Background

The World Health Organization (WHO) defines self-medication as a practice that leads individuals to treat their illnesses with authorized medicines. Self-medication can be defined as "the act of the subject, on his initiative, to consume a drug without consulting a doctor, and whether the drug is already in his possession or whether he obtains it from a pharmacy or another person [1]. It is a very prevalent practice, and the majority of medicines consumed by the population are available without a doctor's prescription [2]. The practice of self-medication is common throughout the world, both in developing and developed countries [3]. The number of people who self-medicate varies significantly by country, for example, in a study conducted in the United States about 71% of men and 82% of women had used self-medication at least once [4]. In the United Kingdom of Britain and Northern Ireland, 41.5% of people had used medicines without a prescription [5].

The use of drugs during pregnancy requires a careful reflection on the benefits to the mother and the risks to the fetus. It is a difficult medical situation for clinicians to select the drug because of the various pharmacokinetic and physiological changes encountered in pregnant women [6, 7]. Self-medication during pregnancy is of great interest; it carries serious risks of drug interactions, misdiagnosis, use of excessive doses of drugs, and prolonged use of drugs. [8, 9]. In Europe, a survey of 740 pregnant women conducted in France revealed that 41.5% of the participants in the study practiced self-medication [3]. In Africa, a systematic review conducted in Ethiopia revealed a prevalence of self-medication of 12.8% to 77.1% [7]. The studies are heterogeneous some of them report a high prevalence and others show the opposite. Also, there is one meta-analysis of the prevalence of self-medication in pregnant women that has been carried out in the world with a small number of included studies [10]. The real objective of this study was to update the published data. The goal of this study was to estimate the pooled prevalence of self-medication in worldwide pregnant women through a systematic review and a meta-analysis of published studies on self-medication during pregnancy.

Methods

The steps of this systematic review complied with the recommendations in the PRISMA (Preferred Reporting for Systematic Reviews and Meta-Analysis) grid. A systematic review protocol was registered by PROSPERO.55 with ID = CRD42022312333. In this study, the focused research question was: “What is the prevalence of self-medication in pregnant women in the world?” [10].

Eligibility criteria

We included all published articles between January 2011 and December 2021. We also included all original studies published in English and French and human studies of pregnant or postnatal women. Studies or scientific medical literature were also included if they described the prevalence of self-medication. Only observational studies of cross-sectional type were included, as well as cohort studies if they mentioned the prevalence of self-medication in pregnant women.

We excluded non-English studies (Persian and Spanish) if they didn’t contain the requested data. We also excluded unpublished reports, pilot studies, conference abstracts, opinion articles, editorial reports, seminal work, systematic reviews, and animal research.

Information sources and search strategy

To collect the maximum of data, we performed an exhaustive bibliographic search in PubMed, Science Direct, Google Scholar, and Web of Science databases started in December 2021. It was performed to identify all relevant studies available from January 2011 to December 2021. The search was performed using Boolean operators AND or OR that narrowed the search and also using a combination of terms and medical subject heading (MeSH). For example, the search strategy in PubMed was as follows: (((self care[MeSH Terms])) OR (self-medication [MeSH Terms])) AND (pregnan*[MeSH Terms]).

All stages of the search were performed by 2 independent researchers (B.A) and (L.L). Any disagreements were resolved by a third researcher (K.Y).

Selection process

A screening strategy was implemented to identify all relevant studies. We used Covidence, software that was developed by an Australian in 2015, it’s the primary screening and data extraction tool for Cochrane authors conducting standard intervention reviews. Covidence is designed to perform the following functions to make review production more efficient. Initially, duplicates were eliminated by the software. Then secondly, the initial selection was made according to the abstract and title. This was followed by reading the full text to identify eligible studies. In case of disagreement, a third reviewer (K.Y) was asked to make a clean agreement. Finally, the articles included in this review were downloaded and the references of each article were manually searched to determine whether other studies met the criteria.

Study risk of bias assessment

The quality assessment of eligible studies was reviewed by two independent reviewers (B.A) and (L.L). This process was conducted using a recent version of the Joanna Briggs Institute's critical appraisal tools (Checklist for Analytical Cross-Sectional Studies and Cohort studies) [11, 12]. Joanna Briggs Institute (JBI) is an international research organization based in the Faculty of Medical and Health Sciences at the University of Adelaide in South Australia. [13]. The purpose of this process is to assess the methodological quality and risk of bias in included studies. This checklist is divided into eight items for cross-sectional studies and ten items for cohort studies. Each item is scored with one point. A study was considered low quality if it had 0–4 points, moderate quality if it had 5–6 points, and high quality if it had 7–8 points.

A third reviewer (Y.K.) was consulted if a consensus could not be reached. When information was missing from the studies, we tried to contact the authors by e-mail. All cross-sectional and cohort studies were included, regardless of their quality score. Articles with missing data were included as long as they reported the prevalence of self-medication.

The results of the quality assessment of eligible studies are presented in Fig. 2.

Effect measures

What is the prevalence of self-medication? The prevalence is the proportion of pregnant women who have self-medicated or taken drugs without a doctor’s prescription during their pregnancy period.

Data extraction

Data covering author, country, year of publication, sample size, age of participants, and prevalence of self-medication were extracted and collated in an Excel table. We also extracted the associated factors of self-medication including the reason for use, source of information, illness of drug used, type of drugs used, and finally sociodemographic characteristics of pregnant women. Any disagreements were resolved by consensus with a third reviewer (K.Y). For non-English studies, we extracted data from the abstracts available in English.

Statistical analysis

To calculate the pooled prevalence, we used a random-effects model, calculating effect sizes with a confidence interval (CI) using Wilson-Score, and Clopper Pearson methods [14] To estimate the true treatment effects that can be expected we calculated the prediction interval using the method that are described by Barker et al. (2021) [15]. The Cochrane criteria such as I2 < 40%—insignificant, 40- < 75% is moderate while 75% + is considerable [16].

To investigate the effect of variables on heterogeneity, we performed a subgroup analysis by ranking eligible studies by income (based on the country in which the study was conducted). This was done by consulting the World Bank website [17], study region, sample size, and study quality. The R software was used for the statistical analysis. In the R4.1.2 version, we used the Meta and Metafor packages for the metanalysis.

Sensitivity analysis

To identify and reduce the source of heterogeneity, we conducted sensitivity analysis based on studies with small sample sizes (≥ 500 participants), and studies with a high risk of bias in any methodological domain (studies that have a low-quality score).

Results

Study selection

The flow chart of the studies included in this systematic review is shown in Fig. 1. The search generated 4475 articles, 151 of which were duplicates and subsequently deleted. We examined the titles and abstracts of 4324 articles. In addition, 4202 were excluded as ineligible based on the inclusion and exclusion criteria of the analysis. Then we examined the full text of the remaining 122 articles for eligibility. We, therefore, excluded 61 studies the reason (of the wrong target population (32), inappropriate study design (22), non-English articles (5), and thesis (2)), therefore 59 studies were eligible and they proceeded to the data extraction after adding four articles from the manual search. Finally, 65 studies were eligible and included in the present systematic review.

Fig. 1
figure 1

Systematic review flowchart

Full size image

Characteristics of included studies

The total sample size analyzed included 42,615 pregnant women. The studies were published between 2011 and 2021. All included studies were cross-sectional observational studies except three studies were Cohort [18,19,20]. The oldest study in terms of publication year (2011) is [21]. The most recent studies published in 2021 are (Table 1).

Table 1 Main characteristics of the included studies
Full size table

One study used the largest sample size of 9,459 participants. This was a multinational study conducted in Europe (western, northern, and eastern), North and South America, and Australia [75]. The study carried out in France recruited only 68 participants [41]. The ages of the participants ranged from 15 to 60 years. Regarding the region where the studies were conducted, 26 (31.32%) studies were conducted in Africa, 25 (30.12%) in Asia 21 (25.30%) in Europe, 10 (12.04%) in America, 1 (1.20%) in Australia.

Of most of the studies 29 (34.93%) were conducted in high-income countries, 28 (33.73%) in low-middle-income countries, 14 (16.86%) in low-income countries, and 12 (14%0.45) in upper-middle-income countries. Regarding the quality of the studies, 42 (50.60%) articles were high quality (score of 7–8), 33 (39.75%) medium quality (6—5), and 8 studies (9.63%) low quality (4—0).

Risk of bias in studies

The findings of the quality appraisal of eligible studies were reported in Fig. 2. The tool is used to indicate the methodological quality and appropriateness of the observational studies, including cross-sectional and cohort studies that were reviewed in this study. We determined the score by counting the asterisks (*) that we gave to each answer to the items in the grids, where a high score (11–8) indicates a higher quality of a study, a Middle score (6–5) indicates middle quality and Low Score (4–0) indicates low quality. The sex of the sixty-five studies was evaluated by the abstract since these articles are in Spanish and Persian language and we have no response from their authors to retrieve the full text. Two reviewers completed this process, and where there were discrepancies, a team of reviewers intervened to resolve them.

Fig. 2
figure 2

Result of quality appraisal using JBI appraisal tool

Full size image

Meta-analysis of the prevalence of self-medication in pregnant women

The overall prevalence of self-medication among pregnant women was 44.50% (95% CI: 38.92–50.23) with a prediction interval (of 8.76–87.00) Overall the prevalence ranged from 2.61% to 85%, as shown in the forest plot (Fig. 3). The I2 test revealed a high statistically significant heterogeneity of 99%.

Fig. 3
figure 3

Forest Plot of the prevalence of self-medication among pregnant women

Full size image

Results of the subgroup analysis are shown in Fig. 4 (A, B, C, D, E). Analysis was based on income level, geographic region, quality of publication study, sample size, and study design. Heterogeneity, assessed by the statistical test I2 was considerable, statistically significant for all subgroup analyses, and ranging from 96 to 99%. By region, the highest prevalence of self-medication was 77.42% (95% CI: 77.27–82.80%) in Australia and the lowest was 33.17% (95% CI: 25.22- 42.22%) in Asia (Fig. 4B). Stratifying by a score of the quality assessment of the studies, the prevalence of self-medication was 33.95% (95% CI: 27.72–40.80%) in studies with high scores. 58.95% (95% CI: 52.03–65.53%), and 42.68% (95% CI: 23.28–64.63%), in middle and low quality respectively.

Fig. 4
figure 4

Forest Plot of subgroups analysis based on Income Level (A), Region (B), Quality of studies (C), Sample size (D), and Study Design (E)

Full size image

For the cohort studies, the prevalence was lower 33.52% (95% CI: 27.68- 39.91%), and 44.93%(95% CI: 39.14- 50.86%) for the cross-sectional studies.

Sensitivity analysis

Due to the high heterogeneity of the results, sensitivity analysis was done after excluding studies with a high risk of bias (Fig. 5A) and studies with small sample sizes (Fig. 5B). The sensitivity analysis showed the stability of the results. The overall prevalence of self-medication based on the random effect model was determined to be 43.70% and 47.34% for the studies with a low score and small sample size, respectively. The results showed that the prevalence of self-medication did not generally change confirming the robustness and reliability of our findings.

Fig. 5
figure 5

Sensitivity analysis of the stability of the results of panel A studies with a high risk of bias and panel B studies with small sample size

Full size image

The reason and motivation for self-medication practice

The duration for making decisions regarding self-medication varied significantly, depending on factors such as women's health requirements, the availability of information, and personal preferences. Several studies found that women's individual choices and decisions were influenced by their personal preferences. There are many factors responsible for self–medication practice among pregnant women. The majority of the respondents pointed out that drugs are easily available in drug stores or pharmacies or by the availability of old prescription that helps themes to purchase drugs. Others think that they have better knowledge about the disease and the treatment, they know that the medication is safe during pregnancy. This report is logical when the respondents see that those illnesses as minor diseases will definitely not worry about visiting the health facility for professional care and will therefore prefer to buy non-prescribed drugs from the patent medicine shop for treatment. Especially when costly medical are expensive and the a lack of healthcare insurance, women consider self-medication as a cheaper practice. Others reported that the previous medication good experiences are one of the reasons for self-medication.

Self-medication for health problem treatments

Most pregnant women practice self-medication on the given health condition/diagnosis and accompanying treatment plans, and the desire to get over the pain, the complication, or if infertility persisted over time. Pregnant women practice self-medication to alleviate pregnancy-associated symptoms. Antiemetics, antiacids, and antipain are the most frequently drugs used to treat gastrointestinal disorders such as nausea, vomiting, abdominal pain, bloating, flatulence, and stomach aches followed by antibiotics and analgesics to treat cold and flu symptoms. Other women treat anemia and weakness with vitamins, calcium supplements, antianemia, folic acid, and iron. Some African studies reported that pregnant women self-medicate with Antimalaria to treat Malaria (Table 2). Whereas others reported self-medication practices with anti-inflammatory antirheumatic, anthelminthic, and antiepileptics. Other uses were specifically for the stimulation of labor or facilitation of labor and delivery. Finally, skin problems, sleep disorders, and weight loss are the fewer reasons that we have marked among some users. Illness drugs used for and family therapeutics of drugs are reported in Table 2.

Table 2 General description of Associated Factors of self-medication among worldwide pregnant women
Full size table

The source and quality of the information received

The choice of treatment for maternal illness was influenced partly by the source and quality of information on drugs received. Differences in over-the-counter drug use patterns and drug prescribing systems that differ from one country to another. For example, in some African countries, all medicines can be purchased without a prescription. However, Pharmacies or drug stores that were authorized to vend and supply drugs and relevant information left significant adverse implications on the overall health outcomes of users. This implies that women who engage in self-medication tend to be primarily guided by the origin of the information they receive and the quality of that information regarding medications. The majority of the studies pointed out various sources and the quality of information as key factors. First, some women opted for drug use during pregnancy based on their previous prescriptions and available medications within their family. Second, certain pregnant women received recommendations from their relatives (mothers), friends, and neighbors. Additionally, healthcare professionals (such as doctors and nurses) and online pharmacies also played a role in influencing women's decisions about medication use. Finally, other users were influenced by recommendations from sources such as the Internet, social media, newspapers, radio, and television.

Discussion

The results of the present study showed that the global prevalence of self-medication among pregnant women in the world is 44.55% although the results show a high heterogeneity. This rate is different from the results of a systematic review and meta-analysis that investigated self-medication among pregnant women worldwide (32%) [10]. Our systematic review differs from the earlier study in several key aspects. Firstly, it provides a more comprehensive and up-to-date analysis of the existing literature, offering a more complete understanding of the subject. We've included a broader spectrum of studies [53], incorporating diverse geographical locations and socioeconomic settings. Moreover, our review delves deeper into the underlying factors and context surrounding self-medication, offering a more holistic perspective on the issue. In essence, this systematic review and metanalysis contributes a more nuanced and comprehensive examination of the relationship between income status, region, and self-medication, providing valuable insights for healthcare policymakers, practitioners, and researchers.

Despite the prevalence of self-medication being higher in pregnant women, due to the complications of pregnancy, Our results compared to the results of other studies showed that self-medication in pregnant women was relatively low in comparison with other groups in general. In Ethiopia, the prevalence of self-medication was found to be 44.0%. Geographical-based subgroup analysis revealed that the highest prevalence was observed in the capital of Ethiopia, Addis Ababa, 62.8%. Population-based analysis indicated that healthcare professionals and students were the main practitioners of self-medication. Besides, the prevalence of self-medication practice in pregnant women is approximately 22.9% [82]. A meta-analysis conducted in Iran among students showed a very high prevalence of 70.1% [83]. The results of the subgroup analysis in our study showed that the prevalence of self-medication varies by region, this finding is attributable to the difference in health-related socio-cultural knowledge, beliefs, attitudes, and behaviors among pregnant women from different geographical areas. The large difference in prevalence rates of self-medication in different regions of the world may be due to differences in over-the-counter drug use patterns and drug prescribing systems that differ from one country to another. In some African countries, all medicines can be purchased without a prescription [26]. As an example, the pooled prevalence of self-medication during pregnancy in Ghana was 65.4% (95% CI = 58.2%–72.6%; I2 = 88.32%; p < 0.001) [84]. Other reasons that may influence self-medication include the high cost of medical visits and limited health insurance coverage. Women with unplanned pregnancies are also more likely to self-medicate [69]. The low prevalence rate noted in Asian countries may be due to the cultural and social habits of using complementary and alternative medicine. The prevalence of the use of herbal medicines among pregnant women varied from 9.2% to 90.2% this was reported in a systematic review conducted by Eastern Mediterranean Regional Office [85]. Most pregnant women believe that herbal medicine is more effective than conventional medicine [85, 86]. Others think that are safe and secure for the mother and her fetus and have fewer side effects than conventional medicine during pregnancy [82, 87]. Stratifying by a score of the quality assessment of the studies, the differences in the prevalence of self-medication can be explained by the fact that high-quality studies have a low risk of bias and therefore the prevalence approaches the pooled prevalence (44.50%). However, the low-quality studies probably overestimated the prevalence because of the low recruited sample size. For the cohort studies, the prevalence was lower this can be explained by the fact that the research methodology of the cohort studies is very rigorous with the best evidence [88]. The results of the subgroup analysis in our study showed that the prevalence of self-medication varies by income status. The high prevalence rate noted in countries with high-income levels was 58.88% compared to the low-income countries the rate of prevalence was 37.07%. We can explain these results that lower-income individuals may have limited financial resources to access healthcare services, including doctor consultations and prescription medications [89, 90]. Health literacy refers to an individual's ability to understand and use healthcare information to make informed decisions. This knowledge and understanding about appropriate medication use could contribute to a higher prevalence of self-medication. Lower-income individuals may have limited access to health education and information, resulting in lower health literacy [89].

In recent years, several initiatives and interventions, such as improving the knowledge of pregnant women about the consequences of self-medication, as well as the provision of brochures and catalogs, have been planned and implemented, which could be very effective in combating this practice. In addition, the continuous training of health professionals on the prescription of drugs and the advice given when dispensing drugs to pregnant women could reduce the prevalence of self-medication. Despite the attempts made by nations to decrease the prevalence of self-medication among expectant women, this behavior continues to increase. Consequently, there is an immediate need to implement novel and more efficient preventive strategies.

Strengths and limitations of the study

The strengths of our study are the large sample size, sample size analysis, and subgroup analysis. All these analyses reflect the methodological rigor of our systematic review and meta-analysis. On the other hand, the included articles provide large and profound information on various aspects of self-medication in pregnant women (prevalence, groups of drugs most used by pregnant women, groups of diseases most often treated by self-medication, and the most common reasons for self-medication) that can be used by health professionals to make decisions and organize effective interventions to prevent self-medication in pregnant women.

The limitation of this study was the fact that we included only articles in English and French, while we excluded some studies of Spanish and Farsi languages after exploiting their abstracts if they only reported the prevalence of self-medication. Another limitation of the present study was few databases were included and some more relevant ones were not searched. The limitations of the subgroup analysis are that gives both false positive and false negative results.

In addition, the quality of the included studies was different and the inclusion of some studies of poor quality may affect the final estimate.

Conclusion

The results obtained from this study showed that the prevalence of self-medication among pregnant women is relatively high. This requires effective measures and interventions to reduce self-medication. We recommend that health professionals consider implementing programs on the risks of self-medication, and strengthening the control and monitoring of over-the-counter sales of drugs. Physicians and pharmacists should also be made more sensitive to prescribing the appropriate medication and avoiding the provision of medication without a prescription.