Background

The World Health Organization (WHO) describes self-medication as taking medications to address illnesses that someone has diagnosed without a doctor’s advice or supervision [1]. Self-medication with antibiotics in particular is a common practice worldwide [2]. In recent times, there has been consistent documentation of increasing rates of antibiotic self-medication worldwide [3]. Consequently, antibiotic self-medication has emerged as a major public health concern, garnering significant attention from researchers in the field of public health [4].

The use of antibiotics for self-identified illnesses without first seeing a trained healthcare provider is known as antibiotic self-medication [5]. This may result in the overuse of antibiotics, as well as other issues such as masking underlying symptoms, postponing or providing a false diagnosis, causing drug interactions, and hastening the development and dissemination of antibiotic resistance (1, 6–7). A larger portion of antibiotic misuse and self-medication is observed in developing countries [8]. Research indicates a higher prevalence of antibiotics misuse and self-medication in developing nations when compared to developed ones [9].

About 80% of antibiotics are thought to be utilized in communities outside recognized healthcare facilities in Low and Middle-Income Countries (LMIC), of which 20–50% are misused [10]. The waste of financial resources from extended treatment periods brought on by improper infection control and unpleasant effects are additional problems associated with antibiotic self-medication [5]. The increasing pandemic of antibiotic resistance has primarily affected Africa [11, 12]. More than half of the antibiotics used in communities, particularly in Africa, were reportedly sold without a prescription in 2011 [13].

Concern over Ethiopians self-medicating with antibiotics has grown recently [14]. Antibiotic self-medication was linked to easy availability of antibiotics without a prescription [15], a lack of knowledge regarding antibiotic resistance [16] and socioeconomic status [16, 17]. Antibiotic self-medication practices may also be influenced by poor healthcare infrastructure and restricted access to healthcare services in rural areas [18].

Ethiopia is known to have a significant burden of infectious diseases, including a high incidence of disease morbidity and mortality. This is likely because of increased rates of antimicrobial resistance (AMR). Additionally, there are indications that people, healthcare professionals, and society as a whole are using antibiotics excessively [19, 20]. The country has been putting numerous initiatives into practice to address the issue, including the responsible use of antibiotics, disease prevention and control, public surveillance suggesting the use of antibiotics, continuous guidelines, and enforcement. Nevertheless, a national study on the scope of antibiotic misuse was not conducted. Therefore, this review aimed to assess the prevalence and associated factors of antibiotic self-medication in Ethiopia.

Methods

Protocol and registration

The review protocol was developed and registered in International prospective register of systematic reviews with registration number CDR42023439111 and available at https://www.crd.york.ac.uk/prospero/#recordDetails. We followed the recommendation of PRISMA statement [21] to report this systematic review and meta-analysis [Supplementary file 1].

Eligibility criteria

All published research on the prevalence of antibiotic self-medication in Ethiopia and its predictors was included in this systematic review. The study covered all cross-sectional observational quantitative studies that were published in English and carried out in households in Ethiopia. Dissertations and masters theses that have not been published were not included. Every study that wasn’t observational was disregarded. Qualitative or mixed method studies were excluded. Non-human studies and conference abstracts were also not included in the review.

Information sources

Our research question focused on repeated database searches to find all the studies that met our inclusion criteria. In systematic reviews, it has been demonstrated that searching for multiple databases yields better results than searching for only one [22]. To find more research that might be included, the references of the identified studies were evaluated.

The inclusion rates of systematic reviews are increased when multiple databases are searched and references are verified [23]. From 2000 to 2024, pertinent research was looked for in the databases of MEDLINE (PubMed), Scopus, Google Scholar, and Web of Science. Additional possible resources, such as conference proceedings and books with abstracts, were also looked up.

Search strategy

Using the PRISMA guidelines [21], an electronic systematic search was conducted on MEDLINE (PubMed), Scopus, Google Scholar, and Web of Science. Both index/subject terms and keywords were employed to expand the search approach. These phrases were combined using boolean operators (“OR,” “AND”) to create a search strategy. The search terms were “prevalence,” “proportion,” “magnitude,” “epidemiology,” “associated factors,” or “determinants,” as well as “antibiotic self-medication” or “self-prescription” or “Non-prescribed use of antibiotics” and “Ethiopia,” and the full syntax used for database search was ((((((((((‘Prevalence’[Mesh]) OR ‘proportion’ [Mesh]) OR ‘magnitude’ [Mesh]) OR ‘epidemiology’ [Mesh]) AND ‘associated factors’ [Mesh]) OR ‘determinants’ [Mesh]) AND ‘antibiotic self-medication’ [Mesh]) or ‘self-prescription’ [Mesh]) OR ‘Non-prescribed use of antibiotics’ [Mesh]) AND ‘Ethiopia’ [Mesh]) The Medical Subject Headings (MeSH) were employed in PubMed to align synonymous phrases. A preliminary scoping search was conducted on PROSPERO to make sure no previous review of a similar nature had been registered. The search was conducted from 01/05/2023 to 30/05/2023.

Study selection

The review was designed by WA, AT, and EA. Independent assessors (WA, AT, EA, TA, YK, WS, and LW) select the study and extract the data. Titles and abstracts were independently evaluated by reviewers (WA, AT, EA, TA, YK, WS, and LW) and vetted against the qualifying criteria. Discrepancies were settled by SZ. The entire texts of the publications were retrieved for quality evaluation after the titles and/or abstracts were changed for potential inclusion.

Data abstraction

The investigators created an Excel data extraction form. Subsequently, this form was used to extract and gather pertinent data. Authors’ names, publication years, regions, study designs, study settings, study participants, sampling methods, sample sizes, response rates, recall periods in months, prevalence (%) with 95% CI, factors associated with self-medication antibiotics, common antibiotics used in self-medication, source of antibiotics, perceived illnesses/symptoms for which antibiotics are used, and reasons for using antibiotics for self-medication are all included in the form. The review’s key outcome, or summary measure, is the prevalence of antibiotic self-medication and its determinants. The most common antibiotics used for self-medication, their source, the ailments or symptoms that people believe warrant their usage, and the motivation for their use were secondary outcomes of this review and meta-analysis.

Assessment of the quality of included studies

The checklist for assessment of bias in systematic review of prevalence studies, created by Damian Hoy in 2012, was used to evaluate the characteristics of the included research. The Hoy checklist is the most popular method for evaluating bias in systematic reviews of prevalence studies. It is ten items total, split into two sections of the checklist. Six components evaluate internal validity (items 5 to 9 evaluate the domain of measurement bias, and item 10 evaluates bias related to the analysis). Four components (items 1–4) evaluate external validity (domains are selection and non-response bias). The total score of 0–4 was regarded as low quality, the total score 5–7 regarded as moderate quality and total score of 8–10 regarded as high quality [24].

Data analysis

The data collected using the data abstraction format in Excel was exported to and analyzed using STATA version 17 statistical software. Data was presented quantitatively and in narrative form. DerSimonian-Laired random effect was performed to estimate the pooled prevalence of antibiotic self-medication in Ethiopia. Cochrane’s Q statistics, I2 and P values were used to check the heterogeneity of the studies. Meta regression analysis, subgroup and sensitivity analysis were performed in order to explain the cause of heterogeneity. The result was presented in a forest plot. The presence of publication bias was presented with a funnel plot.

Results

About 73 articles were identified from PubMed, 151 from Scopus, 246 from Google Scholar, and 214 from Web of Science. 173 articles were duplicates and 511 articles were left for screening the titles and abstracts. About 496 articles were excluded. Then, 15 articles were left for further full text review. From these, 4 articles were excluded after reviewing of the full texts. Finally, 11 articles were eligible for the systematic review and meta-analysis (Fig. 1).

Fig. 1
figure 1

PRISMA flow diagram

Full size image

Characteristics of included studies

All eleven studies selected for this review and meta-analysis were crossectional studies published in English from 2012 to 2024. About 5814 subjects were involved for the study. The samples were drawn using varied sampling methods from the general public (8/10 studies), undergraduate university students (1/10 studies), health professionals (physicians, pharmacists and nurses) (1/10 studies) from different settings such as households, university students, hospitals and drug retail outlets. The recall periods were varied among different studies, which range from 1 month to 12 months (Table 1 and 2).

Table 1 Characteristics of included studies
Full size table

Sampling method

Table 2 Sampling method used by included studies
Full size table

Quality assessment of included studies

Eleven studies were assessed for risk of bias. All studies showed a low-level risk of bias (Table 3).

External validity

  1. 1.

    Was the study’s target population a close representation of the national population in relation to relevant variables?

  2. 2.

    Was the sampling frame a true or close representation of the target population?

  3. 3.

    Was some form of random selection used to select the sample, OR was a census undertaken?

  4. 4.

    Was the likelihood of nonresponse bias minimal?

Internal validity

  1. 5.

    Were data collected directly from the subjects (as opposed to a proxy)?

  2. 6.

    Was an acceptable case definition used in the study?

  3. 7.

    Was the study instrument that measured the parameter of interest shown to have validity and reliability?

  4. 8.

    Was the same mode of data collection used for all subjects?

  5. 9.

    Was the length of the shortest prevalence period for the parameter of interest appropriate?

  6. 10.

    Were the numerator(s) and denominator(s) for the parameter of interest appropriate?

Table 3 Quality assessment of included studies
Full size table

Prevalence of antibiotic self-medication

Nine studies were eligible for meta-analysis. The overall prevalence of antibiotic self-medication in this study is 46.14 [35.71, 56.57]. The prevalence varied across regions which ranged from 18.0% to 0 67.3% (Fig. 2).

Fig. 2
figure 2

A summary of forest plot that showed the overall pooled prevalence of antibiotic self-medication in Ethiopia

Full size image

Subgroup and Meta regression analysis

The selected studies exhibited significant heterogeneity (I2 = 98.99%). This suggests that the inconsistency among studies was greater than what would occur randomly, resulting in an inconsistent overall estimate of the proportion of antibiotic self-medication. This was taken into account while estimating the over prevalence of antibiotic self-medication using a random effect model. Meta regression analysis was evaluated along with a subgroup analysis in order to explain the cause of heterogeneity. Sample size and response rate were used in the Meta regression analysis and none of them were significant and did not explain the source of heterogeneity (Table 4).

Table 4 Meta regression analysis of the studies based on sample size and response rate
Full size table

Subgroup analyzes were carried out based on region and study setting. The analysis showed that the pooled prevalence of antibiotic self-medication is almost similar to the pooled prevalence in Amhara, Tigray and Oromia regions, where studies in Addis Ababa and Sidama are higher and lower than the pooled prevalence of antibiotic self-medication respectively.

Subgroup analysis based on a study setting showed that the pooled prevalence of antibiotic self-medication in household, university and pharmacy retail is similar to the pooled prevalence of antibiotic self-medication. However, heterogeneity obviously not decreased. In addition, a Galbraith plot was drawn to identify some studies that were obviously different from others. But the Galbraith plot showed the absence of substantial heterogeneity since all studies lie within the 95% CI region (shaded area) (Fig. 3).

Fig. 3
figure 3

Galbraith plot

Full size image

Sensitivity analysis

To investigate the impact of each individual study on the pooled prevalence of antibiotic self-medication, a leave-one-out meta-analysis was conducted. When each study was removed from the analysis, the pooled estimate prevalence of antibiotic self-medication fell between the confidence interval of the pooled estimated prevalence of antibiotic self-medication, indicating that no single study could affect the statistically significant difference (Fig. 4).

Fig. 4
figure 4

Leave-one-out sensitivity analysis

Full size image

Publication bias

Nonparametric trim-and-fill analysis of publication bias was performed using funnel plot to confirm the evidence of publication bias. Despite the asymmetry of the funnel plot, Egger and Begg’s tests revealed that publication bias was not statistically significant (P values of 0.2621 and 0.3481 respectively) (Fig. 5).

Fig. 5
figure 5

Funnel plot

Full size image

Common antibiotics used for self-medication

Eight different classes of antibiotics were self-medicated by study participants and the antibiotics commonly used in self-medication include penicillins (10 studies), tetracyclines (6 studies), fluoroquinolones (5 studies), Cephalosporin (2 studies), sulphonamides (1 study), macrolides (1 study), Chloramphenicol (1 study) and aminoglycosides (1 study). The most frequently used classes of self-medicated antibiotics were penicillins followed by tetracyclines (Table 5).

Table 5 Common antibiotics used for self-medication
Full size table

Source of antibiotics

Studies reported that participants were obtained information from various sources for antibiotics used in self-medication in Ethiopia. These include community pharmacies (8/10 studies), family/ relatives/ friends/neighbors (7/10 studies), leftovers from previous treatment(4/10 studies), patent medicine stores (3/10 studies), hospital pharmacies (2/10 studies), health workers such as doctors, nurse (2/10 studies), private health facilities (1/10 studies), sample from medical representatives (1/10 study), by sharing with the others (1/10 study), and kiosks (1/10 study) (Table 6).

Table 6 Source of antibiotics
Full size table

Perceived illnesses/symptoms for which antibiotics are used for self-medication

Four studies reported the perceived illnesses/symptoms for which antibiotics were used for self-medication by study participants. The common indications reported for use for antibiotic self-medication include upper respiratory tract infection (URTI), gastro intestinal problems, common febrile illness, body aches, skin problems, urinary tract problems (Table 7).

Table 7 Perceived illnesses/symptoms for which antibiotics are used for self-medication
Full size table

Reason for which antibiotics are used for self-medication

Seven studies reported the reason for which antibiotics were used for self-medication. The most common reported reasons of antibiotic self-medication include previous experience, to save cost, lack of time and avoiding waiting time (Table 8).

Table 8 Reason for which antibiotics are used for self-medication
Full size table

Factors associated with self-medication antibiotics

All of the studies reported the associated factors with antibiotics self-medication despite differences in factors across studies. Low educational level, age (18–34 years) and gender i.e. being male were common significantly associated factors reported and considered as factors for antibiotic self-medication practice in Ethiopia. Low educational level was the most commonly reported factor associated with self-medication antibiotics (Table 9).

Table 9 Factors associated with antibiotic self-medication
Full size table

Discussion

The prevalence of antibiotic self-medication is a concern in Ethiopia based on the meta-analysis, indicating a high overall rate of 46.14%. The use of antibiotics without a prescription occurs despite their prescription being only legal status in most countries [13]. This self-medication use of antibiotics contributes to accelerating the emergence and spread of antimicrobial resistance (AMR) [1, 36, 37].

Variations across regions from 18.0 to 67.3% suggest differing cultural or healthcare factors influencing this behavior. Numerous studies corroborate this trend. For instance, the prevalence of antibiotic self-medication in Iran was found 53.3% [38], 20–25% in Europe [34], 48.8% in Africa [39]. These discrepancies could be attributed to variations in healthcare accessibility, education, regulatory policies, and cultural beliefs regarding antibiotics. It underscores the global significance of addressing this issue to combat antibiotic resistance.

The overall pooled prevalence in our study is found to be higher than that reported in systematic reviews from South East Asia [40] and the WHO Eastern Mediterranean Region [41]. Poor regulation of antibiotic sales resulting from the absence of policies or laxity in law enforcement makes antibiotics easily available for self-medication [13].

The classes of antibiotics most commonly self-medicated by study participants were penicillin followed by tetracyclines. It is consistent with other systematic reviews reported by the Middle East [42] and Europe [43]. It also aligns with the general knowledge that penicillins are widely used due to their efficacy against a broad range of infections, while tetracyclines are often chosen for their effectiveness against various bacterial illnesses. The varying use across different antibiotic classes might reflect regional availability, familiarity, or perceived effectiveness by users.

Multiple studies have also reported similar trends in the classes of antibiotics used in self-medication. For instance, a study in Saudi Arabia [44] found penicillins to be commonly self-medicated, consistent with our data. Another study in Nigeria [45] observed tetracyclines were among the most frequently self-administered antibiotics. Additionally, the WHO report on antibiotic use highlighted the widespread misuse of penicillins and tetracyclines globally. These studies collectively echo the prevalence of penicillins and tetracyclines in self-medication practices, suggesting a recurring pattern across various regions in the choice of these antibiotic classes (47).

In our study, the most common sources for antibiotics used in self-medication in Ethiopia were community pharmacies. Studies conducted in the Euro-Mediterranean region and developing countries have been reported that pharmacists were the main source of information for SMA (41, 48).

Several studies worldwide have also highlighted similar sources for obtaining antibiotics for self-medication. For instance, a study in Nigeria found community pharmacies and friends/relatives as common sources for self-medicated antibiotics [45]. Moreover, a study in Palestine noted community pharmacies and leftover medications from previous treatments among the primary sources for self-medication with antibiotics (49). Similarly, a study across various European countries identified community pharmacies and obtaining antibiotics from acquaintances as frequent sources for self-medication [34].

These studies emphasizing the role of community pharmacies in facilitating antibiotic self-medication practice, which could contribute to antibiotic misuse and resistance. This shows that community pharmacists are responsible for the extensive antibiotic misuse in the community. Therefore, the laws and regulations the country has should be strongly implemented in the community pharmacies. Because lax regulations or enforcement might allow pharmacies to dispense antibiotics without proper prescriptions, contributing to their frequent use as sources for self-medication.

The reported indications for antibiotic self-medication in this study align with commonly perceived illnesses/symptoms worldwide (Upper Respiratory Tract Infections, Gastrointestinal Problems, Febrile Illnesses, Body Aches, Skin Problems and Urinary Tract Problems). Studies conducted globally corroborate these findings and suggest a consistent pattern where individuals tend to self-medicate with antibiotics for similar perceived illnesses/symptoms across different regions, emphasizing the need for targeted education on appropriate antibiotic use [13, 16, 34, 40].

The reasons behind antibiotic self-medication, including previous experience, cost-saving, time constraints, and avoiding waiting times, align with findings from various studies conducted globally in Iran [38], Saudi Arabia [44], Nigeria [45] and across European countries [34], and Palestine (49). These reasons are recurrent across different regions, indicating common motivations for individuals resorting to self-medication with antibiotics, underscoring the need for improved access to healthcare and education on appropriate antibiotic use.

In the current study, low educational level, age (18–34 years) and gender i.e. being male were, significantly, the most common reported factors for antibiotic self-medication practice in Ethiopia. Low educational level was the most commonly reported factor associated with self-medication antibiotics. This shows the need for promoting literacy among communities and sensitization of the public as a vital strategy to also reduce antibiotic self-medication. Illiteracy is a driver of antibiotic self-medication as individuals and entire communities have less opportunity to be aware of the health risks associated with antibiotic self-medication (50). Special attention should be given to educating the public and healthcare providers on drugs used for self-medication and their impact on the development of antimicrobial resistance should be provided by the community.

This review and meta-analysis has certain limitations. Studies have been concentrated in certain regions, limiting the generalizability of findings to the entire country. Variations in study methodologies and populations could introduce heterogeneity affecting the pooled prevalence.

Conclusions

Antibiotic self-medication is a substantial issue in Ethiopia, with almost half the population engaging in this practice. A prevalence rate of 46.14% indicates a significant public health concern. It is considered high when compared to similar studies conducted in other countries or regions. The World Health Organization (WHO) discourages self-medication with antibiotics due to the risks associated with incorrect usage, such as antibiotic resistance. Any prevalence rate above zero indicates a potential concern, but a rate of 46.14% is particularly high relative to the WHO’s recommendation. Penicillins and tetracyclines were frequently self-medicated. Community pharmacies were a major source, and reasons included past experiences, cost-saving, lack of time, and avoiding waiting times. Lower education levels were the major determinant of antibiotic self-medication.

Recommendations

A targeted interventions such as educating people about the risks associated with using antibiotics without medical guidance which results in reduction in antibiotic resistance is needed. This review and meta-analysis exhibited significant clinical heterogeneity among the studies included, thus it should be considered with caution “Abbreviations.

AMR: Antimicrobial resistance; EFDA: Ethiopian Food and Drug Authority; LMIC: Low and Middle-Income Countries; PRISMA: Preferred Reporting Items for Systematic Review and Meta-Analysis; WHO: World Health Organization.