Ethnicity	Point prevalence (highest to lowest)
Bauer et al. (2022)83	DAWBA DSM-V	Pelotas birth cohort106	Pelotas (RS) South	2010 and 2015	4229	6 and 11 years (two time points) 48.1% Black or mixed: 38.3% White: 61.7%	(At 6 years old) Any psychiatric disorder: 16.3% Anxiety disorders: 8.9% ADHD 2.6% CD/ODD: 2.6% Mood disorders: 1.3%	(At 11 years old) Any psychiatric disorder: 13% Anxiety disorders: 4.3% ADHD: 4% CD/ODD: 3.6% Mood disorders: 3.2%
Álvares et al. (2021)107	MINI DSM-IV	São Luís birth cohort108	São Luís (MA) Northeast	2016	2486	18–19 years 52.4%	PTSD: 2.5%
Anselmi et al. (2010)48	DAWBA DSM-IV	Pelotas birth cohort106	Pelotas (RS) South	2004–2006	4448	12 years 50.3% Black or mixed: 28.5% Asian or native: 4.7% White: 66.8%	Any psychiatric disorder: 10.8% ADHD: 4.1% Anxiety disorder: 2.2% CD: 2.2%	Major depressive disorder: 1.6% Eating disorder: 0.1% ODD: 2.1%
Simões et al. (2020)108	MINI DSM-IV	São Luís birth cohort	São Luís (MA) Northeast	2016	2515	18–19 years 52.4% Black: 16.9% Mixed: 62.4% White: 20.7%	MDD: 11.8% Bipolar disorder: 11.8%	GAD: 3.5%
Gallo et al. (2017)109	MINI DSM-IV	Pelotas birth cohort106	Pelotas (RS) South	2011	3715	18 years 52.6% Non-white: 23.2% White: 76.8%	MDD: 6.8%
Paula et al. (2015)84	K-SADS-PL DSM-IV	Cross-sectional	Caeté (RJ), Goiânia (GO), Itaitinga (CE), Rio Preto da Eva (AM) Southeast, Centre, Northeast and North	2010–2011	1676	6–16 years 51.4%	Any psychiatric disorder: 13.1% Any anxiety disorder: 7.2% Any ADHD: 4.5% Specific phobia: 3.8% Any CD/ODD: 2.3% Social phobia: 2.0% ODD: 1.7% Separation anxiety disorder: 1.3%	Agoraphobia: 0.7% OCD: 0.6% CD: 0.6% Any depressive disorder: 0.5% GAD: 0.4% PTSD: 0.3% MDD: 0.2% Dysthymia: 0.1%
Vivian et al. (2014)47	K-SADS-PL DSM-IV	Cross-sectional	Porto Alegre (RS) South	2009 to 2011	2323	14–17 years 46.8%	OCD: 3.22%
Salum et al. (2021)110	K-SADS-PL DSM-IV	Cross-sectional	RS South	2008–2009	2457	10–17 years	Any anxiety disorder: 5.62% GAD: 3.87% Social Anxiety Disorder: 2.32%	Separation Anxiety disorder: 2%
Fleitlich-Bilyk and Goodman (2004)85	DAWBA DSM-IV	Cross-sectional	Taubaté (SP) Southeast	2000–2001	1251	7–14 years 47%	Any psychiatric diagnosis: 12.7% Any CD/ODD: 7.0% Any anxiety disorder: 5.2% ODD: 3.2% CD: 2.2% Anxiety NOS: 2.1% ADHD: 1.8%	Disruptive disorder NOS: 1.5% Separation anxiety disorder: 1.4% Any depressive disorder: 1% Specific phobia: 1.0% Social phobia: 0.7% PTSD: 0.1% OCD: 0.1% GAD: 0.4%
Salum et al (2015)49	DAWBA DSM-IV	Brazilian High Risk Cohort Study49	RS and São Paulo (SP) South and Southeast	2010–2011	958*	6–14 years 45.2% Non-white: 39.5% White: 60.5%*	Any disorder: 19.9% Any OCD: 5.5% ADHD (combined): 2.9% Panic disorder: 2.9% Any CD/ODD: 2.1% MDD: 1.8% PTSD: 1.6% GAD: 0.8% Social phobia: 0.3%	Bipolar: 0.3% Any PDD/autism: 0.3% Anxiety (others): 0.2% Agoraphobia: 0.2% Eating disorder: 0.1% Tic disorders: 3% Separation Anxiety disorder: 0.00% Psychosis: 0%

Notes: The full dataset with 734 studies on 2576 prevalence estimates can be consulted in online supplemental file 1. Cruzeiro et al¹¹¹ was not included here because, in spite of employing DSM-IV criteria evaluated with a structured interview, it sets two positive answers as the cut-off point for conduct disorder diagnosis, instead of three as established in the manual.

^*The Brazilian High Risk Cohort study is composed of 2512 participants, but we only extracted data for the random, community-based sample (n=958), excluding the high-risk selected sample.

ADHD, attention deficit hyperactivity disorder; CD, conduct disorder; DAWBA, Development and Well-Being Assessment; DSM, Diagnostic and Statistical Manual of Mental Disorders; GAD, Generalised Anxiety Disorder; K-SADS-PL, Schedule for Affective Disorders and Schizophrenia for School Aged Children: Present and Lifetime Version; MDD, major depressive disorder; MINI, Mini International Neuropsychiatric Interview; NOS, not otherwise specified; OCD, obsessive–compulsive disorder; ODD, oppositional-defiant disorder; PTSD, post-traumatic stress disorder.

Table 2 selects nationally representative estimates of mental health-related constructs. Suicide rates per 100 000 indigenous adolescents are alarmingly high (11, age group 10–14, years 2010 to 2015), almost 17 times the equivalent national rate (0.652).⁵⁰ Other studies echo these findings: municipalities with a high density of self-declared indigenous population presented similar rates (11.6 for children aged 5–14),⁵¹ and reports at an indigenous reservation in Dourados (MS) found even higher rates (174 for adolescents aged 14–19).⁵²

Table 2. Nationally representative estimates of mental health-related constructs: selection of 37 studies (N ≥10 000, data collected since 2010, nationwide).

Studies	Data sources	Funding	Years	Sample description	Assessment	Prevalence/rate per 100 000 people (highest to lowest)
5067 112 115	National Information System of Mortality (SIM) National Institute of Geography and Statistics (IBGE)	Public	2010–2020	National system notifications 1–19 years old	Death registers Population census	Suicide mortality rates Indigenous population (10–14 years): 11 15–19 years: 4.3–5.3 10–19 years: 2.43–2.65 1–19 years: 1.67 Under 15 years: 0.59–0.86 10–14 years: 0.6	Homicide-mortality rates 1–19 years: 10.87 Firearm-related death rates 1–19 years: 10.04 Mothor vehicle accident death rates 1–19 years: 5.2
53 66 88	Notifiable Diseases Information System (Sinan) National Institute of Geography and Statistics (IBGE)	Public	2009–2021	National system notifications 10–19 years old	Health notifications Population census	Self-inflicted injury prevalence 2019 – 2021: 27.39% Self-inflicted injury rates 2019: 133.1 2018: 84.2 2017: 59.1 2016: 32.8 2015: 15.3 2010: 4.7	Self-inflicted injury rates (occurred at school) 2018: 2.75 2017: 1.77 2016: 0.51 2015: 0.29 2011: 0.09
116	Hospital Information System of the Brazilian National Health System (SIH/SUS) National Institute of Geography and Statistics (IBGE)	Public	2017–2022	National system notifications 10–19 years old	Hospitalisation notifications (ICD-10: F10 or F19) Population census	Hospitalisation rates for mental and behavioural disorders due to alcohol or psychoactive substances 2022: 13.72 2021: 13.18 2020: 13.10	Hospitalisation rates for mental and behavioural disorders due to alcohol or psychoactive substances 2019: 17.77 2018: 17.53 2017: 16.18
54 5658	2019 National Adolescent School-based Health Survey (PeNSE)117	Public	2019	159 245 adolescents* 13–17 years old	Questions adapted from the Global School Based Student Health Survey (GSHS)	Violence Being bullied (past month): 40% Being bullied twice or more (past month): 23% Physical assault by parents/guardians (past year): 21% Sexual abuse (lifetime): 14.6% Rape (lifetime): 6.3% Physical assault not by parents/guardian (past year): 12.2% Missed class due to insecurity in home-school route (past month): 11.6% Missed class due to insecurity at school (past month): 10.8% Involvement in a physical fight (past month): 10.6% Involvement in a fight in which someone used a melee weapon (past month): 4.8% Involvement in a fight in which someone used a firearm (past month): 2.9%	Substance use Alcohol use (lifetime): 65.15% Alcohol use–being drunk (lifetime): 47% Alcohol use (past month): 27.7% Tobacco use (lifetime): 24.01% E-cigarette use (lifetime): 17.14% Illicit drug use (lifetime): 14.3% Illicit drug use (past month): 4.4%–5.2% Tobacco use (past month): 6.5% well-being Body image dissatisfaction (present): 23.3%
5762 65 118 122	2015 National Adolescent School-based Health Survey (PeNSE)123	Public	2015	118 909 adolescents 13–17 years old†	Questions adapted from the Global School Based Student Health Survey (GSHS)	Violence Being bullied rarely or sometimes (past month): 37.7% Practised bullying (past month): 19.8%–22% (ages 13–15) Physical assault by parent/guardian (past month): 14.5–16.2% Missed class due to insecurity in home-school route (past month): 12.8% Missed class due to insecurity at school (past month): 9.3% Involvement in a fight in which someone used a melee weapon (past month): 8.2% Being bullied most of the time to always (past month): 6.6%	Violence Involvement in a fight in which someone used a firearm (past month): 5.6% Sexual abuse (lifetime): 4% Substance use Alcohol use (past month): 29.3% Tobacco use (lifetime): 18.4% Tobacco use (past month): 6.1% to 8% E-cigarette use (lifetime): 3.2% well-being Laxative misuse/self-induced vomiting (lifetime): 6.9%
124 128	2012 National Adolescent School-based Health Survey (PeNSE)129	Public	2012	109 104 adolescents‡ Age groups 13–15 years: 86% ≥16 years: 13.2% <13 years: 0.8%	Questions adapted from the Global School Based Student Health Survey (GSHS)	Violence Practised bullying (past month): 20.8% Physical assault by parent/guardian (past month): 11.6% Missed class due to insecurity in home-school route (past month): 9.1% Involvement in a fight in which someone used a melee weapon (past month): 8.3% Missed class due to insecurity at school (past month): 8.0%	Violence Involvement in a firearm fight (past month): 6.9% Being bullied most of the times to always (past month): 7.2% Substance use Illicit drug use (lifetime): 7.3% Tobacco use (past month): 5%
130 134	2010 CEBRID study (Brazilian Center on Drug Information)135	Public	2010	50 890 adolescents§ 10–18 years old	Questions adapted from the Global School Based Student Health Survey (GSHS)	Substance use Alcohol use (lifetime): 59.6% Binge drinking (past year): 31.7% (ages 13–18) Tobacco use (lifetime): 25.2%	Substance use Tobacco use (past year): 15.3% Tobacco use (past month): 8.6% Illicit drug use¶ (lifetime): 5.2% Tranquillisers use (lifetime): 3.9% Crack use (lifetime): 0.6%
136	Study of Cardiovascular Risks in Adolescents (ERICA)137	Public	2014	85 000 adolescents** 12–17 years old	Structured questionnaire	Substance use Alcohol use (past month): 21.2%

Note: The full dataset with 734 studies on 2576 prevalence estimates can be consulted in online supplemental file 1.

^*Students from 6th primary to 3rd secondary grade in 4361 public and private schools, urban and rural, including all state capitals.

^†Students from 6th primary to 3rd secondary grade in 3540 public and private schools, urban and rural, including all state capitals.

^‡Ninth primary grade students from 2482 public and private schools across all state capitals and other municipalities.

^§Students from 789 public and private primary and secondary schools across all state capitals.

^¶Illicit drugs groups cannabis, cocaine/crack and ecstasy.

^**Students from 1251 public and private schools in 273 municipalities with more than 100 000 inhabitants, including all state capitals.

ICD-10, International Classification of Diseases–10th edition.

Self-harm notification rates among adolescents have increased markedly, from 15.3 in 2015 to 133.1 in 2019.⁵³ Substance use is also high, with national surveys in 2019 assessing lifetime alcohol use at 65.15% and illicit drug use at 14.3% for adolescents aged 13–17.^{54 55} E-cigarette consumption is an emerging concern, with lifetime use prevalence rising from 3.2% in 2015 to 17.1% by 2019.^{56 57}

Exposure to violence remains alarmingly high and associated with adverse mental health outcomes. Nationwide estimates from 2019 indicate substantial rates among adolescents aged 13–17, including bullying (40% in the past month), physical domestic violence (21% in the past year), sexual abuse (14.6% lifetime prevalence), missing school due to street insecurity (11.6% in the previous month) and involvement in firearm fights (2.9% in the previous month).⁵⁴⁵⁸,⁶²

A racial bias is explicit in the association of violence and mental health. A study of 1686 adolescents in Rio de Janeiro (RJ) found that those in high-crime areas were at elevated risk for mental disorders, with black adolescents experiencing a stronger association.⁶³ Another study of 973 adolescents in Salvador (BA) showed that racial discrimination, rather than race itself, was linked to higher rates of depression.⁶⁴ Racial-based bullying was associated with increased alcohol and tobacco use among racial minorities,⁶⁵ and self-harm rates were higher among black and brown adolescents.⁶⁶ Firearm-related death rates reached 10.04 in 2020 (ages 10–14), predominantly composed of homicides of black and brown male youth.⁶⁷ Notwithstanding, research focused on the black population remains limited, and there are only two studies assessing quilombola mental health.^{68 69}

Assessment instruments

Appraisal of the literature

Replication issues were evident: 282 (30.92%) out of 912 unique instruments were reported in only one study, and only 20 (2.19%) were featured in more than five studies. Many studies employed different versions of the same instrument, only validated subscales, or created new instruments instead of existing validated ones.

Psychometric assessment represents a significant quality gap. Positive evidence of internal consistency was reported for 277 instruments (30.37%), construct validity for 112 (12.28%), structural validity for 80 (8.77%), criterion validity for 39 (4.28%) and cross-cultural validity for only 20 tools (2.19%).

Instruments available in Brazil are mostly translations of international tools: 745 instruments (81.69%) were translated from English or other foreign languages. Only 120 of these had at least one validated procedure for translation and cross-adaptation, while the others used free translation or unidentified previous translations. Often, instruments had previous Portuguese versions for adults and were directly validated for children/adolescents without prior adaptation.⁷⁰

There were 167 instruments (18.31%) originally developed in Brazilian Portuguese, and development quality and content validity were considered adequate for only 5 and 4 of those. With few exceptions,^{71 72} development of locally constructed instruments missed essential steps, such as involving patients and professionals in assessing item relevance/comprehensibility.

Findings of validation

Table 3 presents a selection of 14 instruments that were positively validated across at least five psychometric properties (see online supplemental file 1 for all 991 tools). There are valid, reliable tools for attention deficit hyperactivity disorder, anxiety, ASD or psychosocial functioning, while there are scarcer options for substance use and suicidality. Only two tools were cross-culturally validated to a native Brazilian language (Karajá), the Children Behaviour Checklist (CBCL) and the Teacher Report Form.⁷³ The Everyday Discrimination Scale for Adolescents and Adults is the sole tool addressing racism.⁴⁵

Table 3. Assessment instruments for mental health: selection of 14 tools (positive validation reports on at least five psychometric properties).

Instrument	Population	Setting	Rater	Items	Translation	Psychometrics*									References
Diagnostic interviews for mental disorders					Translation	CC	CR	CV	IC	IR	ME	R	SV	TR	References
Development and Well–Being Assessment	Children adolescents	Clinical community	Caregiver clinician		Valid 85	NA	+	NA	+	+	NA	+	+	NA	4985 138 145
Autism Diagnostic Interview–Revised	Children	Clinical	Caregiver		Valid 146	NA	+	+	+	+	NA	+/–	NA	NA	146 149
Screening instruments for mental disorders					Translation	CC	CR	CV	IC	IR	ME	R	SV	TR	References
Swanson, Nolan and Pelham Scale	Children adolescents	Clinical community	Caregiver teacher	18	Valid 150	+/–	+	+	+	+/–	NA	+/–	+/–	NA	151 172
Child Behaviour Checklist	Children adolescents	Clinical community	Caregiver	118	Valid 173 174	NA	+/–	+/–	+/–	NA	?	+	+/–	+/–	73140 154 161 174 211
Screen for Child Anxiety Related Emotional Disorders	Children adolescents	Clinical community	Self-report	41	Valid 212	NA	+	+	+/–	–	NA	+	+/–	+/–	110151 207 212 219
Mood and Feelings Questionnaire	Adolescents	School	Caregiver	33	Valid 220	NA	NA	+	+	+	NA	NA	+	+	221 224
Scales on well-being					Translation	CC	CR	CV	IC	IR	ME	R	SV	TR	References
Paediatric Quality of Life Inventory	Children adolescents	Clinical community	Caregiver self-report	23	Valid 225	NA	NA	+	+/–	+/–	NA	+	+	NA	225 230
Brief Multidimensional Students’ Life Satisfaction Scale	Children adolescents	School	Self-report	6	Uncertain	+	NA	+	+/–	+/–	NA	NA	+	NA	231 237
Rosenberg Self-Esteem Scale	Adolescents	Clinical community	Self-report	10	Valid 238	NA	NA	+	+	NA	NA	+	+	+/–	5221 238 246
Body Esteem Scale for Adolescents and Adults	Adolescents	Community	Self-report	23	Valid 247	NA	NA	+	+	NA	NA	+	+	+	247 248
Children’s Sleep Habit Questionnaire	Children	Clinical	Caregiver	33	Valid 249	NA	NA	+/–	+/–	NA	NA	+	+	+	190249 251
Neurodevelopmental assessment					Translation	CC	CR	CV	IC	IR	ME	R	SV	TR	References
Bayley Scales for Infant Development	Infant	Clinical	Health professional		Valid 252	NA	+	+	+/–	NA	NA	+	+	+	252
Social Skills Rating System	Children	School	Self-report teacher	30	Valid 253	NA	+	+	+/–	NA	NA	NA	+/–	+/–	242254 260
Childhood Executive Functioning Inventory	Children adolescents	School	Health professional	26	Valid 155	NA	+	+/–	+	+	NA	+/–	?	NA	155 157 170 261

Note: The full dataset with 991 tools can be consulted in online supplemental file 1.

^*‘+’ stands for a property with positive evidence for validation, ‘+/–’ for mixed evidence, ‘–’ for negative evidence, ‘?’ for available evidence that cannot be classified as positive or negative, ‘NA’ for non-available evidence (see online supplemental table 2).

CC, cross-cultural validity; CR, criterion validity; CV, construct validity; IC, internal consistency; IR, interrater reliability; ME, measurement error; PTSD, post-traumatic stress disorder; R, responsiveness; SV, structural validity; TR, test–retest reliability.

Interventions

Appraisal of literature

Psychosocial interventions were reported in 139 studies (72.40%), followed by 30 pharmacological or neuromodulation studies (15.63%) and 22 lifestyle intervention studies (11.46%). School-based programmes were reported in 78 studies (40.63%).

RCTs accounted for 90 studies (46.88%), followed by 65 quasi-experimental (33.85%) and 36 uncontrolled pre–post design studies (18.75%). A high risk of bias was observed in 164 studies (90.48%), including 70 RCTs, most of which raised significant concerns about procedures for blinded outcome assessment (detection bias) and/or blinding of subjects and researchers (performance bias). Large trials were notably lacking, with 139 studies (72.40%) involving fewer than 100 participants.

Interventions generally lacked input from individuals with lived experience, except for the single study focused on quilombolas, which employed a participatory design to develop an alcohol use intervention for adolescents.⁷⁴

Funding sources were predominantly public, reported in 171 studies (89.06%). Considering only RCTs with at least 100 participants, public funding was present in 34 out of 35 studies.

Findings

Online supplemental table 8 presents a selection of interventions reported in 35 RCTs with more than 100 participants (see online supplemental file 1 for the full dataset of 192 studies). No large trial has demonstrated an effective universal school or community intervention programme ready to scale in Brazil. Some interventions with null effects lacked sufficient sample power to detect outcomes in universal prevention (eg, PERAE alcohol use prevention programme with 348 participants).⁷⁵

The largest school-based programmes focused on substance use prevention, including #TamoJunto, #TamoJunto 2.0, and PROERD (Drug and Violence Resistance Educational Programme), which were modelled after the European Unplugged and the USA DARE programmes.^{76 77} These initiatives were government-endorsed, with nationwide implementation of PROERD, yet analysis showed partial to worsening effects on drug use.^{78 79}

There was a single study focusing on quilombola groups,⁷⁴ and no interventions specifically addressed the impact of racism on mental health or targeted indigenous mental health.

Discussion

This systematic review provides a compendium of evidence-based resources on child and adolescent mental health in Brazil. A dataset is made freely accessible to navigate 2576 prevalence estimates, 912 assessment instruments, and 173 intervention trials, facilitating the uptake of resources to practice and offering a wealth of data for scholar research and policymaking. The best data on the prevalence of mental disorders in Brazil comes from regional cohort studies, which estimate that 10.8% to 19.9% of children and adolescents meet clinical criteria for at least one diagnosis.^{48 49} At the national level, health records reveal suicide rates about 17× times higher among Indigenous youth,⁵⁰ as well as a 28-fold rise in self-harm notifications among all Brazilian adolescents over the past decade.⁵³ School-based surveys raise concerns about substance use (eg, lifetime alcohol use: 65.2%; e-cigarettes: 17.1%; illicit drugs: 14.3%) and exposure to severe violence (eg, physical assault by a guardian in the past year: 21%; lifetime sexual abuse: 14.6%).⁵⁴,⁵⁶⁵⁸ Some internationally used instruments such as the Development and Well-Being Assessment, Swanson, Nolan and Pelham Scale, and CBCL have been systematically validated, yet most tools lack replication to ensure reliable assessment locally. To date, no community-based intervention has demonstrated effectiveness for large-scale implementation in Brazil. Here, we debate these findings and analyse the current state of national science in this field.

Brazil lacks a nationwide prevalence study on child and adolescent mental disorders as have been conducted in the USA and UK.^{80 81} Best available data estimate the prevalence of mental diagnoses comparable to global figures, usually around 13% for age groups from 7 to 11,⁴⁸⁸²,⁸⁵ yet literature overly represents samples from south and southeast states. Alcohol consumption among Brazilian adolescents is a cause for concern, placing the country in the upper range of multicountry rankings.^{86 87} The sharp rise in self-harm notifications between 2010 and 2019 is partly attributed to the 2014 policy mandating the reporting of self-harm cases within the Brazilian health system, yet underreporting likely remains.⁸⁸ While global estimates suggest an upward trend during recent time series, direct comparisons are limited due to differences in assessing self-inflicted injury.⁸⁹

Brazil’s health surveillance systems stand out in several areas of vital statistics.⁹⁰ This yields high-quality assessment on suicide mortality, which shows that rates among Brazilian children and adolescents are broadly in line with global estimates,⁹¹ yet disproportionately higher among Indigenous youth.⁵⁰ A similar pattern has been reported for Indigenous peoples globally, including Australian Aboriginal, Māori and Native American groups.^{91 92} Nevertheless, a substantial paucity of suicide data persists for Indigenous populations. This gap stems from fragile vital statistics systems in the Global South (where the majority of Indigenous people live), and from a lack of ethnic identifiers in national databases of high-income nations such as Canada.⁹² Consequently, these findings from the Brazilian surveillance system provide a crucial contribution to the epidemiology of suicide among Indigenous groups, particularly youth.⁹² Despite these insights, our review highlights that research focused on the mental health of Indigenous youth in Brazil remains particularly scarce, reflecting a history of exclusion on health data and statistics.¹¹

Brazil’s burden of sociocultural violence profoundly impacts mental health. Children and adolescents face high rates of severe street, school, domestic and sexual violence, increasing risks of unfavourable mental health outcomes. Systemic racism underlies many of these associations: racial discrimination, poorer living conditions and higher exposure to violence place Black youth at elevated risk for depression, self-harm and substance use. Notably, Brazil has a world-record rate of firearm-related adolescent deaths (10.04 per 100 000), predominantly involving homicides of young Black males.⁶⁷ These findings add crucial evidence to the emerging debate on the mental health of Black children and adolescents in Brazil,⁹³ supporting calls for intervention programmes that explicitly incorporate anti-racist goals and culturally responsive tools tailored to diverse contexts such as quilombola communities.^{64 65 69}

Public funding is the primary driver of scientific production on child and adolescent mental health. Despite growth, concerns about quality and scope persist. Many widely used international tools lack consistent psychometric and cross-cultural validation, reflecting common challenges in non-English-speaking countries.²⁸ There are no large-scale trials for anxiety, depression, stress or bullying as reported across international literature.⁹⁴,⁹⁶ The largest school-based interventions in Brazil have adapted substance-use prevention programmes from European and North American interventions, lacking efficacy. Notably, #TamoJunto 2.0 improved cross-cultural adaptation to mitigate harmful effects on alcohol initiation outcomes reported in #TamoJunto 1.0, suggesting the complexity and risk of cultural adaptation of interventions developed in high-income countries.^{78 97}

Improving research quality also requires addressing major replication issues. This review identified a vast number of mental health publications scattered across a fragmented and uncoordinated body of literature. There were hundreds of duplicated efforts and isolated findings, many only available as theses and dissertations. By appraising and organising this research body, the compendium offers a unified catalogue intended to foster a common language and hopefully optimise resources within the scholarly community.

Strengths and limitations

This broad-scope review adopts a comprehensive search strategy, encompassing a range of databases, snowballing inclusions, expert consultations and framed search for gaps, without restrictions of time or language. It adheres to evidence synthesis guidelines across three domains, rigorously appraising studies according to established manuals such as Cochrane and COSMIN.^{40 41} Our database synthesises a substantial amount of data in an accessible manner for consultation, potentially representing the largest compilation effort in this field.

We also encountered limitations. To manage the large-scale scope of the review, full-text screening and data extraction were conducted by a single reviewer, with independent accuracy checks limited to a random 20% sample of the dataset. Inclusions from grey literature were not peer-reviewed, compromising the quality of evidence from this datasource. The broadness of scope implies that many studies might not be located. As highlighted in the COSMIN manual,⁴⁰ capturing all data sources for instruments is challenging, and many translations and validations may exist in sources beyond the reach of our search strategy (eg, books, developer manuals and websites). Works with relevant information were not included due to age-related inclusion criteria, such as the gender dysphoria prevalence studies that do not discriminate information for participants under 19.⁹⁸,¹⁰¹ Most prevalence and intervention studies also reported information on instruments, potentially overrepresenting the latter’s numbers in the final aggregation. Due to substantial heterogeneity in scope, populations, regions and methodologies, a meta-analysis was not conducted. We only assessed quantitative works. It is important to note that qualitative social and psychosocial research is robust in Brazil and provides culturally sensitive insights to illuminate quantitative exploration.¹⁰²,¹⁰⁵

Conclusions

Brazilian research on child and adolescent mental health misrepresents the country’s specific struggles, as it is largely based on international perspectives that are insufficient to address local context.¹⁰ Brazilian-specific challenges impact youth’s mental health, particularly violence, racism and indigenous seclusion. As the primary funder of national science, public funding should direct incentives to develop culturally sensitive assessment and scalable interventions focusing on social minorities.

The funder had no role in the methodology, execution, analyses, or interpretation of the data.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.