Access barriers to support services for people who use illicit drugs in Lamu County, Kenya

¹Department of Community Health, School of Public Health, Amref International University, Nairobi, Kenya, ²Department of Pharmacy, School of Health Sciences, University of Nairobi, Nairobi, Kenya

^&Corresponding author
Ahmed Mohamed Salat, Department of Community Health, School of Public Health, Amref International University, Nairobi, Kenya

Introduction    

Support services for people who use illicit drugs (PWUD) require a holistic approach integrating medical, mental health, and social components. Ayenew et al. [1] in Ethiopia emphasise the importance of medical care, while Behrooz et al. [2], highlight the comorbidity between substance use and mental health disorders in Bangladesh. Social support services, including housing and employment assistance [3], peer support [4], and community-based programmes, are essential to address stigma and social isolation. Lennox et al. [5] advocate for a combined approach that addresses the multifaceted nature of drug use in Canada. Drug use represents a global challenge affecting communities worldwide, with particular impact on youth. In the United States of America (USA), substance use is associated with poor academic performance, health deterioration, and criminal justice involvement [6]. Drug use is also linked to social isolation [7] and community instability.

In Kenya, substance abuse is associated with addiction, disease burden, and social problems including crime and violence [8]. The National Authority for the Campaign Against Alcohol and Drug Abuse (NACADA) [9] reports that 5.2% of Kenyan youth use alcohol, 3.2% tobacco, 3.6% khat, and 2.7% cannabis, with 8.9% of youths aged 15 to 24 years affected by at least one substance.

Significant barriers to accessing support services exist across multiple countries. In India, Indu et al. [10] report drug abuse among slum-dwelling juveniles with limited service access. In Pakistan, determinants include drug availability, poverty, and low literacy [11]. Behrooz et al. [2] notes that in Bangladesh, socioeconomic factors like poverty and unemployment are significantly associated with drug use, while financial barriers and limited healthcare coverage impede access to support services; however, both studies are cross-sectional, limiting causal inference, and differ substantially from coastal Kenya in terms of healthcare infrastructure and cultural context.

Regional studies reveal distinct patterns across Africa. In Uganda, Nakibuuka et al. [12] found that sociocultural, economic, and individual factors significantly influence drug abuse among youth; in Nigeria, Lumun et al. [13] confirm poverty and education level as significant factors; in Kenya, family socioeconomic factors [14], stress levels [8], and sociodemographic factors [15] are all associated with substance abuse, particularly in slum areas. These African studies share methodological similarities with the current study but focus on urban informal settlements, while Lamu County presents an island and coastal rural context not yet adequately studied.

The drug situation in Kenya worsens with age. NACADA [9] reports that among 25 to 35 year olds, alcohol use stands at 14.0%, tobacco at 6.0%, khat at 4.8%, and cannabis at 2.1%, with 18.5% affected by at least one substance. Lamu County's proximity to Somalia (approximately 100 km) creates additional challenges, as the United Nations Office on Drugs and Crime (UNODC) [16] identifies the Somali border as a significant transit point for illicit drugs, with weak border controls enabling smuggling and contributing to organised crime and terrorism financing [17].

Access barriers in Lamu County encompass awareness, satisfaction, and socioeconomic dimensions. Awareness barriers involve lack of information about available services [18], satisfaction barriers relate to service quality and appropriateness [19], and socioeconomic barriers include income, education, and employment status [20]. The National Crime Research Centre (NCRC) [21] reports illicit brew possession at 46.6% and narcotic drug cases at 27.6% in Lamu County, while the Kenya National Bureau of Statistics (KNBS) [22] indicates a 35.1% poverty rate, with 37.5% among teenagers aged 14 to 17 years. These statistics underscore the urgency of addressing access barriers and highlight the evidence gap this study seeks to address.

The prevalence of drug use and its associated challenges in Kenya constitute a serious public health problem. According to NACADA [9], one in six Kenyans aged 15 to 65 years engages in the use of illicit drugs or substances of abuse, and Kenya records one of the highest rates of Disability-Adjusted Life Years (DALYs) from alcohol use disorders in Africa [9]. Drug and substance abuse remain major problems along the coastal region [23]. In Lamu County, the NCRC [21] documents a 46.6% rate of illicit brew possession and a 27.6% rate of narcotic drug cases, pointing to systemic failures in both regulatory frameworks and the availability of effective support services.

The socioeconomic conditions in Lamu County compound the problem. According to KNBS [22], the poverty headcount rate stands at 35.1% overall, with 37.5% of teenagers aged 14 to 17 years living below the poverty line. Financial hardship, geographic isolation, and inadequate health infrastructure combine to create structural barriers that deter help-seeking, particularly among low-income and marginalised populations [9,24]. Social stigma further discourages individuals from accessing services [16], while centralised urban service models are poorly suited to Lamu's dispersed island and mainland communities [24]. Despite these documented challenges, there remains a critical gap in evidence specifically examining awareness of, satisfaction with, and socioeconomic barriers to support services for PWUD in Lamu County.

The study was guided by the Social Determinants of Health (SDH) theory, shaped by Marmot's study [25], and the Health Belief Model (HBM) developed by Janz et al. [26]. The SDH theory locates drug use and service access within broader social, economic, and environmental structures and informed the assessment of socioeconomic barriers. The HBM explains help-seeking through perceived susceptibility, severity, benefits, barriers, cues to action, and self-efficacy, and informed the assessment of awareness of and satisfaction with support services. Because the HBM focuses on individual cognition, it was applied alongside the SDH theory to account for structural determinants of access.

Empirical studies on substance use and service access report varied findings. Kupi et al. [27] found that literacy levels, drug availability, and county-level strategies had no significant effect on youth substance abuse in Lamu County; while this study shares the same geographic context, it focused on supply-side and demographic factors rather than access barriers to support services, leaving a gap this study directly addresses. Voon et al. [28] reported that 34.8% of PWUD in Vancouver faced healthcare access barriers, with pain severity significantly correlated with access difficulties; although the Canadian healthcare infrastructure differs substantially from Lamu County, the finding that financial and service-related barriers deter access has clear relevance to the current study.

Enkelmann et al. [29] documented high hepatitis C virus (HCV) prevalence (36%) among people who recently started injecting drugs across eight German cities, with 41% unaware of their positive status despite regular healthcare contact; the study's large multi-city sample is a methodological strength, though its focus on a single clinical indicator limits broader generalisability to service access barriers. In Norway, Stallvik et al. [19] found that satisfaction with health services among 491 service users with substance use disorders was negatively associated with older age, larger municipality size, psychological distress, and recent substance use, with only 54% reporting high overall satisfaction; the study's longitudinal design and large sample are methodological strengths, though findings from a well-resourced Scandinavian healthcare system may not translate directly to low-resource coastal Kenyan contexts.

Maccagnan et al. [30] analysed United Kingdom (UK) data and found that people who did not use cannabis scored higher on life satisfaction scales than those who did, with well-being losses valued at £5,600 (approximately US$7,200) annually per person using cannabis; this economic focus on national well-being costs differs from the current study's focus on service access barriers, limiting direct comparability. Al Ghafri et al. [31] reviewed 24 studies from Muslim communities and identified psychosocial barriers such as denial, lack of trust, and privacy concerns, alongside organisational obstacles including affordability, stigma, and awareness gaps; this systematic review provides the most directly comparable framework given Lamu County's predominantly Muslim population, though the review's broad geographic scope means context-specific Kenyan evidence remains limited.

Several Kenya-focused studies provide additional evidence. Okoyo et al. [32] documented 86% substance use prevalence across selected counties among 3,600 participants, with higher rates among males aged 45 to 53 years, providing a valuable national benchmark, though the study did not examine access barriers in depth. Wainaina et al. [33] identified unmarried status and peer pressure as significant relapse factors among alcohol rehabilitation patients in Kiambu County; the study's focus on a Kenyan rehabilitation setting is a strength, though its concentration on relapse rather than initial access limits comparability. Makau et al. [34] found that 45% of hospital inpatients in Kitui County reported recent substance use, with 45% meeting criteria for substance use disorder, demonstrating clinical burden in a Kenyan facility context.

Collectively, these studies confirm that while Kenyan evidence exists, none specifically examines access barriers in the geographically distinct and underserved context of Lamu County. Common themes across all settings include awareness gaps, socioeconomic barriers, and the need for contextually sensitive approaches. This study therefore aimed to assess the level of awareness of support services among PWUD in Lamu County, to evaluate their level of satisfaction with the support services they receive, and to investigate socioeconomic factors affecting their access to support services.

Methods     

The study employed a cross-sectional mixed-methods design integrating qualitative and quantitative approaches to examine access barriers to support services for PWUD in Lamu County, Kenya. Three strategically selected wards, Mkomani (urban), Hindi (mainland), and Faza (island), were included to capture geographic diversity. Using Cochran's formula [35] with a 95% confidence level and 5% margin of error, a minimum sample size of 207 participants was calculated and adjusted to 228 to account for non-response. Of the 228 targeted participants, 207 PWUD were recruited and 206 returned complete questionnaires, giving a response rate of 99.52% among recruited participants and an overall enrollment rate of 90.35% (206 of 228). Data were collected from these 206 PWUD and from 48 key stakeholders through structured questionnaires and semi-structured key informant interviews.

A two-stage sampling approach combined stratified sampling across the three wards and snowball sampling to access this hard-to-reach population. Anonymous questionnaires and trained peer-linked research assistants were used to reduce social desirability bias. The study included participants aged 15 to 65 years, with special ethical protections for minors, who represented a critical demographic given the finding by NACADA [9] that 8.9% of youths aged 15 to 24 years were affected by at least one substance. Prior to data collection, all participants provided written informed consent. For participants with low literacy, witnessed verbal consent was obtained in the presence of a research assistant who was not a member of the participant's household. Minors (aged 15 to 17 years) provided assent, with parental or guardian consent also obtained. Confidentiality was maintained by anonymising all data at the point of collection, assigning participant codes in place of names, and storing data on password-protected devices accessible only to the principal researcher. Participation was entirely voluntary, and participants were informed of their right to withdraw at any time without consequence. All procedures were conducted in accordance with approvals granted by the Amref Ethics and Scientific Review Committee (ESRC) and the National Commission for Science, Technology and Innovation (NACOSTI).

Research instruments underwent pilot testing with 23 PWUD in Mombasa County, achieving Cronbach's alpha values between 0.856 and 0.931 [36]. Quantitative data were analysed using the Statistical Package for the Social Sciences (SPSS) version 28, including descriptive statistics, Pearson correlation analysis to examine bivariate associations between study variables, and multiple linear regression to identify significant independent predictors of service access, reported with 95% confidence intervals (CI). Prior to regression, multicollinearity was assessed using Variance Inflation Factor (VIF) values, and normality of residuals, homoscedasticity, and independence of errors were examined through residual plots. The mediation analysis followed the four-step approach of Baron et al. [37] to examine the intervening effects of government policies, resource availability, and community support, with model significance assessed through Analysis of Variance (ANOVA).

Qualitative data from 48 key informant interviews underwent thematic analysis following the six-phase framework of Braun et al. [38]: (1) familiarisation with data; (2) generating initial codes; (3) searching for themes; (4) reviewing themes; (5) defining and naming themes; and (6) producing the report. Qualitative and quantitative findings were subsequently triangulated to provide a comprehensive account of access barriers.

Results     

Response rates were 99.52% among recruited PWUD (206 of 207), 100% among government officials, 80% among healthcare professionals, 88.89% among administrative staff, and 77.78% among community representatives. The 206 complete responses represented 90.35% of the adjusted sample target of 228 and formed the analytical sample for all quantitative analyses. All measurement scales exceeded the conventional Cronbach's alpha threshold of 0.70 [36], with coefficients of 0.901 for awareness of support services, 0.896 for level of satisfaction, 0.931 for socioeconomic factors, 0.929 for intervening variables, and 0.856 for access to support services (Table 1).

Access to support services showed strong positive correlations with awareness (r = 0.742), satisfaction (r = 0.726), socioeconomic factors (r = 0.767), and the intervening variables of government policies, resource availability, and community support (r = 0.721), all significant at p < 0.001. The strongest correlation among the explanatory variables was between awareness and the intervening variables (r = 0.881, p < 0.001) (Table 2).

Regression assumptions were satisfied: all VIF values were below 5 (awareness 3.14; satisfaction 3.67; socioeconomic factors 4.02), residuals were normally distributed, and residual plots confirmed homoscedasticity and independence of errors. The three predictors together explained 64.1% of the variance in access to support services (R² = 0.641; F(3, 202) = 120.097, p < 0.001), a large effect (Cohen's f² = 1.79). Awareness showed the strongest standardised association with access (β = 0.328, B = 0.293, 95% CI (0.160, 0.426), t = 4.281, p < 0.001), followed by socioeconomic factors (β = 0.274, B = 0.270, 95% CI (0.074, 0.466), t = 2.712, p = 0.007) and level of satisfaction (β = 0.258, B = 0.251, 95% CI (0.098, 0.404), t = 3.237, p = 0.001) (Table 3).

All four mediation models were statistically significant. Awareness, satisfaction, and socioeconomic factors explained 64.1% of the variance in access (model 1: R² = 0.641; F(3, 202) = 120.097, p < 0.001) and 59.1% of the variance in the intervening variables (model 2: R² = 0.591; F(3, 202) = 254.213, p < 0.001). The intervening variables alone explained 52.0% of the variance in access (model 3: R² = 0.520; F(1, 204) = 220.686, p < 0.001), while the combined model explained the highest proportion, 64.8% (model 4: R² = 0.648; F(4, 201) = 92.485, p < 0.001) (Table 4, Table 5).

In step 2, awareness (p < 0.001) and socioeconomic factors (p = 0.001) significantly predicted the intervening variables, whereas satisfaction did not (p = 0.422). In step 4, awareness (p = 0.047), satisfaction (p = 0.001), socioeconomic factors (p = 0.030), and the intervening variables (p = 0.044) all remained significant predictors of access, with smaller coefficients for awareness (B = 0.177 compared with 0.293) and socioeconomic factors (B = 0.223 compared with 0.270) than in step 1, a pattern consistent with partial mediation (Table 6).

Discussion     

This study examined access barriers to support services for PWUD in Lamu County, Kenya, focusing on awareness of available services, satisfaction with service delivery, and socioeconomic factors affecting accessibility. The findings indicate that individual, structural, and systemic factors are jointly associated with reduced service utilisation, with direct implications for how primary healthcare services in the county identify, reach, and retain PWUD. All findings are interpreted as associations because the cross-sectional design does not permit causal inference.

Awareness emerged as the factor most closely linked to service access. Participants demonstrated limited knowledge of service locations, types of services offered, and access procedures, consistent with Grieb et al. [18], who documented low awareness of mobile health clinics among PWUD as a major utilisation barrier, and Enkelmann et al. [29], who found that many people who inject drugs were unaware of their hepatitis C status despite regular healthcare contact. In Lamu County, dispersed island and mainland settlements constrain information dissemination. The close relationship between awareness and the structural variables suggests that supportive policies, adequate resources, and community support may extend the reach of awareness campaigns. For primary healthcare, this implies that facilities and community health workers are well placed to provide routine information on where and how PWUD can obtain support.

Low satisfaction with support services was associated with lower access. Participants cited concerns about service quality, respect and dignity from providers, privacy, and service availability. These findings resonate with Stallvik et al. [19], who found that satisfaction among service users with substance use disorders in Norway was negatively associated with psychological distress and recent drug use, and with Al Ghafri et al. [31], who reported that affordability, awareness gaps, and stigma impeded treatment access in Muslim communities. Because satisfaction remained associated with access after structural factors were considered, the experience of care at the point of service appears to shape whether PWUD continue to engage. Primary healthcare providers can therefore influence continued engagement by offering confidential, respectful, and nonjudgmental care.

Socioeconomic hardship, including low income, unemployment, and poor housing, was associated with reduced service accessibility. This aligns with the SDH framework [25], which emphasises how social, economic, and environmental conditions shape health outcomes, and is notable given the high poverty rate reported for Lamu County [22]. Poverty and unemployment are associated with drug use [1,2], while financial constraints are simultaneously associated with reduced access to support services, a compounding disadvantage also documented in Pakistan [11], Uganda [12], and other Kenyan settings [14,15].

Kozhimannil et al. [20] and Ayenew et al. [1] similarly reported greater barriers among low-income, unemployed, and less educated populations, Voon et al. [28] found that financial barriers influenced access to drug-related health services, and Al Ghafri et al. [31] identified fear of legal repercussions as a major obstacle. For primary healthcare provision, these findings support fee waivers, transport support, and integration of drug-related services into facilities closer to where PWUD live.

The mediation findings indicate that government policies, resource availability, and community support partially account for the relationship between individual-level factors and service access. Awareness and economic circumstances appear to operate partly through the structural environment while retaining an independent association with access. Interventions directed at individuals are therefore unlikely to achieve their full effect without parallel investment in policy, resources, and community support, and embedding drug-related services within primary healthcare and community health structures offers one route to align these levels.

Lamu County's awareness deficit mirrors findings from Germany, where Enkelmann et al. [29] found that many people who recently started injecting drugs were unaware of their hepatitis C status despite healthcare contact, although awareness in Lamu County appears lower, potentially reflecting limited health infrastructure and communication channels. Satisfaction appears lower than that reported in Norway [19], suggesting that service quality challenges may be more severe in resource-limited settings. Lamu County's position as a border region with Somalia adds further difficulty, as UNODC [16] identifies weak border controls as enabling drug smuggling, which intensifies both supply and demand challenges.

This study has several limitations. The cross-sectional design measured exposures and outcomes at a single point in time, so the temporal sequence linking awareness, satisfaction, and socioeconomic conditions to service access cannot be established, and no causal inference is possible. Because drug use is criminalised in Kenya, PWUD may have underreported drug use or overstated positive engagement with services owing to social desirability bias, and recall bias may have affected reports of past service experiences. Snowball sampling through chain referrals within tightly knit social circles introduces selection bias, as participants with strong peer connections may be overrepresented while isolated PWUD, who may face the greatest barriers, may be underrepresented, limiting generalisability. Finally, the Baron et al. [37] is exploratory, and bootstrap-based mediation analysis would provide stronger evidence.

Conclusion     

Awareness of support services among PWUD in Lamu County was low, with marked gaps in knowledge of service locations, types of services offered, and access procedures, making awareness the most important intervention point identified in this study. Satisfaction with support services was low and closely associated with service utilisation, indicating that improvements in service delivery quality are linked to access. Socioeconomic factors were also associated with service accessibility, as poverty, unemployment, and financial constraints create compounding barriers. Government policies, resource availability, and community support play partial intervening roles, and the combined model explained a substantial proportion of variance in service access. All findings should be interpreted as associations. In the immediate term, awareness of available support services for PWUD should be increased across Lamu County. Campaigns should be comprehensive, clear, and tailored to the needs of different community groups across urban (Mkomani), mainland (Hindi), and island (Faza) settings, and should provide detailed information on service types, access procedures, and eligibility criteria. Local leaders, health professionals, and non-governmental organisations (NGOs) should actively disseminate information through community meetings, health barazas, radio programmes, and social media platforms, given the strong association between awareness and access documented in this study. In the medium term, service providers should prioritise enhancing user satisfaction by addressing quality of care, respect and dignity, privacy, and service availability. Regular training in interpersonal communication, cultural sensitivity, and confidentiality should be implemented, alongside feedback mechanisms to monitor satisfaction and identify areas for improvement. Socioeconomic barriers should be addressed through targeted financial assistance programmes covering treatment costs and transportation for low-income individuals, complemented by employment training and vocational skills programmes for PWUD. In the long term, government policies should be strengthened to improve rehabilitation service accessibility, including increased treatment centres, improved staffing, and greater integration of drug treatment within primary healthcare systems. Policies should also address stigma through public awareness campaigns. Community-based organisations, health professionals, and NGOs should collaborate to create inclusive support networks offering peer mentorship, counselling, and family reintegration services. Future research should employ longitudinal designs to establish temporal relationships and causal pathways, intervention studies to evaluate the effectiveness of specific awareness campaigns and service quality improvements, and participatory action research engaging PWUD in the codesign of more accessible and responsive services in Lamu County and similar coastal Kenyan settings. Bootstrap-based mediation analyses are also recommended to strengthen mediation evidence.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

All the authors read and approved the final version of this manuscript.

Tables     

Table 1: internal consistency reliability (Cronbach's alpha) of the questionnaire scales measuring awareness, satisfaction, socioeconomic factors, intervening variables, and access to support services for people who use illicit drugs in Lamu County, Kenya

Table 2: Pearson correlation coefficients between access to support services, awareness, satisfaction, socioeconomic factors, and intervening variables among people who use illicit drugs in Mkomani, Hindi and Faza Wards, Lamu County, Kenya (N = 206)

Table 3: multiple linear regression of awareness, satisfaction and socioeconomic factors as predictors of access to support services among people who use illicit drugs in Mkomani, Hindi and Faza Wards, Lamu County, Kenya (N = 206)

Table 4: coefficients of determination (R²) of the four-step Baron and Kenny mediation models testing government policies, resource availability, and community support as intervening variables between awareness, satisfaction, socioeconomic factors and access to support services among people who use illicit drugs in Mkomani, Hindi and Faza Wards, Lamu County, Kenya (N = 206)

Table 5: analysis of variance (ANOVA) results of the four-step Baron and Kenny mediation models testing government policies, resource availability and community support as intervening variables between awareness, satisfaction, socioeconomic factors and access to support services among people who use illicit drugs in Mkomani, Hindi and Faza Wards, Lamu County, Kenya (N = 206)

Table 6: regression coefficients of the four-step Baron and Kenny mediation models testing government policies, resource availability, and community support as intervening variables between awareness, satisfaction, socioeconomic factors, and access to support services among people who use illicit drugs in Mkomani, Hindi, and Faza Wards, Lamu County, Kenya (N = 206)

References