I am doing research on pathway to care and treatment barriers in substance use disorder. I have used who Encounter Form and BACE v3 . I want prevalence of barriers for sample size calculation
Searching PubMed
"substance use disorder" AND "barriers to treatment" AND prevalence
Searching PubMed
BACE "barriers to care" substance use disorder prevalence
Searching PubMed
"pathway to care" substance use disorder barriers prevalence
Searching the Web
BACE v3 "barriers to care evaluation" substance use disorder prevalence sample size
Reading Article
PMID: 27736680
Reading Article
PMID: 31640647
Searching PubMed
"substance use disorder" AND "treatment gap" AND barriers prevalence
Reading Web Page
https://pmc.ncbi.nlm.nih.gov/articles/PMC3379935
Searching PubMed
Clement "Barriers to Access to Care Evaluation" BACE development psychometric
Searching PubMed
"WHO encounter form" "pathway to care" barriers treatment substance use
Searching the Web
prevalence treatment barriers substance use disorder Africa LMIC BACE "pathway to care" sample size calculation
Excellent - I now have enough data from multiple strong sources. Let me compile the comprehensive answer.
Here is a structured synthesis of prevalence estimates for treatment barriers in substance use disorder (SUD), directly usable for your sample size calculation, organized by instrument and source.
Prevalence of Barriers to Care in SUD - Reference Data for Sample Size Calculation
1. BACE v3 Item-Level Prevalence Data
The original BACE v3 development paper (Clement et al., 2012, PMC3379935) provides the most direct prevalence figures from a secondary mental health service population. Key findings for individual barrier items:
| Barrier Item | % Rating as Any Barrier | % Rating as Major Barrier |
|---|---|---|
| Fear of employment consequences (stigma) | ~89% | 39% |
| Concern about being seen as a bad parent (stigma) | ~88% | 38% |
| Difficulty taking time off work | moderate-high | ~30-35% |
| Too unwell to ask for help | moderate-high | ~30% |
| Previous bad experiences with services | moderate | ~25-30% |
| Wanting to solve the problem alone | moderate | ~25% |
| Not wanting it on medical records | moderate | ~25% |
| Items dropped (endorsed as major barrier by <10%) | - | <10% |
Key rule for BACE v3 item retention: Items were dropped if <10% endorsed them as a major barrier. This implies all retained items have a major-barrier prevalence of ≥10%, typically ranging from 20-39% in mental health/SUD populations.
2. Meta-Analytic Estimates for SUD Barriers (Best for Sample Size Formula)
From Cumming et al. (2016) - systematic review and meta-analysis of methamphetamine treatment barriers (PMID: 27736680, Drug Alcohol Depend, N = 6 quantitative studies):
| Barrier | Pooled Prevalence | 95% CI |
|---|---|---|
| Embarrassment / stigma | 60% | 54-67% |
| Belief treatment unnecessary | 59% | 54-65% |
| Prefer to withdraw alone | 55% | 45-65% |
| Privacy concerns | 51% | 44-59% |
These four are the most-endorsed barriers and represent the highest-powered estimates for sample size purposes.
3. BACE Used in SUD Population - Ethiopian Study (Rural LMIC)
From Zewdu et al. (2019), Substance Abuse Treatment, Prevention, and Policy, population-based study using BACE in alcohol use disorder (AUD):
| Barrier | Prevalence |
|---|---|
| Wanting to handle the problem alone | High (most common) |
| Thinking problem would resolve itself | High |
| Not feeling bothered enough | High |
| Unsure where to go | Moderate |
| Financial / cost concerns | Moderate |
This LMIC-specific data is most relevant if your study setting is Africa or South Asia.
4. South African Young Adult SUD Study (Most Directly Relevant if COSUP-type setting)
From PMC9675246 (Tshwane, South Africa, N = 206 young adults on treatment):
| Barrier | % Endorsed (any degree) |
|---|---|
| Substance use sites too few/far | 54.4% |
| Healthcare sites lack enough workers | 56.8% |
| Liked using, not ready to give up | 57.8% |
| Didn't think I needed help | 56.3% |
| Substance use seemed normal | 51.9% |
| Didn't want to share personal life | 45.6% |
| No family moral support | 36.9% |
5. Nepal BACE Study - Treatment Coverage in AUD (PMID: 31640647)
Luitel et al. (2019), BMC Public Health, population-based survey (N = 1983 baseline / 1499 follow-up):
- Treatment coverage for AUD at baseline: 5.1% (i.e., ~95% untreated)
- Treatment gap confirms the vast majority encounter barriers
- Study was noted as underpowered to detect change - highlighting the importance of an adequate sample size
Recommended Prevalence Estimate for Your Sample Size Calculation
Conservative choice (most stable, based on meta-analysis):
Use p = 0.50 (50%) - this gives the maximum variance and the largest, most conservative sample size. This is the standard recommendation when:
- You expect a moderate average barrier prevalence
- You are unsure which single barrier is primary
- You want to power the study to detect any barrier
If using a specific primary barrier:
| Barrier Domain | Recommended p | Source |
|---|---|---|
| Stigma/embarrassment | 0.60 | Cumming et al., meta-analysis |
| Perceived no need for treatment | 0.56-0.59 | Multiple sources |
| Privacy concern | 0.51 | Cumming et al. |
| Financial barriers | 0.40-0.50 | Ethiopia/Nepal/SA studies |
| System/structural barriers | 0.50-0.57 | South Africa study |
Sample Size Formula (Cross-Sectional Prevalence Study)
$$n = \frac{Z^2 \cdot p(1-p)}{d^2}$$
Where:
- Z = 1.96 (95% confidence), or 2.576 (99%)
- p = expected prevalence of the target barrier (use values above)
- d = acceptable margin of error (typically 0.05)
Example using p = 0.50, d = 0.05:
$$n = \frac{(1.96)^2 \times 0.50 \times 0.50}{(0.05)^2} = \frac{3.84 \times 0.25}{0.0025} = \mathbf{384}$$
If using p = 0.60 (stigma barrier):
$$n = \frac{3.84 \times 0.60 \times 0.40}{0.0025} = \frac{0.9216}{0.0025} = \mathbf{369}$$
Add 10-15% for non-response/attrition, giving a final target of approximately 385-440 for most designs.
Key Citations
- Cumming et al. (2016) - Meta-analysis of treatment barriers, PMID 27736680
- Clement et al. (2012) - BACE v3 development and prevalence data, PMC3379935
- Luitel et al. (2019) - BACE in AUD Nepal, PMID 31640647
- Zewdu et al. (2019) - BACE + AUDIT in rural Ethiopia, PDF
- South African COSUP study (2022), PMC9675246
Practical recommendation: If your study is in an LMIC or Sub-Saharan Africa context, use p = 0.55 (moderate barriers) or p = 0.50 (conservative) with d = 0.05, yielding n ≈ 384. If you are using the WHO Encounter Form to characterize the pathway to care and BACE v3 as the primary barrier measure, cite the Luitel Nepal study and the Clement BACE validation paper as your prevalence justification.
This is a shared conversation. Sign in to Orris to start your own chat.