English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
INTRODUCTION
Self-medication involves using medicinal products to treat self-recognized disorders or symptoms or the intermittent or continued use of a medication prescribed by a physician for chronic or recurring diseases or symptoms [1]. According to the World Health Organization, responsible self-medication is carried out with over-the-counter (OTC) drugs, using medications for the intended indications, and reading and following the manufacturer's instructions on the label and package insert [2]. Unresponsible self-medication can result in health risks, such as delayed diagnosis and treatment of disease, adverse drug reactions, drug-drug interactions, and antimicrobial resistance [3],[4].
Although most of the world population self-medicates or has done so at some point, the frequency of self-medication and its associated factors varies depending on the population studied. Indeed, a recent global meta-analysis of 69 articles published between 2000 and 2018, including a total of 27,890 individuals, estimated a prevalence of self-medication of 55.9% (95% CI 42.4%-68.5%) in Africa; 71.2% (95% CI 63.0%-78.3%) in Asia; 74.0% (95% CI 56.2%-86.4%) in Europe; and 60.0% (95% CI 40.2%-77.0%) in South America [5]. In Colombia, different studies about self-medication have been conducted and reported frequencies ranging from 32.3% [6] to 84.7% [7], depending on the study sample and time frame used. However, there are no systematic reviews, nor is there a global quantitative estimate (meta-analysis) of its prevalence in the Colombian population. Therefore, this systematic review and meta-analysis aimed to estimate the pooled prevalence of self-medication and its associated factors in young people and adults in Colombia.
METHODS
Comprehensive systematic literature search
Search strategy
A comprehensive systematic literature review was conducted and reported according to the standards set out in the PRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analyses[8] and guided by the overview of Munn and colleagues [9]. The search was performed to identify studies published from January 2000 – June 2022 in English, Spanish, and Portuguese. The search was conducted in multiple electronic bibliographic databases: MEDLINE, PubMed, Academic Search, Web of Science (including Science Citation Index, Social Sciences Citation Index, and Arts and Humanities Citation Index), LILACS, the catalog of the Universidad Nacional de Colombia (UNAL), BIBLAT, and Google Scholar. Multiple combinations of the following keywords were used: cross-sectional study, statistics and numerical data, prevalence, multivariate analysis, bivariate analysis, and Colombia. In addition, the citations in the relevant articles were manually screened.
Study selection
The inclusion criteria were: (i) primary and secondary sources studies; (ii) descriptive studies (cross-sectional, cohort, case-control, and prospective), carried out in Colombia and reporting the prevalence of self-medication in young people and adults, as well as its associated factors; and (iii) articles, theses, and book chapters. Studies were excluded if: (i) they were review articles, letters to the editor, editorials, comments, expert opinions, case studies, case series, or conference abstracts, and (ii) they were published in languages other than English, Spanish, or Portuguese. The selection of the studies was carried out by two investigators independently, and discrepancies were resolved by consensus.
Quality assessment of studies
Using a standardized form, each study's bias risk was assessed using Hoy et al.'s criteria [10]. Each criterion was assessed as 0 if the risk of bias was low; 0.5 if the risk was intermediate or unclear; and 1 if the risk was high. The sum of the scores for the ten criteria indicated the overall risk of bias of the study, was defined as criterion 11, and was used to determine whether a study was included in the analysis. Studies with scores of 0-3 were considered low risk and were included in the quantitative analysis; studies with scores of 4-6 had moderate risk and were excluded, as well as those with scores of 7-10, which had a high risk of bias. This process of assessing the quality and risk of bias of the studies was carried out independently by two investigators, and when discrepancies were found, the final scoring and the decision to include or exclude the study was made by consensus. The summary figure of the studies' risk of bias was prepared using the RevMan v5.0 package.
Data extraction
Using a standardized format in Microsoft Excel, two authors independently extracted data on study characteristics (first author, publication date, characteristics) and outcomes of interest (sample size, self-medication time frame, prevalence of self-medication, women who self-medicated, men who self-medicated, lack of time and delays in medical care as causes of self-medication, and use of analgesics, anti-inflammatories, antihistamines, antibiotics and antiparasitics). When the prevalence of self-medication was evaluated in several time frames in the same study, the data were extracted separately.
Data processing and meta-analysis
Relevant data extracted from the included studies in the Microsoft Excel format were exported to the statistical package Stata v17. A random-effects meta-analysis model was used to estimate DerSimonian and Laird's pooled effect to show heterogeneity. Subgroup analysis was conducted to adjust random variation between point estimates of the original studies and investigate how the prevalence of self-medication fluctuates across subgroups depending on the self-medication time frame. The outliers within the included articles were checked using sensitivity analysis. Publication bias across studies was assessed using a funnel plot, Egger's regression test, and Begg's test (Z statistic) at 0.05. A forest plot format presented the point prevalence (percent) and 95% confidence intervals (CIs). In this plot, each crossed line referred to a 95% CI. The odds ratio (OR) was used to determine the association between self-medication and sex, and point prevalence (percent) was used for causes of self-medication and medications groups.
RESULTS
Search results
Initially, the search yielded 31 studies. After removing one duplicate, 30 studies were screened using titles and abstracts, 24 of which (16 articles and eight theses) underwent methodological evaluation. Eleven studies with moderate and high risk of bias were excluded, and the reasons for their exclusion are presented in Supplementary Table 1. Ultimately, 13 were included for the qualitative and quantitative synthesis: 12 about general self-medication with 5,668 participants were used for the pooled prevalence estimation, and one about self-medication with antibiotics and 140 participants was used only when analyzing medication groups (Figure 1).
Study characteristics
Table 1 shows that 12 studies with 5,668 participants were included in the systematic review and meta-analysis of self-medication prevalence. One additional study with 140 participants was included to analyze antibiotic and antiparasitic use through self-medication [11]. All the included studies had a cross-sectional design and were published between 2002 and 2021. All the included studies were conducted in urban areas and used randomized sampling, except one, whose sample selection was by convenience [12]. The sample sizes ranged from 140 in Cali [11] to 1,127 in Pereira [6]. The prevalence of self-medication varied from 32.3% in Pereira [6] to 84.7% in Neiva [7]. Regarding geographical distribution, five studies were conducted in the Caribbean region [12]-[16], three in the Andean region [6],[7],[17], four in the capital city Bogotá [18]-[21], and one in the Pacific region [11]. There were no studies from the Amazonia and Orinoco regions. Nine studies included adults from the general population [6],[11]-[15],[17]-[19], and four included university students [7],[16],[20],[21]. Six studies used instruments with no known validity or reliability [7],[16]-[19],[21], and seven used instruments with partially documented validity and reliability [6],[11]-[15],[20]. The overall risk of bias score of the included studies was between 0.5/10 and 2.5/10 (Supplementary Figure 1).
Each study used a different self-medication time frame (Table 1), so they were categorized into three subgroups. Five studies with 2,344 participants were included in the subgroup that evaluated current self-medication in the last 30 days. Five studies with 2,355 participants were included in the subgroup that evaluated self-medication in the previous year and throughout life. Four studies with 1,383 participants were included in the subgroup that evaluated self-medication in university students during the last six months and throughout life. Two studies assessing self-medication in various time frames [12],[17] were simultaneously included in subgroups 1 and 2, and their data were extracted separately for meta-analysis. One study evaluated self-medication with antibiotics and was included only in the analysis by medication groups [11].
Meta-analysis
Self-medication prevalence
The pooled prevalence of self-medication among the Colombian youth and adult population was 62.4%, with a 95% CI from 47.2 to 77.7%. As the I2 statistic revealed, there was a high degree of heterogeneity across studies (I2= 99.6%, p<0.001). The random effects model was assumed for this meta-analysis.
Subgroup and sensitivity analysis
No significant change in heterogeneity was observed when trying to exclude outliers, or one or more studies. For this reason, all the studies were included in the meta-analysis, and a subgroup analysis was performed according to the self-medication time frame and the type of population (general population or university students). The subgroup analysis revealed a lower frequency of self-medication among the general population when evaluated during the last 30 days (32.3%; 95% CI 25.4%-39.3%). In contrast, the prevalence of self-medication was significantly higher when assessed in the general population during the last year or lifetime (74.6%; 95% CI 61.3%-87.9%) and in university students in the previous six months or lifetime (84.7%; 95% CI 75.1%-94.3%), with no significant difference between them (Figure 2). Therefore, the effect of the time frame during which the self-medication behavior is evaluated seems more important than the effect of age or the university student status. We did not find evidence of publication bias (Egger's regression test, one-tailed, p= 0.2924; Begg's correlation test, one-tailed, p= 0.3811).
Prevalence of self-medication according to sex
Data on the frequency of self-medication in women and men were extracted from 10 studies. In the meta-analysis, no significant association was observed between sex and self-medication, with a pooled OR of 1.12 (95% CI 0.87-1.44). Moreover, the subgroup analysis showed no association between self-medication and sex among the general population in the last 30 days (OR= 0.87; 95% CI 0.57-1.32); nor in the previous year or at some point in life (OR= 1.11; 95% CI 0.61-2.03). In contrast, female university students had a greater probability of self-medication in the last six months or throughout their lives than males (OR= 1.72; 95% CI 1.17-2.53) (Figure 3). We did not find evidence of publication bias (Egger's regression test, one-tailed, p= 0.7983; Begg's correlation test, one-tailed, p= 1.00).
Determinants of self-medication
Due to the lack of a standardized and validated instrument to evaluate self-medication, each study addressed different determinants for this phenomenon. However, among these, the most recurring determinants reported were the lack of time and delays in medical care. Data on these causes could only be extracted from eight studies. In the meta-analysis, the lack of time and delays in medical care were reported with an overall frequency of 35.2% (95% CI 25.6%-44.7%). Moreover, the subgroup meta-analysis showed that the frequency with which these causes were reported by university students who self-medicated in the last six months or throughout their lives was significantly lower (19.7%; 95% CI 10.3-29.2) than that of the general population who self-medicated in the last 30 days (47.8%; 95% CI 31.5-64.2) (Figure 4).
Figure 2 Subgroup analysis of studies describing the prevalence of self-medication disaggregated by time frame and the type of population (general population or university students).
Figure 3 The association between self-medication and sex, disaggregated by time frame and the type of population (general population or university students), with results presented as the nepe- rian logarithm of the OR.
Figure 4 The frequency of reporting the lack of time and delays in care as causes of self-medication.
Medication groups used for self-medication. Data on the medication groups used for self-medication could be extracted from nine studies. The meta-analysis showed that the pooled frequency of self-medication with analgesics was 37.7% (95% CI 15.8%-47.6%), anti-inflammatories 33.2% (95% CI 15.4%-51.0%), and antihistamines 14.8% (95% CI 6.8%-23.0%), while the pooled frequency for antibiotics and antiparasitics was 12.1% (95% CI 7.4%-16.9%) (Table 2).
DISCUSSION
A systematic review and meta-analysis were carried out to estimate the prevalence of self-medication in Colombia and its associated factors in the population over 16. The pooled prevalence of self-medication found in young people and adults from urban areas of Colombia (64.2%; 95% CI 50.8%-77.5%) was similar to that recently found in a global meta-analysis, according to which the prevalence of self-medication was 60.0% (95% CI 40.2%-77.0%) in South America [5]. All the studies included in this meta-analysis were conducted in the urban population, which, according to the 2018 Colombian population census, constitutes 77% of the total Colombian population [22]. Additionally, 11 of the 12 included studies used randomized sampling. Only one study used a convenience sample, which had a low weight on the pooled calculations because of its small sample size (n= 150 subjects) [12]. Therefore, the pooled prevalence could be considered representative of the Colombian population, specifically those living in municipal capitals.
On the other hand, since there is no universal and standardized definition of self-medication, most studies use different time frames to measure its frequency [23]. This has led most of the systematic literature reviews to establish comparisons of the prevalence of self-medication between diverse populations without analyzing in depth the effect that these different time frames could have on the reported frequency. To address the problem of the different time frames used for measuring self-medication, which can undoubtedly contribute to the high heterogeneity observed, we performed a meta-analysis by subgroups, according to the time frame and the study population (General population vs. University students). Our results showed that when self-medication was evaluated for long periods, such as the previous 6-12 months or throughout life, the reported frequency was significantly higher, both in the general population (74.6%; 95% CI 61.3%-87.9%) and in university students (84.7%; 95% CI 75.1%-94.3%). In contrast, when self-medication was assessed during the previous 30 days, a significantly lower prevalence of self-medication was found (32.3%; 95% CI 25.4%-39.3%). This suggests that, at least in the urban population of Colombia, the time frame during which the self-medication behavior is evaluated could be more important than the effect of age or the university student status. However, the impact that the variables, time frame, age, socioeconomic status, and educational level have on the frequency of self-medication is complex and requires in-depth study. Previous studies have already reported an effect of time frame on the reported frequency of self-medication. This is the case of a systematic review from Brazil that reported a pooled prevalence of self-medication in the last 15 days of 35.0% (95% CI 39.0%-40.0%) [24]; meanwhile, a study evaluating self-medication during lifetime found a frequency of 66.6% (95% CI 61.3%-71.4%) [25]. In another case, a community-based cross-sectional study conducted in southwestern Ethiopia reported that the prevalence of self-medication changed depending on the time frame (58.2% for lifetime vs. 31.2% for current use) [26]. Indeed, when a behavior is evaluated by self-reporting, a long time frame does not allow recent changes to be detected and is associated with recall bias. For this reason, to assess the frequency of self-medication through questionnaires, surveys, and interviews, it is essential to select the recall period appropriately. Generally, a short recording period is preferable to a long one, especially when asking about recurring or frequent events [27].
Although the second objective of this study was to evaluate the factors associated with self-medication in Colombia, only some of the included studies assessed the same characteristics. However, it was possible to extract and analyze the data related to sex, groups of medications used for self-medication, and some causes associated with self-medication behavior. Regarding self-medication in men and women, we did not find differences within the general population, but we did find a higher frequency of self-medication in female university students compared to males (OR= 1.72; 95% CI 1.17-2.53), a finding that other authors have previously reported, and that can be explained as a strategy for relieving headaches, colds, and dysmenorrhea [28].
This meta-analysis found that, within the population that self-medicated, 37.7% used analgesics, 33.2% anti-inflammatories, and 14.8% antihistamines. Other studies have also shown these drugs to be the most used for self-medication. For example, in Ethiopia, the frequency of analgesics and anti-inflammatories was 46.1% [29], and in Iran, the frequency of analgesic use was 53.2%, with 39.3% for anti-inflammatories and 24.3% for antihistamines [30]. The non-responsible use of OTC medicines is an essential cause of adverse effects and severe drug interactions [31]-[33]. It is favored by selling in large-area stores without mediating the dispensing process and providing the user with the required information rather than by the insert that needs to be read.
The use of antimicrobials without a medical prescription is a cause for concern due to its implications for developing antimicrobial resistance, to which 1.27 million deaths were attributed in 2019 [34]. We found that antibiotics and antiparasitics were used for self-medication in 12.1% (95% CI 7.4-16.9%) of the cases, a lower frequency than that reported in other studies and systematic reviews. Within low- and middle-income countries, the pooled prevalence of self-medication with antimicrobials has been found to range from 39 to 78%, depending on the countries included in each analysis [35]-[37]. Therefore, given this high heterogeneity, it is essential to conduct investigations by countries within the same income level group [38]. In Colombia, in particular, it was observed that by 2011, up to 80% of pharmacies did not comply with the rule of dispensing antibiotics only with a medical prescription, and the pharmacy staff did not provide the required information to the user [39]. This phenomenon contributes substantially to using antibiotics through self-medication and is added to other associated factors, such as those discussed below.
Finally, some of the determinants of self-medication were explored; however, despite these causes being so varied and complex, it was only possible to extract information about two: lack of time and delays in care. These two situations were reported by those who self-medicated in 35.2% (95% CI 25.6%-44.7%) of the cases, although their frequency in university students was significantly lower (19.7%; 95% CI 10.3%-29.2%). Another systematic review also reported that the lack of time to see a doctor is reported recurrently by those who self-medicate [40], reaching a frequency of up to 68% in a high-income country such as Saudi Arabia [41]. Therefore, shortening the long waiting times for healthcare services and improving pharmaceutical services might encourage the proper use of medications [42]. However, these strategies must be accompanied by others that involve pharmacy staff, who must have the appropriate formation and training, comply with the rules for dispensing prescription drugs, and provide the user with the required information on the adequate use of drugs [43],[44]. As limitations of the present study, it should be considered that some studies included in the meta-analysis evaluated self-medication by self-report (memory bias). In contrast, others did so by direct pharmacy observation (information bias), which may contribute to the observed heterogeneity. Additionally, the studies analyzed did not use a standardized time frame. These limitations are common to all studies on self-medication and point to the need for standardized definitions and tools. Therefore, within the recommendations and prospects for future studies, the need to develop and validate instruments to assess the occurrence of self-medication, especially responsible self-medication, as defined by the World Health Organization, is emphasized. These instruments should set a specified time frame, ideally short, as this minimizes recall bias, allows for population assessments, and can detect recent changes in behavior, for example, after an intervention. Additionally, given the impact of self-medication globally, these instruments must be developed and validated in different languages and cultural contexts, and they must evaluate the medications used.
CONCLUSIONS
A systematic review and meta-analysis were carried out to estimate the prevalence of self-medication in Colombia and its associated factors in the population over 16. A total of 12 studies with 5,668 were included, and a pooled prevalence of self-medication in young people and adults in Colombia (64.2%; 95% CI 50.8%-77.5%) was calculated. However, the prevalence was significantly lower when self-medication was evaluated during the last 30 days (32.3%) compared to more extended time frames (six months, one year, or throughout life). The self-medication frequency was higher in female university students than in males. The most frequently used medications were analgesics (37.7%), anti-inflammatories (33.2%), antihistamines (14.8%), and antibiotics/antiparasitics (12.1%). As determinants of self-medication, lack of time and delays in medical care were reported in 35.2% (95% CI 25.6%-44.7%) of the cases. The reported frequency of self-medication in Colombia changed across the studies depending on the time frame used; therefore, this should be considered when conducting and comparing studies on self-medication prevalence. Although over-the-counter drugs were the most frequently involved (85.7%), the reported use of prescription drugs such as antibiotics/antiparasitics was 12.1%. Moreover, considering that lacking time and delays in medical care were reported in a third of cases of self-medication, shortening the long waiting times for healthcare services might contribute to the proper use of medications.
