Barriers to 12-month treatment of common anxiety, mood, and substance use disorders in the World Mental Health (WMH) surveys

The study addresses why many individuals with common mental disorders do not receive timely treatment despite evidence-based interventions being available. Using the WMH Surveys, prior work showed low perceived need and multiple bottlenecks in the treatment cascade. This report focuses on barriers among individuals with 12-month anxiety, mood, or substance use disorders who recognize a need for care, aiming to quantify the distribution of barriers, identify correlates, and assess how barriers mediate predictors of treatment and delays. Understanding these barriers is critical to designing interventions to reduce the global treatment gap.

Prior frameworks distinguish barriers by predisposing factors (beliefs about effectiveness; stigma), enabling factors (financial and other practical obstacles), and low perceived disorder severity. Previous WMH and other studies found that lack of perceived need is common, each barrier type is prevalent, and many individuals report multiple barriers. Country-level healthcare resources (e.g., physician density, health spending) have shown modest associations with treatment receipt. The current study builds on this literature by examining barrier distributions and their mediating roles across multiple disorders and countries.

Design and sample: Secondary analysis of WMH Surveys: 22 community-based surveys in 19 countries (administered 2001–2019), combined n=102,812 adults (Part I); Part II sample n=51,520 included all with any lifetime disorder plus a subset of others. Weighting adjusted for sampling, subsampling into Part II, and post-stratification; design effects accounted for clustering. Focused analysis on respondents with 12-month DSM-IV disorders who perceived a need for treatment (n=5,136). Subgroups: those with perceived need who did not receive 12-month treatment and were asked reasons (n=2,444; excluding n≈1,097 with very recent perceived need and reason-skip and additional small missing), and those who received treatment with delays (>30 days after recognizing need) and were asked reasons (n=926). Assessment: Trained lay interviewers administered the CIDI 3.0. Disorders included five anxiety disorders (GAD, Panic/Ago, SP, SoP, PTSD), two mood disorders (MDD, BD spectrum), and two substance use disorders (AUD, DUD). Disorder severity classified as severe, moderate, or mild based on diagnostic and impairment criteria. Perceived need assessed differently among treated vs. untreated respondents via standardized questions. Twelve-month treatment defined as any contact in past 12 months with general medical, psychiatrist, other mental health, human services, or complementary/alternative providers for emotional or substance problems. Barriers: Fourteen items grouped into five domains: (i) low perceived severity (problem will improve on its own; not very bothersome; prefer to handle alone); (ii) financial (insurance would not pay; cannot afford); (iii) other enabling/practical (transportation/scheduling; unsure where to go; time/inconvenience; unable to get appointment); (iv) low perceived treatment effectiveness (dissatisfaction with available treatments; belief treatment won’t work; prior unhelpful treatment); (v) perceived stigma (concerns about others finding out; worry about involuntary hospitalization). Binary coding per domain (any item endorsed = yes). Parallel items assessed reasons for delayed help-seeking among those treated with delay. Predictors: Socio-demographics (sex; age 18–29/30–44/45–59/60+; education categorized to country-specific low to high; employment status: employed, homemaker, retired, student, disabled/other/unemployed; insurance: private and/or public/universal), disorder type/number/severity, prior treatment history (types and number of provider types seen before past 12 months; past treatment modalities; perceived helpfulness), and country-level indicators (per capita non-psychiatrist MDs; proportion of GDP spent on healthcare). Statistical analysis: Cross-tabulations estimated perceived need and treatment by disorder. Multivariable Poisson regression with robust standard errors produced risk ratios (RRs) with design-based 95% CIs for outcomes: (1) receiving any 12-month treatment (among those with perceived need), and (2) receiving treatment without delay vs. with delay (among treated with perceived need). Pooled within-country models included country dummies; multi-level models for country-level predictors adjusted for individual-level composition. Mediation by barrier domains assessed via sequential exclusion: re-estimating models after removing respondents who endorsed a given barrier domain and comparing changes in predictor associations. Two-sided α=0.05; design-based Wald χ² tests used.

- Prevalence and cascade: Across WMH, 14.2% had any 12-month disorder; 13.9% received any 12-month treatment. Perceived need among cases: 40.7%. Treatment receipt given perceived need: 30.6% overall (vs. 2.5% without perceived need). Higher clinical severity associated with greater perceived need and treatment (e.g., treatment given perceived need: severe 39.6%, moderate 28.2%, mild 18.1%). - Barriers among those with perceived need who did not receive treatment (n=2,444): low perceived severity 52.9%; low perceived treatment effectiveness 44.8%; other enabling/practical 40.2%; financial 32.9%; stigma 20.6%. Mean number of barriers 1.9; 14.8% reported none; 32.9% reported 3+. - Barriers among those who delayed treatment (n=926): low perceived severity 83.4%; other enabling/practical 51.0%; low perceived treatment effectiveness 49.4%; stigma 37.0%; financial 36.8%. Mean number of barriers 2.6; 2.3% reported none; 50.1% reported 3+. Each barrier type was significantly more prevalent for delays than for not receiving treatment (χ² range 3.8–199.8; p=0.050–<0.001). - Predictors of receiving treatment (among perceived-need cases; n≈4,039): higher per capita non-psychiatrist MDs (RR≈1.1); having health insurance (RR=1.5); higher education (education not high associated with lower treatment, RR=0.9); disorder type (Panic/Ago, GAD, PTSD, MDD, BD; RR≈1.1–1.4), and greater severity (RR≈1.2–1.3) increased treatment probability. Prior treatment history: having seen only one provider type previously associated with lower odds (RR≈0.5–0.6 vs. none), but greater number of prior provider types strongly increased treatment (2 types RR=2.3; 3 RR=4.4; 4 RR=7.9; 5 RR≈15.9). Prior treatment perceived helpful: RR=3.1; even prior unhelpful: RR=1.9. - Mediation by barriers (sequential exclusion): The education effect (lower education → lower treatment) was largely mediated by low perceived treatment effectiveness and nonfinancial enabling/practical barriers (education χ² dropped to non-significance when excluding these barrier reporters). The positive effect of insurance on treatment disappeared after excluding respondents endorsing financial barriers. For disorders: low perceived effectiveness and stigma mediated associations for GAD and BD; additional mediation by nonfinancial enabling for BD and low perceived severity for GAD. Elevated RRs by severity and most disorder-type differences were not fully mediated by any single barrier. Strong effects of multiple prior provider types and perceived helpful prior treatment were reduced but remained significant after excluding each barrier, especially when excluding low severity. - Predictors of timely (vs. delayed) treatment (n=1,595): younger age 18–59 had lower probability of timely care (RR=0.8); being a student (RR=1.6), homemaker (RR=1.3), retired (RR=1.3), and disabled/other/unemployed (RR≈1.2; some CIs include 1.0) associated with greater timeliness relative to employed. GAD and SP associated with lower timeliness (RR=0.8 each). Higher national healthcare spending share (GDP) associated with slightly lower timeliness (RR=0.9). Mediation: age association mediated by perceived treatment effectiveness; employment status associations mediated by financial barriers and low perceived severity; GAD association mediated by low perceived severity and nonfinancial enabling; the healthcare spending association mediated by low perceived severity.

Even after individuals recognize a need for mental healthcare, diverse and often multiple barriers impede care initiation and timeliness. Low perceived disorder severity was the most common barrier for both not seeking care and delaying care, suggesting that beliefs about natural remission and self-management critically deter timely help-seeking. Nonfinancial practical barriers and skepticism about treatment effectiveness were also prominent, while financial barriers, though important for some, were less prevalent than often assumed. The mediation analyses indicate that socio-economic and system-level predictors (e.g., insurance, physician supply, education) operate partly through specific barrier domains, especially perceived effectiveness, financial constraints, and practical access issues. These findings imply that policies focusing solely on financial coverage are insufficient; multi-component strategies addressing mental health literacy (severity and chronicity), treatment effectiveness beliefs, and practical access barriers are required to reduce unmet need and delays.

A wide range of barriers persist among people with common mental disorders even after acknowledging need for treatment, and most individuals report multiple barriers. Barriers are widespread and not predominantly socio-economic in nature. Efforts to reduce unmet need must be multi-pronged: increase recognition of need, counter inaccurate beliefs that treatments are ineffective or problems resolve on their own, and address practical access issues alongside financial obstacles. Future work should tailor multi-component interventions and evaluate their impact on both initiation and timeliness of care across diverse health systems.

- Self-reported service use, perceived need, barriers, and helpfulness may be affected by recall and reporting biases; temporal ordering between perceived need/barriers and treatment seeking cannot be definitively established. - Focused on selected DSM-IV anxiety, mood, and substance use disorders; findings may not generalize to other conditions (e.g., psychotic or externalizing disorders beyond SUDs). - Some analyses had limited power due to small subsamples. - Barriers were measured at the respondent level rather than per disorder, limiting insight into disorder-specific help-seeking among comorbid cases; number of 12-month disorders not fully modeled beyond severity classification. - Barriers were only assessed among respondents with perceived need; predictors of treatment among those without perceived need were evaluated separately in prior work, limiting direct comparability. - Dropout from treatment (another potential barrier to adequate care) was not analyzed. - Cross-national differences in barrier distributions and mediation effects were not examined; pooled analysis may mask country-specific patterns.