Introduction

Antibiotics, that is, antimicrobials targeting bacteria, represent one of the most significant medical breakthroughs of the 20th century and have revolutionised treatment approaches, saving countless lives from potentially life-threatening bacterial infections.1 However, the phenomenon of antimicrobial resistance (AMR), where bacteria, viruses, fungi and parasites become resistant to these medicines, poses a growing global health concern.2 3 This challenge is exacerbated by the excessive and inappropriate use of antimicrobial drugs, largely influenced by the unregulated sale and dispensing of antibiotics. The implications of AMR are far-reaching, restricting treatment choices and complicating the management of infections that were once easily treatable.3 4

The global trend in antibiotic consumption is on the rise, presenting a particularly alarming situation in LMICs, where the mortality burden from AMR surpasses that in high-income nations.5 6 Research published in The Lancet Planetary Health reveals a significant surge in antibiotic use across 204 countries between 2000 and 2018.7 Specifically, worldwide antibiotic consumption rates surged by 46% during this period, rising from 9.8 to 14.3 Defined Daily Doses (DDD) per 1000 population per day. In LMICs, a more pronounced increase of 76% was noted, with consumption expanding from 7.4 to 13.1 DDD per 1000 population per day.7

In this review, we focus on informal healthcare providers (IPs), a group that plays a significant role in the healthcare systems of many LMICs. IPs often serve as the first point of contact for healthcare-seeking individuals, delivering a substantial portion of primary care services—estimates range from 9% to 90%—particularly in underserved and impoverished communities.8–12 The term ‘informal healthcare providers’ encompasses a broad spectrum of practitioners, including drug sellers, village doctors, medicine vendors, traditional healers and traditional birth attendants, without a standardised definition. Common attributes used to identify these providers include the absence of formal qualifications, operation outside the formal healthcare system, and lack of registration or affiliation with any regulatory body.8 9 13

A critical aspect of IPs' practice is the widespread dispensing of antibiotics, a practice documented in various countries, including Bangladesh, India, Zambia and Congo, where a significant share of antibiotic dispensing is attributed to IPs.14–21 A comprehensive survey conducted through supplier mapping, customer exit interviews and household surveys across six countries—three in Africa (Mozambique, Ghana and South Africa) and three in Asia (Bangladesh, Vietnam and Thailand)—revealed that antibiotics were readily available in these countries through both formal and informal channels. Importantly, this study found that the practice of dispensing antibiotics without a prescription was notably prevalent,22 underscoring the critical role of the informal sector in the broader landscape of antibiotic dispensing.

Despite their lack of formal training and qualifications, IPs in many LMICs engage in antibiotic dispensing practices that pose substantial risks to individual patients and the broader health system.9 14–21 However, it is crucial to acknowledge the well-established position of IPs as healthcare providers within their communities. This recognition is based on several key factors, including the trust they garner, their proximity to and long-standing presence within the community—often spanning multiple generations—the affordability of their services, and their availability around the clock. Collectively, these factors establish IPs as the first and preferred providers for many patients in their care pathways.10 23–25 Therefore, given the integral role of IPs in primary care delivery within communities, it is imperative to prioritise these providers in both global and national antibiotic policies, research and interventions.

The issue of inappropriate antibiotic dispensing as well as usage is multifaceted, influenced by various factors across different levels (patient, provider and health system) in both the formal and informal healthcare sectors. Among patients, factors such as limited educational attainment, younger age, dissatisfaction with healthcare services and insufficient awareness about antibiotic usage contribute to inappropriate use. On the provider side, factors such as inappropriate drug prescribing and/or dispensing habits, inadequate knowledge and training, financial incentives, the influence of pharmaceutical representatives and unauthorised antibiotic sales are significant contributors.3 19 26–29 Furthermore, the lack or inadequacy of regulatory frameworks and their enforcement in many LMICs exacerbates the challenge, leading to the sale of antibiotics without prescriptions across various outlets, including local pharmacies, shops and clinics operated by IPs.16 30

IPs operate outside the formal health system’s purview, often resulting in their exclusion from national health policies. In many contexts, IPs are either overlooked, not officially recognised or even considered illegal by the health system. This situation complicates the documentation of their practices in national reports and surveys, thereby hindering the accurate assessment of antibiotic dispensing and its related practices within this group. Despite these challenges, IPs play a crucial role in primary care delivery, especially in LMICs, where they are known to dispense antibiotics. However, a comprehensive understanding of these practices and the factors influencing them remains elusive.

While numerous previous reviews have concentrated on formal healthcare providers,31 32 as far as we know, no reviews have specifically focused on IPs. Despite a growing body of research documenting antibiotic dispensing among IPs, there persists a notable gap in synthesising this evidence through a comprehensive review approach. Therefore, this review aims to map out the existing literature to offer a detailed overview of antibiotic dispensing and the factors influencing these practices among IPs within the context of LMICs.

Methods

The methodology for this review follows the Joanna Briggs Institute (JBI) guideline for conducting scoping review.33 The scoping review methodology is particularly well-suited for situations where the literature has not been extensively reviewed, and the goal is to offer an overview of a specific field.34 This approach aligns with the objectives of our review, which aims to scope the current state of knowledge and provide a summary of antibiotic dispensing and its determinants among IPs.

The JBI guideline for scoping review consists of five sequential stages, each of which is elaborated below.

Stage 1: research questions

The research questions for this review are as follows:

What is the extent of antibiotic dispensing among IPs? Given the diversity in methodologies used to assess antibiotic dispensing among IPs—ranging from prescription reviews and patient exit interviews to retrospective surveys of providers—this question aims to collate and present these varied reporting measures to offer a comprehensive descriptive overview of antibiotic dispensing at the level of IPs.

What are the patterns of antibiotic dispensing among IPs? This includes an exploration of the types of antibiotics most frequently dispensed, categorised according to the WHO’s AWaRe (Access, Watch and Reserve) system, where the ‘Access’ group primarily includes narrow-spectrum agents. Additionally, it seeks to investigate the conditions for which these antibiotics are dispensed and the routes of administration used, including any additional details on dispensing practices gleaned from the included studies.

What are the main factors driving antibiotic dispensing among IPs? This question aims to identify and categorise the factors influencing the dispensing of antibiotics among IPs. It distinguishes between intrinsic factors, which are internal to the providers themselves (such as their knowledge, attitudes and experiences), and extrinsic factors, which are external influences (such as patient demand, community expectations, and broader health system dynamics).

Stage 2: identification of relevant studies

The inclusion criteria for this review are developed according to the Population–Concept–Context framework.33

Population

As there is no standardised definition for IPs, we referred to our previously published review on this subject as well as other review papers focusing on IPs,8 9 13 to construct a working definition tailored for this review. We characterise IPs as providers who satisfy all the criteria outlined below:

Recognition/affiliation: healthcare providers who are not registered or affiliated with a government or a recognised private institution. This excludes community health workers (CHWs) integrated into the formal health system or affiliated with a non-government organisation (NGO). For example, the Building Resources Across Communities (BRAC) NGO in Bangladesh has trained and deployed more than 50 000 CHWs as part of their programme.35 This category also excludes alternative providers if they are trained in accredited institutions and are officially recognised by the local government. For example, the Ayurveda, Yoga & Naturopathy, Unani, Siddha and Homeopathy (AYUSH) system is recognised by the Government in India and, therefore, will not be considered in this review.36

Independent self-established practice: healthcare providers who have independently established their practice and offer health services in the community. Our previous review on a related subject has identified that practitioners engaged in spiritual, religious and traditional practices (such as traditional healers and traditional birth attendants) are also classified as IPs.8 However, as we are interested in antibiotic dispensing, IPs who base their practice on allopathic medicine (often referred to as untrained allopathic practitioners or village doctors) will only be included in this review. Therefore, the use of the term IPs in this review represents only this group of providers.

Appropriate qualification: healthcare providers who lack appropriate medical qualifications to prescribe or dispense allopathic medications, including antibiotics, as part of their accredited qualification (eg, providers often locally termed as ‘village doctors’, although they do not have a medical degree).

Concept

This review focuses on two main concepts: first, antibiotic dispensing and practices, and second, the drivers of antibiotic dispensing among IPs. To classify the types of antibiotics dispensed by IPs, we will employ the WHO’s AWaRe framework.37 This framework categorises antibiotics into three main groups: ‘Access’ denotes antibiotics that should be widely available at the primary care level and are recommended as the first-line treatment for common conditions; ‘Watch’ encompasses those with a higher risk of leading to resistance, and their use should be more restricted and monitored closely; and ‘Reserve’ refers to antibiotics that should be used only when there is no alternative owing to the pathogen’s susceptibility pattern, and should be confined to hospital settings. The drivers of antibiotic dispensing will be classified as intrinsic (provider-specific, such as age, training, education, knowledge) and extrinsic (external influences on the provider, such as patient demand, market availability, incentives) factors, based on lessons learnt from a previous review, which adopted a health system approach to understand drivers related to antibiotics dispensing.38

Context and design

The World Bank classification system will serve as the reference framework for categorising LMICs in this review. This system categorises countries into four income groups based on their Gross National Income per capita, expressed in US dollars. Our review will encompass countries classified as low-income, lower-middle-income and upper-middle-income. To accommodate the annual updates of the World Bank classification, we will use the study’s start year or the publication year (if the start year is not specified) as the basis for determining a country’s LMIC status. Given this consideration, the term 'LMIC' will not be applied as a search filter during the initial database search. Instead, the classification of countries as LMICs will be verified during the study selection process, referencing the World Bank’s historical data for income classification.

This review will impose no restrictions regarding study design, methodology, publication year or language. However, grey literature will not be considered in this review. In our study, we will use translation software, such as Google Translate or DeepL Translate, to process articles written in languages other than English. We acknowledge that while machine translations may not always yield perfect results, they provide a feasible means to access and include non-English studies, thereby reducing language bias in our review, which is also supported by existing research.39 Specific exclusions will apply to review papers, editorials, commentaries, study protocols, conference abstracts, perspective pieces, and multiple publications from a single study.

Search strategy

The literature search for this review will be conducted by an experienced research librarian across nine electronic databases: MEDLINE, EMBASE, SCOPUS, Global Health, CINAHL, Web of Science, LILACS, African Journals Online (AJOL) and Index Medicus for South-East Asia Region (IMSEAR). Two key concepts will be used in the literature search: ‘Antibiotics’, ‘Informal Healthcare Providers’. For an extensive list of key search terms associated with each of these concepts, please see table 1, which was compiled by referencing previously published reviews on topics related to IPs,8 9 13 as well as from a preliminary search of MEDLINE. A complete search strategy for MEDLINE (via Ovid) is available as a online supplemental file S1. Additionally, hand-searching of references from studies selected for full-text review will be conducted to identify additional relevant literature.

Table 1

Key search terms for this review

Stage 3: study selection

All records identified through the search will be imported into Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia), where duplicate entries will be automatically detected and removed. The study selection process will consist of two main steps:

Title and abstract screening: four reviewers (DT, SB, ST and AB) will independently screen titles and abstracts against predetermined inclusion criteria. To refine and streamline the screening process, a pilot test will be conducted using 15–20 studies. Each record will undergo independent review by two authors to ensure a thorough and unbiased selection process. Any record deemed potentially relevant by at least one reviewer will automatically advance to the full-text review stage, preventing the premature exclusion of potentially eligible studies.

Full-text review: for records identified as potentially relevant during the initial screening, the same four reviewers (DT, SB, ST and AB) will independently assess the full texts to determine final eligibility. Each full-text article will undergo independent review by two authors. Similar to stage 1, a pilot test with 10–15 studies will be conducted to refine and validate the selection criteria. Reasons for exclusion at this stage will be documented and reported. Discrepancies in the selection process will be resolved through discussion involving a third author (PT or MT) when necessary to achieve consensus. The study selection process will be documented and presented using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews flow diagram.

Assessment of methodological quality

The methodological quality of all included studies will be evaluated using the JBI critical appraisal tools.40 It is important to note that the primary purpose of this quality assessment is to gain insight into the methodological strengths and limitations of the studies, rather than to serve as a criterion for inclusion in the review. The quality assessment will involve three reviewers (PT, BB and SS), with each study being independently appraised by two authors. Any disagreements during this process will be resolved through discussion until a consensus is reached.

Stage 4: charting the data

A data extraction form will be developed and preliminarily tested on 5% of the included articles by two coauthors (PT and MT) using Covidence. Subsequently, data extraction will be conducted by four coauthors (DT, SB, ST and AB), such that each paper is assigned to two independent extractors. Key data points to be captured include study objectives, design, timing and location, sampling strategy, data sources, method used to assess antibiotics dispensing, provider characteristics, details on dispensed antibiotics, dispensing practices (including types of antibiotics, indications, route of administration and patient characteristics if available), as well as intrinsic and extrinsic factors influencing antibiotic dispensing.

Stage 5: collating, summarising and reporting the results

Essential information from the included studies, such as year of publication, research design, sample size, study population demographics, geographic location, and related outcomes, will be systematically displayed in a table.

For the first and second objectives, we will provide a descriptive summary of the extent of antibiotic dispensing and its patterns among IPs. We will compile dispensing estimates and patterns as reported in the included studies. Given the diverse methodologies applied in these studies and the lack of a standardised system for reporting antibiotic dispensing by IPs, our review will refrain from aggregating data into quantitative metrics, such as combined prevalence proportions. The findings will be synthesised and reported using descriptive tables, visuals and a narrative summary, as recommended for scoping reviews.41 42 This approach will offer a comprehensive overview of antibiotics dispensing by IPs, reflecting the varied contexts, methods, and reporting practices documented in the literature.

For the third objective, which aims to elucidate the factors influencing antibiotic dispensing among IPs, we will employ both an inductive and deductive content analysis method.41 First, the data will be deductively coded according to the framework outlined by Rodrigues et al,38 classifying reported factors as either intrinsic (relating to the providers themselves) or extrinsic (external influences on providers). Second, any factor not captured by this framework will be discussed among the authors to determine whether they should be merged into existing categories or if a new category should be established. This comprehensive method allows for an in-depth exploration of the complex factors influencing antibiotic dispensing practices among IPs.

The overall proposed tentative timeline for the various stages and completion of this review is detailed in the online supplemental file S2, which provides an understanding of the project timeline.