Harmful substance use (HSU) is a major challenge globally with far-reaching detrimental effects on social, economic, and health systems.¹ According to the World Health Organization (WHO), in 2019, more than 180 000 deaths were directly linked to drug use disorders.² Recent WHO Global Health Estimates and Global Burden of Disease analyses indicate that, while cardiometabolic and tobacco-related exposures remain major global risk factors, drug use disorders are still an important contributor to morbidity and mortality – particularly in younger age groups and certain regions – and were ranked among the top 20 causes of Disability-Adjusted Life Years (DALYs) for people aged approximately 10–49 years in 2019.³ According to the WHO’s 2024 Global Status Report, harmful alcohol consumption in 2019 accounted for an estimated 2.6 million deaths globally (≈ 4.7% of all deaths) and about 115.9m DALYs (≈ 4.6% of all DALYs).³ A report by the United Nations Office on Drugs and Crime⁴ provides global estimates of substance use prevalence according to the substance. According to their report, the best estimate for global prevalence of cannabis use is at 4.12%, opioids 1.21%, cocaine 0.42%, amphetamines and prescription stimulants 0.68% and ecstasy 0.39%.⁴ In low- and middle-income countries (LMICs), the best estimate for the prevalence of cannabis use ranges from 2.77% to 12.00%, opioids from 1.00% to 3.20%, cocaine from 0.04% to 2.70%, amphetamines and prescription stimulants from 0.28% to 1.26% and ecstasy from 0.23% to 2.84%.⁴

Harmful substance use is an umbrella term used to refer to licit and illicit substances, including alcohol, nicotine, and drugs such as heroin, cocaine, cannabis, amphetamines, methamphetamines, and pharmaceutical drugs such as opioids and benzodiazepines.⁵ Harmful substance use can therefore be defined as the use of substances harming an individual, which may include problems in the areas of health, social life, finances, emotional problems, and possibly legal problems.⁵ The social burden of HSU is significant and is commonly associated with health concerns, impairment in education, and an increasing tendency towards crime and unemployment.⁶ This burden weighs more heavily on developing countries as they do not have the infrastructure for managing the consequences of substance use.⁷ The World Mental Health Survey showed that 76% to 85% of the patients with severe mental illnesses (e.g. substance use disorders [SUDs], anxiety or mood disorders) in LMICs received no treatment for their conditions over 12 months.⁸

In high-income countries, several evidence-based brief intervention models are used for HSU, including motivational interviewing (MI) alone or with cognitive-behavioural therapy (CBT), stress management, problem-solving, case management and community contingency therapy.^9,10 However, fewer studies have been conducted in low-resource settings.¹¹ Multiple factors make delivering effective treatment for HSU challenging in LMICs.¹² These challenges include the large geographical spread of health facilities, poorly trained treatment providers to deliver the interventions, stigmatising attitudes by providers towards those presenting with HSU, governance that accords low priority for HSU, heavy staff workload,¹³ poorly resourced health facilities, and minimal funding to employ adequate staff among other factors; all contribute towards insufficient services offered to patients presenting with HSU.¹¹

Motivational interviewing is a collaborative, client-centred counselling approach focused on strengthening an individual’s motivation and commitment to change.¹⁴ It is not a single uniform intervention, but encompasses several formats that differ in dose, structure, and application, including full or standard MI, typically delivered over multiple sessions, brief MI, which is shorter and adapted for time-limited settings, and MI-informed or MI-consistent approaches, in which practitioners apply MI principles without delivering a full protocol.^14,15,16 These variations are widely recognised in both foundational MI literature and empirical reviews and are important to consider when comparing outcomes across studies.^17,18

The effectiveness of MI as an intervention for substance use in high income countries (HICs), in its various formats as highlighted above, is mainly positive, but has also produced mixed results. Studies from HICs demonstrate that standard MI improves outcomes for non-injection drug users⁹; brief MI is effective for patients presenting with AUDs¹⁹; and MI-consistent interventions have been shown to improve treatment response²⁰ and are useful when used for co-occurring alcohol use and depression^21,22 and have been found effective in the treatment of first-episode psychosis and SUDs in young adults.²³ However, in a meta-analysis by Wang and colleagues,²⁴ the use of pure MI to treat psychosis and SUDs was found to be varied with relatively modest positive outcomes. Moyers and Houck²⁵ point out that MI combined with other treatments may at times be challenging, as the difference in approaches may be contradictory. They further state that combined treatment, although effective, may not always be seamless or without difficulty.²⁵

The evidence of the effectiveness of MI for SUD treatment in LMICs is substantially less. In a meta-analysis on psychosocial interventions (including three MI-informed and/or MI-consistent interventions) for reducing alcohol consumption in sub-Saharan Africa, the outcomes were mixed.²⁶ The meta-analyses concentrating on Alcohol Use Disorders Identification Test (AUDIT) scores discovered no statistically significant distinctions between the intervention and the comparison group at intervals of 2–3 months, 6 months, or 12 months following the intervention. However, reviewing trials related to alcohol abstinence revealed a positive impact of psychosocial interventions compared to the impact on the comparison group at the 3–6-month mark post-intervention, with sustained effect at longer-term follow-up periods (12–60 months).²⁶

Outcomes of intervention research in HICs may not yield similar results in LMICs even after being adapted to the local context in low-income populations.²⁷ Given the various challenges faced in LMICs, of utmost importance is to identify and implement brief, innovatively delivered, cost-effective solutions that empower providers in primary care settings to treat HSU effectively.¹¹ This review, therefore, aims to determine whether HSU interventions incorporating MI are effective with adults in LMICs. We further aim to highlight any limitations within the literature and offer opportunities for future research. This information will be of value to relevant stakeholders working in the field of addiction in LMICs to help make treatment and policy decisions within their given context.

Methods were pre-specified and documented in a protocol (PROSPERO–CRD42022329196). The inclusion and exclusion criteria are discussed in the next section.

Figure 1 shows the article selection process for this review, according to the PRISMA guidelines. The initial searches yielded 518 records, and 495 remained after removing duplicates. After screening titles and abstracts, 18 were eligible for full review. Full texts were then reviewed, and nine did not meet the inclusion criteria (one article focused on adolescents and not adults; one was a published protocol; one was an evaluation of a training programme; the study sample in one article was the partner and children of the substance-using patient rather than the patient himself; and all others were excluded because the study sites were not based in an LMIC). Examining the reference lists from the remaining seven articles identified four additional studies.^33,34,35,36 Thus, 11 studies were included in this review.^{33,34,35,36,37,38,39,40,41,42,43}

In this review, only one of the included studies used standard MI, not demonstrating significant reductions in substance use.³⁵ The remaining studies were MI-informed, and these MI-informed interventions accounted for significant reductions in substance use and related health outcomes observed in six of the 11 studies.^{33,35,37,38,39,42,43} These findings were consistent with results found in HICs.^{9,19,20,21,22,23} The potential reasons for the remaining four studies^34,36,40,41 yielding less favourable outcomes are various.

The study by Wandera et al.³⁶ reported mixed results, with the only favourable outcomes for brief MI found among female participants. The authors proposed that one reason for their results might be that studies reporting positive outcomes in similar populations included participants whose HIV status was unknown or who were HIV negative, coming from regions with low HIV prevalence. In contrast, their study focused solely on HIV-positive participants. Furthermore, trials involving HIV-positive participants yielded similar findings of no intervention effect.^44,45,46

In contrast to the brief single-session, MI-informed approach in this study, research using standard multi-session MI has demonstrated reductions in both the number of drinks consumed and in heavy drinking episodes. This finding suggests that multi-session MI, which aligns more closely with the original structure and phases of MI, may produce stronger behavioural outcomes than single-session adaptations do.^6,47 Sample size may have been another contributing factor. The relatively small sample sizes in each of the four studies may not have provided sufficient statistical power to demonstrate significant treatment differences between intervention and control groups, possibly yielding imprecise effect estimates.

Regarding the interventions themselves, Nadkarni and colleagues⁴¹ suggested that alcohol dependence might necessitate a more intensive psychosocial treatment. The CAP intervention utilised in their study embeds the core ‘spirit’ and some techniques of MI, including client centredness, motivational enhancement, feedback, readiness to change, empathy, and collaboration. The intervention also adds substantial cognitive-behavioural and relapse-prevention content. Therefore, conclusions about MI effectiveness should be qualified as the effects of CAP reflect outcomes of an MI-informed, hybrid intervention and not necessarily what a full and/or standard MI intervention may achieve. Nadkarni et al. (2019; p.114) further proposed that additional strategies may have improved outcomes, such as potentially involving discussions on barriers to care and treatment efficacy, among others.⁴¹ Sileo et al.,²⁶ in their meta-analysis of psychosocial interventions for alcohol use in sub-Saharan Africa, similarly argue that research on alcohol-focused interventions in this region would benefit from including more structural-level interventions such as the reduction of alcohol outlet density, enhanced alcohol regulation, and policy enforcement.

Wandera et al.³⁶ found reductions in alcohol consumption as measured by AUDIT-C scores in both the intervention and the control arm of their study, implying that both interventions contained partially effective components. The comparator in their study may therefore be considered an active intervention as participants received one-on-one counselling, which differs from most other studies that use patient information leaflets only in the control arm.⁴⁸ Prior research has shown that the simple act of participation and answering alcohol assessment questions in itself may lead to reduced reporting of drinking behaviour.⁴⁸ The authors argue that this may have distorted the measured intervention effects and could account for the lack of difference noted in their study.³⁶ Similarly, in the study by Madhombiro et al.,³⁴ the comparator was the World Health Organization Mental Health Action Programme Intervention Guide, which may also be considered an active intervention and may have skewed outcomes.

This review highlights that overall, MI-informed interventions demonstrated positive effects in improving substance use outcomes in the majority of the studies reviewed. Although the only study that utilised standard and/or pure MI did not demonstrate significant reductions in substance use, significant reductions were observed in several MI-informed studies. The reviewed studies employed a range of intervention strategies, including MI-blended with other interventions (such as CBT, BA, and imaginal desensitisation, and milieux therapy, among others) or using MI-informed approaches. Despite the heterogeneity, MI-informed interventions demonstrated potential effectiveness in reducing substance use. The studies were conducted in diverse settings, including primary healthcare centres, emergency centres, specialist clinics, hospitals, and Non-Governmental Organisation (NGOs). This result suggests that MI-informed interventions can be implemented effectively across various healthcare settings and different geographic locations.

The primary outcome measures varied across the studies, with the AUDIT being the most commonly used measure. While many studies reported significant reductions in substance use based on these measures, some studies showed nonsignificant differences between intervention and control groups. Future research should address the gaps highlighted above by clearly defining interventions, specifying interventionists, and evaluating the unique contribution of MI within combined programmes. Studies with larger sample sizes and more rigorous study designs are also necessary to strengthen the evidence base and address potential sources of bias. Nonetheless, MI-informed interventions can be considered a valuable approach for substance use interventions in LMICs within a range of treatment settings and available resources, complementing existing treatment modalities and promoting positive outcomes.