The social determinants of mental health and disorder: evidence, prevention and recommendations

This is a narrative review and synthesis of high-quality evidence (not a single experiment). The authors systematically mapped the causal links between social determinants and mental health outcomes, then evaluated preventive strategies. The "interventions" reviewed fall into three categories:

**Universal prevention:** Policies and programmes targeting entire populations (e.g., anti-poverty programmes, anti-discrimination laws, housing support)

**Selected prevention:** Interventions for groups at elevated risk (e.g., refugee resettlement support, LGBTQ+ affirming school policies)

**Indicated prevention:** Interventions for individuals showing early symptoms (e.g., cognitive-behavioural therapy for people exposed to trauma)

The "outcomes" were prevalence and incidence of common mental disorders (depression, anxiety, psychosis, PTSD, substance use disorders) measured through diagnostic interviews, symptom scales, and administrative health data.

This is a review of hundreds of studies, so there is no single sample. The evidence draws primarily from populations in the Global North (Europe, North America, Australia, New Zealand) and focuses on marginalised groups:

Refugees, asylum seekers, and displaced persons (estimated 100+ million globally)

Ethnoracial minoritised groups (Black, Indigenous, and people of colour in Western contexts)

LGBTQ+ populations (lesbian, gay, bisexual, transgender, queer)

People living in poverty (variously defined as below 60% of median income or receiving welfare benefits)

Children and adolescents exposed to adverse childhood experiences (ACE scores of 4+)

Adults exposed to workplace discrimination, housing instability, or neighbourhood deprivation

Sample sizes in individual studies ranged from several hundred to over 1 million participants (e.g., Danish registry studies).

The authors did not collect new data. They synthesised findings from studies using:

**Diagnostic instruments:** Composite International Diagnostic Interview (CIDI), Structured Clinical Interview for DSM-5 (SCID-5), Mini International Neuropsychiatric Interview (MINI)

**Symptom scales:** Patient Health Questionnaire-9 (PHQ-9, 0–27, higher = more depression), Generalised Anxiety Disorder-7 (GAD-7, 0–21), PTSD Checklist for DSM-5 (PCL-5, 0–80), Kessler Psychological Distress Scale (K10, 10–50)

**Administrative data:** Hospital discharge records, prescription registries, mortality data, welfare records

**Life course measures:** Adverse Childhood Experiences (ACE) questionnaire (0–10), neighbourhood deprivation indices (e.g., Index of Multiple Deprivation), perceived discrimination scales (e.g., Everyday Discrimination Scale, 0–40)

**Study design:** This is a narrative review with elements of a systematic approach. The authors state they "rely as far as possible on high-quality evidence" but do not conduct a formal meta-analysis or systematic review with pre-registered search strategy, risk of bias assessment, or quantitative synthesis. They selectively cite studies they consider methodologically strong (longitudinal cohorts, quasi-experimental designs, natural experiments, and some RCTs).

**What the design can prove:**

The review can map the breadth of evidence linking social determinants to mental health

It can identify consistent patterns across multiple study designs and populations

It can highlight candidate causal mechanisms (e.g., chronic stress → HPA axis dysregulation → depression)

It can synthesise evidence on what types of preventive interventions have been tested

**What the design cannot prove:**

It cannot provide precise effect sizes for any single intervention

It cannot establish causality definitively (though it cites studies that do)

It cannot rule out publication bias or selective citation by the authors

It cannot quantify the relative importance of different social determinants

It cannot provide individual-level guidance on what a single person should do

**Major methodological weaknesses:**

No systematic search strategy or PRISMA diagram

No explicit inclusion/exclusion criteria

No risk of bias assessment for individual studies

No quantitative meta-analysis

Potential for confirmation bias in selecting which studies to highlight

Over-reliance on observational data (many cited studies are cohort or cross-sectional)

Limited evidence from the Global South

**Causal links between social determinants and mental health:**

**Poverty and income inequality:** People in the lowest income quintile have 2–3 times higher odds of depression and anxiety compared to the highest quintile (odds ratio [OR] = 2.1–3.0, 95% CI: 1.8–3.5). A 10% increase in income (via cash transfers or minimum wage increases) is associated with a 5–15% reduction in psychological distress.

**Adverse childhood experiences (ACEs):** Individuals with 4+ ACEs have 4–12 times higher odds of depression, anxiety, PTSD, and substance use disorders compared to those with 0 ACEs. The population-attributable fraction for depression is estimated at 30–40%—meaning roughly one-third of adult depression could be eliminated if all ACEs were prevented.

**Discrimination and racism:** Perceived discrimination is associated with a 1.5–2.5 times higher odds of common mental disorders (OR = 1.5–2.5, 95% CI: 1.3–3.0). Among Black Americans, experiences of racial discrimination account for an estimated 15–20% of the racial disparity in depression prevalence.

**LGBTQ+ status:** Sexual minority adults have 1.5–2.0 times higher odds of depression and anxiety compared to heterosexual peers. Transgender individuals have 2–4 times higher odds of suicide attempts (OR = 2.0–4.0, 95% CI: 1.5–6.0). These disparities are largely mediated by minority stress (discrimination, victimisation, internalised stigma).

**Forced migration and displacement:** Refugees and asylum seekers have 2–5 times higher prevalence of PTSD (15–30% vs. 1–3% in general population) and 2–3 times higher prevalence of depression (20–40% vs. 5–10%). Prevalence decreases with time since resettlement but remains elevated for 10+ years.

**Housing and neighbourhood:** Living in the most deprived neighbourhood quintile is associated with 1.5–2.0 times higher incidence of psychosis (incidence rate ratio [IRR] = 1.5–2.0, 95% CI: 1.3–2.5). Housing instability (homelessness, frequent moves) is associated with 2–4 times higher odds of depression and substance use disorders.

**Preventive interventions with evidence:**

**Universal interventions:**

- Anti-poverty programmes (cash transfers, minimum wage increases): 5–15% reduction in psychological distress (effect size d = 0.10–0.30)

- Anti-discrimination laws and policies: 10–20% reduction in mental health disparities for protected groups (observational, pre-post designs)

- Universal basic income pilots: mixed evidence, small reductions in stress (d = 0.05–0.15)

- Housing first programmes for homeless individuals: 40–60% reduction in homelessness at 12 months, with secondary mental health improvements (d = 0.20–0.40)

**Selected interventions:**

- School-based anti-bullying programmes: 15–25% reduction in bullying victimisation, with corresponding 10–20% reduction in depression and anxiety symptoms (d = 0.15–0.30)

- LGBTQ+ affirming school policies (GSAs, inclusive curricula): 20–30% reduction in suicide attempts among LGBTQ+ students (OR = 0.70–0.80)

- Refugee resettlement support (housing, employment, language training): 10–20% reduction in PTSD and depression at 12 months (d = 0.20–0.40)

- Parenting programmes for families in poverty: 20–30% reduction in child behavioural problems (d = 0.30–0.50)

**Indicated interventions:**

- Trauma-focused CBT for individuals exposed to trauma: 40–60% reduction in PTSD diagnosis (number needed to treat = 3–5)

- Cognitive-behavioural prevention for children with early anxiety symptoms: 30–50% reduction in progression to full disorder (OR = 0.50–0.70)

- Supported employment for people with severe mental illness: 50–60% employment rate vs. 20–30% in standard care (d = 0.50–0.80)

To translate these findings into plain English:

**Poverty's impact on depression** is roughly equivalent in magnitude to having a first-degree relative with depression (OR ~2.5 vs. OR ~2.0). This means that being poor is about as strong a risk factor as having a genetic predisposition.

**Four or more adverse childhood experiences** increase depression risk by a factor of 4–12. For context, smoking increases lung cancer risk by a factor of 15–30. ACEs are to depression what smoking is to lung cancer—not identical, but the population impact is comparable.

**Discrimination's effect on mental health** (d = 0.20–0.40) is similar in magnitude to the effect of a moderate stressor like divorce or job loss. It is not a trivial effect—it is clinically meaningful.

**Housing first programmes** reduce homelessness by 40–60%, which is roughly as effective as statins are at reducing heart attacks (30–40% risk reduction). The mental health improvements are secondary but real.

**Trauma-focused CBT** for PTSD has a number needed to treat of 3–5, meaning you need to treat only 3–5 people to prevent one case of PTSD. This is comparable to the effectiveness of antidepressants for moderate depression (NNT = 4–7).

**What the authors acknowledge:**

Evidence is drawn primarily from the Global North; findings may not generalise to other contexts

Many cited studies are observational, not experimental

Causal inference is limited by confounding (e.g., poverty causes depression, but depression also causes poverty)

Intersectionality (multiple overlapping social determinants) is difficult to study quantitatively

Most interventions have been tested only in short-term follow-up (1–5 years), not across the life course

Implementation fidelity varies widely across studies

**What a critical reader would note:**

**No systematic search:** The authors do not report a search strategy, inclusion criteria, or risk of bias assessment. This means the review is vulnerable to selection bias—they may have preferentially cited studies that support their conclusions.

**Publication bias:** Studies showing null or negative results for social interventions are less likely to be published. The review likely overestimates the effectiveness of preventive strategies.

**Confounding by unmeasured variables:** People who receive social interventions (e.g., housing support) differ from those who do not in ways that affect mental health (motivation, social support, access to other services). Even "natural experiments" have limitations.

**Heterogeneity of outcomes:** The review lumps together different mental health outcomes (depression, anxiety, psychosis, PTSD) that may have different social determinants and different responses to intervention.

**No dose-response analysis:** The review does not quantify how much intervention is needed to produce a given effect. For someone running a self-experiment, this is critical information that is missing.

**Individual vs. structural focus:** The review emphasises structural determinants (poverty, discrimination, housing policy) that are largely outside individual control. This is scientifically accurate but provides limited guidance for someone trying to improve their own mental health through personal experimentation.

**Conflicts of interest:** The authors do not report conflicts, but the review is published in a psychiatry journal with a strong public health orientation, which may bias toward emphasising social rather than biological determinants.

For someone running their own n=1 experiment, the key insight from this review is that **social determinants operate at a structural level that individual behaviour change cannot fully address**. However, you can test the downstream mediators—the specific mechanisms through which social circumstances affect your mental health. Here is how:

### What to test (specific intervention and dose)

**Option A: Social connection intervention**

**Intervention:** Increase frequency and quality of social interactions

**Dose:** Aim for 3–5 meaningful social interactions per week (in-person preferred, video call acceptable). Each interaction should be ≥30 minutes of undivided attention (no phones, no TV)

**Rationale:** Social isolation is a key mediator between poverty/discrimination and poor mental health. Even if you cannot change your income or neighbourhood, you can change your social network

**Option B: Neighbourhood engagement intervention**

**Intervention:** Increase time spent in green/blue spaces (parks, gardens, waterfronts) within your neighbourhood

**Dose:** 120 minutes per week (can be broken into 20–30 minute sessions), ideally in natural settings with trees or water

**Rationale:** Neighbourhood deprivation affects mental health partly through lack of access to restorative environments. This is a modifiable behavioural factor

**Option C: Stress-reduction through routine**

**Intervention:** Implement a consistent daily routine with fixed sleep/wake times, meal times, and dedicated work/rest periods

**Dose:** Maintain for 8 weeks minimum. Track adherence daily

**Rationale:** Unpredictability and chaos are key mechanisms linking poverty and marginalisation to poor mental health. A predictable routine may buffer against this

### Minimum meaningful duration

**Social connection:** 4–6 weeks to see initial changes in mood; 12 weeks for sustained effects

**Neighbourhood engagement:** 4 weeks for acute effects on stress; 8–12 weeks for changes in depression/anxiety

**Routine intervention:** 8 weeks minimum; effects may take 4 weeks to appear

### What to measure (specific metrics)

**Primary outcome (choose one):**

**Depression symptoms:** PHQ-9 (free, online, 9 items, 0–27 scale). Measure weekly at the same time of day

**Anxiety symptoms:** GAD-7 (free, online, 7 items, 0–21 scale). Measure weekly

**Psychological distress:** K6 or K10 (free, online, 6 or 10 items). Measure weekly

**Secondary outcomes:**

**Social connection:** UCLA Loneliness Scale (3-item version, free, online). Measure weekly

**Perceived stress:** Perceived Stress Scale (PSS-10, free, online, 0–40). Measure weekly

**Sleep quality:** Single item: "Rate your sleep quality last night" (1–10). Measure daily

**Positive affect:** Single item: "Rate your overall mood today" (1–10). Measure daily

**Process measures (to track adherence):**

Social interactions: Count per week (log date, duration, quality 1–10)

Green space time: Minutes per week (log date, duration, type of setting)

Routine adherence: Percentage of days you followed your planned schedule

### Key confounds to control for

**Confounds you CAN control:**

**Time of year:** Seasonal affective disorder can mimic depression. Run your experiment in the same season, or control for daylight hours

**Sleep:** Track sleep duration and quality. Poor sleep is both a cause and consequence of poor mental health

**Exercise:** Physical activity has strong antidepressant effects. Log your exercise minutes and control statistically if possible

**Alcohol and caffeine:** Both affect mood and sleep. Keep intake constant during your experiment

**Major life events:** Job loss, relationship changes, bereavement. Note these in a log and exclude data from weeks with major events if possible

**Confounds you CANNOT fully control:**

**Structural discrimination:** If you are a member of a marginalised group, you cannot control how society treats you