Self-medication practices to prevent or manage COVID-19: A systematic review

Background Previous studies have assessed the prevalence and characteristics of self-medication in COVID-19. However, no systematic review has summarized their findings. Objective We conducted a systematic review to assess the prevalence of self-medication to prevent or manage COVID-19. Methods We used different keywords and searched studies published in PubMed, Scopus, Web of Science, Embase, two preprint repositories, Google, and Google Scholar. We included studies that reported original data and assessed self-medication to prevent or manage COVID-19. The risk of bias was assessed using the Newcastle–Ottawa Scale (NOS) modified for cross-sectional studies. Results We identified eight studies, all studies were cross-sectional, and only one detailed the question used to assess self-medication. The recall period was heterogeneous across studies. Of the eight studies, seven assessed self-medication without focusing on a specific symptom: four performed in the general population (self-medication prevalence ranged between <4% to 88.3%) and three in specific populations (range: 33.9% to 51.3%). In these seven studies, the most used medications varied widely, including antibiotics, chloroquine or hydroxychloroquine, acetaminophen, vitamins or supplements, ivermectin, and ibuprofen. The last study only assessed self-medication for fever due to COVID-19. Most studies had a risk of bias in the “representativeness of the sample” and “assessment of outcome” items of the NOS. Conclusions Studies that assessed self-medication for COVID-19 found heterogeneous results regarding self-medication prevalence and medications used. More well-designed and adequately reported studies are warranted to assess this topic.

•R1C2: Introduction doesn't give enough back-ground information about the topic. It should have included a background on burden of COVID 19, self-medication, rationale/justification to conduct this review, availability of previously published review article. oWe agree with the recommendation; we added the following text in the second and third paragraphs of the introduction: Line 36-40: "This may lead to unintended consequences, such as adverse events, unnecessary expenses, delay in attending professional evaluation, masking of symptoms, and drug interactions [17][18][19]. Self-medication prevalence varies according to several factors, such as the methodology used to assess self-medication [20], the population characteristics [20][21][22], and across different countries and contexts [23][24][25]." Line 41-43: "Previous studies have assessed the prevalence and characteristics of COVID-19 self-medication, reporting the use of several medications, herbal products, and dietary supplements as treatment or prevention for . However, to date, no systematic review has summarized their findings." •R1C3: Line 36-37, is the objective of this review just assessing the frequency of selfmedication for COVID-19? Many things are mentioned in the result section. Please include all the specific objectives of this review at the end of the introduction section. oWe agree with the recommendation and added these items to the objective (end of the introduction), as following: Line 44-46: "we aimed to assess the prevalence of self-medication for the prevention or management of COVID-19. In addition, we assessed the type of medication used, reasons to practice self-medication, from where such medications were obtained, and adverse events due to its practice." •R1C4: Better if 'frequency' is replaced by 'prevalence' throughout the document. oWe agree with the recommendation and replaced "frequency" with "prevalence" throughout the manuscript.
•R1C5: Line 50-51, in the inclusion criteria you gave the definition of self-medication while you said the difference in definition is one of the limitations that hinder you from doing meta-analysis. so, remove the definition in the inclusion criteria. Ok to retain it in the introduction. oIn order to clarify this, we have reformulated the text at the end of the "inclusion criteria" subheading of the methods, as follows: Line 63-65: "We considered that a study assessed the use of self-medication in any of the following situations: 1) the study reported explicitly that self-medication was assessed (regardless of the definition used), or 2) study assessed the use of medications without medical prescription." •R1C6: Include the following in the methods section, key words used for searching. oThank you for your observation. The keywords used for the search in each database are available in the supplementary material (S1 Table). This is mentioned now in the second paragraph of the Methods: Line 52-53: "Search strategies were constructed a priori using different terms related to 'COVID-19' and 'self-medication' (Supplementary Table S1)." •R1C7: In your inclusion criteria nothing is said about the study participants in the retrieved studies (age (adult, children, or all), COVID status (positive, negative unknown)), region or geographic location where the studies were conducted, publication status (do you consider grey literature?). oWe agree with the recommendation. We have added the following in the "inclusion criteria" subsection of the Methods: Line 61-62: "in participants of any age and from any location." oAlso, we have added the following in the "Data sources" subsection: Line 54-55: "…and grey literature sources (Google and Google Scholar)." •R1C8: 'Drug' and 'medication' are used interchangeably in the document, be consistent. I prefer 'medication' than 'drug' in your context. oWe agree with the recommendation and replaced 'drug' by 'medication' throughout the manuscript where appropriate.

Introduction
•R2C1: Paragraph 1. The introduction part could be written with more information. For example, the authors may explain more on the examples of prevalence difference between countries, age group and occupation, in terms of the highest self-medication practices. A more thorough definition on self-medication would also be very useful, as this is the main focus of this paper, yet it is still unclear as it is only briefly mentioned in the introduction. •R2C2: Paragraph 2. It would also be good if the authors described on the examples of drugs that has been proposed as the potential candidate for Covid-19, and the outcome of such treatments. The examples of direct and indirect consequences can also be further explained here, to make the introduction more informative. •R2C3: Paragraph 3. Some description on the examples of medication used as selftreatment and indication in Covid would also be informative and will make the introduction more dynamic. oWe agree with these 3 recommendations. Accordingly, we added the following text in the first paragraph of the introduction: Line 25-32: "With the progression of the COVID-19 pandemic, several medications have been proposed as potential candidates for this disease [1], most of which resulted in little or no benefit for the patients [2,3] or even in harms [4]. For example, hydroxychloroquine gained wide attention as a possible treatment for COVID-19 due to favorable results found in in-vitro or small uncontrolled studies [5]. However, later, randomized trials in hospitalized patients, such as the RECOVERY trial [6] and the Solidarity trial [3], failed to find any clinical benefit compared to usual care. This is similar to what happened with azithromycin [7-10], while there are still few welldesigned trials that have assessed other medications such as ivermectin [11][12][13] or vitamins supplements [14,15]." oAlso, we added the following to the second and third paragraphs of the introduction: Line 36-40: "This may lead to unintended consequences, such as adverse events, unnecessary expenses, delay in attending professional evaluation, masking of symptoms, and drug interactions [17][18][19]. Self-medication prevalence varies according to several factors, such as the methodology used to assess self-medication [20], the population characteristics [20][21][22], and across different countries and contexts [23][24][25]." Line 41-43: "Previous studies have assessed the prevalence and characteristics of COVID-19 self-medication, reporting the use of several medications, herbal products, and dietary supplements as treatment or prevention for . However, to date, no systematic review has summarized their findings." Methods •R2C4: Data source. It would be good if the authors can include the terms used for the papers search, as this is an important component for a systematic review. •Thank you for your observation. The keywords used for the search in each database are available in the supplementary material (S1 Table). This is mentioned now in the second paragraph of the Methods: oLine 52-53: "Search strategies were constructed a priori using different terms related to   Table S1)." Results •R2C5: Definition of self-medication. If 9 of the 11 studies did not mention on the definition of self-medication, how did the author include these papers in the study? More explanation would be very useful. Please explain this in the manuscript. This showed the importance of providing the keywords for article search. oThank you for your observation. To clarify this, we have added the following at the end of the first paragraph of the "inclusion criteria" subheading: Line 63-65: "We considered that a study assessed the use of self-medication in any of the following situations: 1) the study reported explicitly that self-medication was assessed (regardless of the definition used), or 2) study assessed the use of medications without medical prescription." •R2C6: Who suggested the practice of self-medication. In the results it stated that seeking advice from doctors and medical guideline are included as the source to the advice on self-medication. But if someone get an advice from a doctor, would it still be defined as self-medication? This goes back to the concern on the definition of the selfmedication and the selection of papers in this study. Please explain. oThanks for the observation. We have discussed the subject, and we consider that the self-medication definition is incompatible with the situations in which participants consulted with physicians regarding their medications. oAccordingly, we have added this definition in the "inclusion criteria" subsection of the methods: oLine 63-65: "We considered that a study assessed the use of self-medication in any of the following situations: 1) the study reported explicitly that self-medication was assessed (regardless of the definition used), or 2) the study assessed the use of medications without medical prescription." oAlso, we reviewed again all the studies, and excluded 3 of them which did not match with this self-medication definition: Ahmed (2020), Kamarli Altun (2020), and Chauhan (2020). Accordingly, the review now includes only 8 studies, and all the articles have been updated.

Discussion
•R2C7: It would be good if the author may emphasize on the importance of the findings and why this paper is worth to be published. oWe agree with the recommendation. Accordingly, we added the following as the last paragraph of the discussion (right before the conclusion): oLine 244-249: "The findings of this review suggest that there is an important though heterogeneous prevalence of self-medication to prevent or manage COVID-19. This includes some medications that have not shown any benefit so far and may therefore expose people to unnecessary adverse events. The results vary widely across studies, suggesting that each context (region, country) has its own self-medication patterns and impact, calling on local health authorities to promote research and interventions to reduce potential self-medication adverse consequences." •R2C8: It would be good if the author may discuss on the future study or research gap related to the topic. oThanks for the suggestion. We added the following as the penultimate paragraph of the discussion: oLine 239-243: "Due to the limitations of the primary studies included in this review, more well-designed and adequately reported studies are needed, as well as the use of standardized self-medication definitions across studies. Moreover, since the COVID-19 status is changing rapidly, future studies are required to assess if there is any variation in self-medication trends between COVID-19 waves and after introducing the massive vaccination." Reviewer #3 •R3C1: Line 5 and 6 "Suppose the study aimed to assess the frequency of selfmedication for the prophylaxis or management of COVID-19. In that case, I was wondering why the authors included studies that assessed self-medication for all reasons?" oWe understand the concern of the reviewer. To clarify our inclusion criteria, we have changed the redaction as follows: Line 59-62: "We included original studies, either published in scientific journals or with full text available in preprint repositories, which reported original data and assessed the use of self-medication to prevent or manage COVID-19 as a primary or secondary outcome, in participants of any age and from any location." Line 63-65: "We considered that a study assessed the use of self-medication in any of the following situations: 1) the study reported explicitly that self-medication was assessed (regardless of the definition used), or 2) the study assessed the use of medications without medical prescription." •R3C2: Line 12 "The authors stated that they "identified 11 studies that assessed selfmedication for the prevention or management of COVID-19." It is unclear how the authors linked the self-medication in the 7 studies that assessed "self-medication for any reason" to COVID-19. The authors' definition of self-medication did not specify COVID-19, e.g., Quispe-Cañari (2021). The respondents may be self-medicating for another ailment". oThank you for your observation. To clarify this, we have added the following at the end of the first paragraph of the "inclusion criteria" subheading: Line 63-65: "We considered that a study assessed the use of self-medication in any of the following situations: 1) the study reported explicitly that self-medication was assessed (regardless of the definition used), or 2) study assessed the use of medications without medical prescription. " oRegarding the study of Quispe Cañari et al, it assessed self-medication for COVID-19 prevention and management (table 2 of the results of the study, link: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7832015/). These are the results that we took into account.
•R3C3: Line 19 "It is unclear why the authors chose the term frequency instead of prevalence throughout the manuscript. 'Prevalence' may be more appropriate to use in this context". oWe agree and have made such changes throughout the manuscript.
•R3C4: Line 29 "Replace 'now' with 'later'" oAs we modified the writing of the manuscript, this suggestion is not valid anymore.
•R3C5: Line 33 to 35 "The sentence needs to be rephrased to reveal the intended meaning. Consider rephrasing to: "Previous studies have assessed the frequency and characteristics of COVID-19 self-medication to figure out which medications are being used that are ineffective or potentially dangerous and which factors predispose people to self-medication" oThanks for the observation. We have changed the following: Line 2-3: "Previous studies have assessed the prevalence and characteristics of selfmedication in COVID-19. However, no systematic review has summarized their findings." Line 41-43: "Previous studies have assessed the prevalence and characteristics of COVID-19 self-medication, reporting the use of several medications, herbal products, and dietary supplements as treatment or prevention for . However, to date, no systematic review has summarized their findings." •R3C6: Line 34 "The use of the term 'useless' is unsuitable for scientific writing. Consider using 'ineffective'" oAs we modified the writing of the manuscript, this suggestion is not valid anymore.
•R3C7: Line 43 "I wondered if it is possible to search all these databases in a single day without any pre-defined search guide. It will be good to state that the search terms were developed a priori for better clarity to any reader". oWe agree with the recommendation, So, in order to guide the readers, we added the following line in "Data sources" section: Line 52-53: "Search strategies were constructed a priori using different terms related to 'COVID-19' and 'self-medication' (Supplementary Table S1)." •R3C8: Line 44 to 45 "This is very nice. By not restricting, you avoid what is called "language bias". However, since no language restriction was applied, it will be good for the authors to state how articles found that are not written in English were interpreted? Assuming none was found, the interpretation criteria should have been mentioned in the protocol, but unfortunately, it was not stated either in the protocol". oThanks for the observation. We added the following lines in the "Data sources" section to clarify this: Line 55-56: "We planned to hire a translation service if any study written in a language other than English or Spanish was found." •R3C9: Line 120 "The Table title does not tally with the content. It states, "…. assessed self-medication of any drug for any reason". The title will assume that studies that assessed self-medication for the prevention and treatment of COVID-19 will not be included. The title needs to be rephrased to be all-encompassing. "Characteristics and findings of studies that assessed self-medication practices during the COVID-19 pandemic"" oWe merged tables 2 y 3, and this new table 2 have the following title: Line 131-132: "Characteristics and findings of studies that assessed self-medication to prevent or manage COVID-19." •R3C10: Table 3 "In the 3rd column in the 1st row, i.e., Mansuri (2020), it is unclear how you presented the Subjects (age, sex). It seems to be in percentage. The same applies to Zavala-Flores (2020). The unit of age should be stated, i.e., years, months, e.t.c." oWe agree and have made the change, in the first row of the third column of table 3 now it says "age: 60.8% were less than 40 years, female: 60.3%". Then, in the last row of the third column of table 2, now it says, "Mean age: 60.3% years".
•R3C11: Line 174 "Kindly rephrase to "Only one study [22] specified the source of the patients' self-medications"" oWe agree and have made such change in the results. Please, see the second paragraph (line 180) of the "Reasons to practice self-medication, from where were the medications obtained, and adverse effects" subheading in the results.
•R3C12: Line 203 "Current guidelines??? It will be good to mention which guideline the authors are referring to. COVID-19 guideline?? By which body?? WHO, CDC etc" oWe consider the recommendation, so we rephrased the sentence (4th paragraph of the discussion). Now it says: Line 215-216: "As stated by the current COVID-19 guidelines by WHO and IDSA [41,42]." •R3C13: Throughout the manuscript "The manuscript needs detailed proofreading and revision for its English" oWe agree with the recommendation, and a translator has checked the revised version of this manuscript.

Reviewer #4
•R4C1: Page 3, paragraph 3: "…we aimed to assess the frequency of self-medication…". Please explain the reason of using the term "frequency" rather than "prevalence". oWe have reviewed the differences between the two terms and agree with your suggestion. So that, we replaced 'frequency' by 'prevalence' throughout the manuscript.
•R4C2: Page 5-6, Table 1: Both the final score for Ahmed (2020) and Chauhan (2020) are zero. Please explain the rational of include these two articles in the systematic review although the scores are zero. oDue to the commentaries of another reviewer, we have realized that 3 of the initially included studies did not fulfil the self-medication definition because they included medication taken due to a prescription. We have excluded these 3 studies from the results: Ahmed (2020), Kamarli Altun (2020), and Chauhan (2020).
•R4C3: Page 8, paragraph 2: "Of these seven studies, only two specified which question was asked, whereas the other five studies did not detail whether participants could report self-medication only for a prespecified list of drugs or whether it was an open question". The used of a prespecified list of drugs might narrow down the answers from the respondents. This subsequently might lead to biases in the study findings. This issue needs to be discussed in the discussion section. oWe agree with your suggestion and decide to add the following sentence (6th paragraph of the discussion): Line 232-234: "In addition, it is possible that studies that evaluated self-medication through a pre-specified list of medications reported biased prevalences, since these medications may not have included all the most commonly used medications [47]  Describe where the data may be found in full sentences. If you are copying our sample text, replace any instances of XXX with the appropriate details.
If the data are held or will be held in a public repository, include URLs, accession numbers or DOIs. If this information will only be available after acceptance, indicate this by ticking the box below. For example: All XXX files are available from the XXX database (accession number(s) XXX, XXX.).
• If the data are all contained within the manuscript and/or Supporting Information files, enter the following: All relevant data are within the manuscript and its Supporting Information files.
• If neither of these applies but you are able to provide details of access elsewhere, with or without limitations, please do so. For example: Data cannot be shared publicly because of [XXX]. Data are available from the XXX Institutional Data Access / Ethics Committee (contact via XXX) for researchers who meet the criteria for access to confidential data.
The data underlying the results presented in the study are available from (include the name of the third party • All relevant data are within the paper and its Supporting Information files. and contact information or URL). This text is appropriate if the data are owned by a third party and authors do not have permission to share the data.

• * typeset
Additional data availability information: Tick here if your circumstances are not covered by the questions above and you need the journal's help to make your data available. such as ivermectin [11][12][13] or vitamins supplements [14,15]. 32 Nonetheless, the fear of contracting the virus, low access to health services, and massive 33 misinformation have led some people to self-medicate. According to the World Health Organization 34 (WHO), self-medication "involves the use of medicinal products by the consumer to treat self-35 recognized disorders or symptoms." [16]. This may lead to unintended consequences, such as adverse 36 events, unnecessary expenses, delay in attending professional evaluation, masking of symptoms, and 37 drug interactions [17][18][19]. Self-medication prevalence varies according to several factors, such as the 38 methodology used to assess self-medication [20], the population characteristics [20][21][22], and across 39 different countries and contexts [23][24][25]. 40 Previous studies have assessed the prevalence and characteristics of COVID-19 self-medication, 41 reporting the use of several medications, herbal products, and dietary supplements as treatment or 42 prevention for . However, to date, no systematic review has summarized their findings. 43 Thus, in this systematic review, we aimed to assess the prevalence of self-medication for the prevention 44 or management of COVID-19. In addition, we assessed the type of medication used, reasons to practice 45 self-medication, from where such medications were obtained, and adverse events due to its practice. 46

Data sources 51
Search strategies were constructed a priori using different terms related to 'COVID-19' and 'self-52 medication' (Supplementary Table S1). On February 4th, 2021, we searched PubMed, Scopus, Web 53 of Science, Embase, two preprint repositories (MedRxiv and SciELO Preprints), and grey literature 54 sources (Google and Google Scholar). No language restriction was applied. We planned to hire a 55 translation service if any study written in a language other than English or Spanish was found. The 56 search was restricted to include documents published in 2020 or 2021. 57 58 We included original studies, either published in scientific journals or with full text available in 59 preprint repositories, which reported original data and assessed the use of self-medication to prevent 60 or manage COVID-19 as a primary or secondary outcome, in participants of any age and from any 61

location. 62
We considered that a study assessed the use of self-medication in any of the following situations: 1) 63 the study reported explicitly that self-medication was assessed (regardless of the definition used), or 64 2) the study assessed the use of medications without medical prescription. 65

Study selection 66
Two authors (AQL and MHG) independently selected the studies. For this, they first screened the 67 search results by title and abstract according to the inclusion criteria using the web-based tool Rayyan 68 (http://rayyan.qcri.org). Later, they reviewed the full text of the relevant studies to determine whether 69 they were appropriate for study inclusion. Discrepancies were consulted with another author (ATR) 70 and resolved by consensus. 71 Two authors (CBI and RQE) independently extracted the relevant data for the review using a pre-73 piloted Microsoft Excel spreadsheet. Again, any discrepancies were discussed with another author 74 (ATR) and resolved with consensus. 75 The following variables were assessed for each study: author, year of publication, country, study 76 design, setting, population, sample technique, date of data collection, number of participants, age, sex, 77 how was self-medication assessed, the prevalence of self-medication, reasons for practicing self-78 medication, who recommended such medication, from where was the medication obtained, and 79 adverse events related to its use. 80

Risk of bias assessment 81
Two authors (CBI and RQE) independently evaluated the risk of bias for each included study using 82 the Newcastle-Ottawa Scale (NOS) adapted for descriptive cross-sectional studies [28]. This scale 83 comprises two criteria: the selection criteria (representativeness of the sample, sample size, and 84 nonresponders) and the outcome criteria (outcome assessment and statistical tests). Any discrepancy 85 was discussed with another author (ATR). 86

Statistical analysis 87
Because studies were performed in different populations and using different definitions of self-88 medication, we did not perform meta-analyses and decided to present the results of each study 89 separately.  100 We assessed the risk of bias using the NOS. The study performed by Sadio [29] included two samples: 101 one obtained using a probabilistic sample, and the other used a convenience sample. Thus, we assessed 102 the risk of bias of these two samples separately. 103

Methodological quality of the included studies
Of the eight studies, only one (Sadio probabilistic sample) fulfilled the five stars in the NOS scale. In 104 addition, the Sadio non-probabilistic sample, along with the other three studies, received four out of 105 five stars. All studies except one (Sadio probabilistic sample) failed in representing the sample item of 106 the selection criteria (Table 1). 107 1 Representativeness of the sample: One star was given to studies with random sampling or census. 111 2 Sample size: One star was given to studies with justified and satisfactory sample size.

112
3 Non-respondents: If comparability between respondents' and nonrespondents' characteristics was established and the 113 response rate was satisfactory, one star was given.
114 4 Outcome evaluation: If the study explicitly mentioned how self-medication was defined and how long the recall period 115 was, one star was given.
116 5 Statistical test: One star was given if it did not use a complex sample and the sample had been calculated correctly, or if it 117 used a complex sample and such sampling was considered to estimate the self-medication prevalence.

Studies characteristics 123
Studies characteristics are shown in and one in workers from five sectors (health care, air transport, police, road transport, and informal) 137 The eight studies included between 132 and 3,792 participants. Five studies mentioned the summary 139 measure (mean or median) of the participants' age. Such summary measure (either mean or median) 140 ranged from 21 to 60.3 years. All studies were performed in adults. The percentage of females ranged 141 from 28.4% to 69.1% (Table 2). 142

Definition of self-medication 143
All studies assessed self-medication through self-report. The questions used to assess self-medication 144 were explicitly mentioned only in the one study [36], while other five studies report having asked for 145 the use of medicine without prescription [29,[31][32][33][34], and two studies gave no details regarding how 146 self-medication was measured in their surveys [30,35]

153
As showed in Table 2, We classified the eight included studies into three groups: 1) those that assessed 154 self-medication for COVID-19 prevention or management without focusing on a specific symptom, 155 and 2) those that assessed self-medication for a specific COVID-19 symptom. 156 Seven studies were included in the first group (assessed self-medication for COVID-19 prevention or 157 management without focusing on a specific symptom). Of these, four were performed in the general 158 population (with a range of self-medication prevalence between <4% and 88.3%) and three in specific 159 populations (which reported a self-medication prevalence of 33.9% for hospitalized adults with 160 COVID-19, 34.2% in adults who worked in five assessed sectors, and 51.3% in undergraduate students 161 of health-related careers). 162 Of these seven studies, six mentioned the list of medications reported by the participants. The reported 163 medications were different across the studies: six studies reported the consumption of antibiotics, five 164 of chloroquine or hydroxychloroquine, three of acetaminophen, three of vitamins or supplements, two 165 of ivermectin, and two of ibuprofen. 166 In addition, consumption preferences varied across studies. For example, when medications were 167 sorted in the order of highest to lowest consumption, vitamins/supplements occupied the first place in 168 two studies, but the fourth place in another study (Table 2). 169 The second group (studies that assessed the use of medication for a specific symptom) included one 170 study performed on the general population of Saudi Arabia, which found that 35.1% of the surveyed 171 individuals used self-medication for fever [36] (Table 2). 172

obtained, and adverse effects
Four of the eight included studies mentioned the reasons for self-medication. However, each study 175 assessed the reasons differently. Thus, one study referred only to the symptoms that motivated 176 participants to self-medicate, and the other three collected a plethora of reasons, including the fear of 177 stigmatization, fear of quarantine, affordability, the convenience of self-medication, and patients 178 believing that the symptoms were not severe (Table 3). 179 Only one study [34] specified the source of the patients' self-medications. This study was performed 180 on adults from Nigeria. The most common sources reported were the pharmacy (73.9%) and the patent 181 medicine vendor (23.6%). 182 Only two studies reported whether adverse events occurred in those who self-medicate. One study 183 performed on Undergraduate students of health-related careers found that 11.7% of those who self-184 medicate presented with side effects of self-medication in the past three months [32]. Nonetheless, this 185 study did not specify the adverse effects that occurred. The second study, performed in adults from 186 Nigeria, reported body rash (23.1%) followed by worsened condition (17.3%), yellowish eyes (7.7%), 187 swollen face (3.8%), and vomiting of blood and severe diarrhea (5.8%) [34]. 188 Table 3. Reasons to practice self-medication, from where were the 189 medications obtained, and adverse effects

Dare (2021)
Self-medication is affordable (37%) and convenient of 53%, 44%, and 53.6%, respectively; for a recall period that ranged from a single day to 6 months. 201 In our review, we only found eight studies. Although the included studies were pe rformed in six 202 countries, the lack of studies in some regions such as North America, the Middle East, North Africa, 203 and Oceania is evident. It is expected that the intercountry differences in drug promotion, regulations, 204 and the possibility of accessing some medications without a prescription can influence the self-205 medication patterns [38,39]. Moreover, seven of the eight included studies were performed in low-and 206 middle-income countries. While self-medication may be higher in such countries, due to the conditions 207 and structure of the health system [40], it is not possible to make solid comparisons due to the 208 heterogeneity in the self-medication assessment across studies. 209 Some of the included studies assessed over-the-counter medications such as acetaminophen or 210 nonsteroidal anti-inflammatory drugs, which are used for the symptom management of COVID-19, 211 along with several other diseases. However, the included studies have also found an heterogeneous 212 prevalence of self-medication with medications that have not proven to benefit in the prevention or 213 management of COVID-19 (such as antibiotics, chloroquine or hydroxychloroquine, vitamins or 214 supplements, ivermectin, and antiasthmatics), as stated by the current COVID-19 guidelines by WHO 215 and IDSA [41,42]. 216 In fact, the adverse effects of some of these medications are of great concern, such as antibiotic 217 resistance due to extensive antibiotic use, bleeding caused by aspirin use, inhibition of the immune 218 system caused by corticosteroids, or arrhythmia caused by hydroxychloroquine [4]. Some healthcare 219 organizations have issued statements on self-medication. For instance, the WHO recognizes that a 220 "successful" (i.e., beneficial) self-medication can be achieved in many countries only by improving 221 individuals' knowledge and education level in such a way that the potential damages of this practice 222 can be avoided [16]. Similarly, the International Pharmaceutical Federation, in tandem with the World 223

Self-Medication Industry [43], and the World Medical Association [44], emphasizes the responsible 224 use of non-prescription medications. 225
Similar to that found in previous systematic reviews [20,21], the included studies in our review tended 226 to use different questions to assess self-medication, or they did not specify the question used. adequately reported studies are needed, as well as the use of standardized self-medication definitions 240 across studies. Moreover, since the COVID-19 status is changing rapidly, future studies are required 241 to assess if there is any variation in self-medication trends between COVID-19 waves and after 242 introducing the massive vaccination. 243 The findings of this review suggest that there is an important though heterogeneous prevalence of self-244 medication to prevent or manage COVID-19. This includes some medications that have not shown any 245 benefit so far and may therefore expose people to unnecessary adverse events. The results vary widely 246 across studies, suggesting that each context (region, country) has its own self-medication patterns and 247 impact, calling on local health authorities to promote research and interventions to reduce potential 248 self-medication adverse consequences. 249

250
We identified eight studies that assessed the use of self-medication for the prevention or management 251 of COVID-19. Only one detailed the question used to assess self-medication. The recall period was 252 heterogeneous across studies. With respect to the seven studies that assessed self-medication without 253 focusing on a specific symptom: five performed in the general population (self-medication prevalence 254 ranged between <4% -88.3%) and three in specific populations. In these studies, the most used 255  such as ivermectin [11][12][13] or vitamins supplements [14,15]. 32 Nonetheless, the fear of contracting the virus, low access to health services, and massive 33 misinformation have led some people to self-medicate. According to the World Health Organization 34 (WHO), self-medication "involves the use of medicinal products by the consumer to treat self-35 recognized disorders or symptoms." [16]. This may lead to unintended consequences, such as adverse 36 events, unnecessary expenses, delay in attending professional evaluation, masking of symptoms, and 37 drug interactions [17][18][19]. Self-medication prevalence varies according to several factors, such as the 38 methodology used to assess self-medication [20], the population characteristics [20][21][22], and across 39 different countries and contexts [23][24][25]. 40 Previous studies have assessed the prevalence and characteristics of COVID-19 self-medication, 41 reporting the use of several medications, herbal products, and dietary supplements as treatment or 42 prevention for . However, to date, no systematic review has summarized their findings. 43 Thus, in this systematic review, we aimed to assess the prevalence of self-medication for the prevention 44 or management of COVID-19. In addition, we assessed the type of medication used, reasons to practice 45 self-medication, from where such medications were obtained, and adverse events due to its practice. 46

Data sources 51
Search strategies were constructed a priori using different terms related to 'COVID-19' and 'self-52 medication' (Supplementary Table S1). On February 4th, 2021, we searched PubMed, Scopus, Web 53 of Science, Embase, two preprint repositories (MedRxiv and SciELO Preprints), and grey literature 54 sources (Google and Google Scholar). No language restriction was applied. We planned to hire a 55 translation service if any study written in a language other than English or Spanish was found. The 56 search was restricted to include documents published in 2020 or 2021. 57 58 We included original studies, either published in scientific journals or with full text available in 59 preprint repositories, which reported original data and assessed the use of self-medication to prevent 60 or manage COVID-19 as a primary or secondary outcome, in participants of any age and from any 61

location. 62
We considered that a study assessed the use of self-medication in any of the following situations: 1) 63 the study reported explicitly that self-medication was assessed (regardless of the definition used), or 64 2) the study assessed the use of medications without medical prescription. 65

Study selection 66
Two authors (AQL and MHG) independently selected the studies. For this, they first screened the 67 search results by title and abstract according to the inclusion criteria using the web-based tool Rayyan 68 (http://rayyan.qcri.org). Later, they reviewed the full text of the relevant studies to determine whether 69 they were appropriate for study inclusion. Discrepancies were consulted with another author (ATR) 70 and resolved by consensus. 71 Two authors (CBI and RQE) independently extracted the relevant data for the review using a pre-73 piloted Microsoft Excel spreadsheet. Again, any discrepancies were discussed with another author 74 (ATR) and resolved with consensus. 75 The following variables were assessed for each study: author, year of publication, country, study 76 design, setting, population, sample technique, date of data collection, number of participants, age, sex, 77 how was self-medication assessed, the prevalence of self-medication, reasons for practicing self-78 medication, who recommended such medication, from where was the medication obtained, and 79 adverse events related to its use. were cross-sectional studies (Fig. 1). 96  (Table 1). 107 117 used a complex sample and such sampling was considered to estimate the self-medication prevalence.

Studies characteristics 123
Studies characteristics are shown in and one in workers from five sectors (health care, air transport, police, road transport, and informal) 137 The eight studies included between 132 and 3,792 participants. Five studies mentioned the summary 139 measure (mean or median) of the participants' age. Such summary measure (either mean or median) 140 ranged from 21 to 60.3 years. All studies were performed in adults. The percentage of females ranged 141 from 28.4% to 69.1% (Table 2). 142

Definition of self-medication 143
All studies assessed self-medication through self-report. The questions used to assess self-medication 144 were explicitly mentioned only in the one study [36], while other five studies report having asked for 145 the use of medicine without prescription [29,[31][32][33][34], and two studies gave no details regarding how 146 self-medication was measured in their surveys [30,35]

153
As showed in Table 2, We classified the eight included studies into three groups: 1) those that assessed 154 self-medication for COVID-19 prevention or management without focusing on a specific symptom, 155 and 2) those that assessed self-medication for a specific COVID-19 symptom. 156 Seven studies were included in the first group (assessed self-medication for COVID-19 prevention or 157 management without focusing on a specific symptom). Of these, four were performed in the general 158 population (with a range of self-medication prevalence between <4% and 88.3%) and three in specific 159 populations (which reported a self-medication prevalence of 33.9% for hospitalized adults with 160 COVID-19, 34.2% in adults who worked in five assessed sectors, and 51.3% in undergraduate students 161 of health-related careers). 162 Of these seven studies, six mentioned the list of medications reported by the participants. The reported 163 medications were different across the studies: six studies reported the consumption of antibiotics, five 164 of chloroquine or hydroxychloroquine, three of acetaminophen, three of vitamins or supplements, two 165 of ivermectin, and two of ibuprofen. 166 In addition, consumption preferences varied across studies. For example, when medications were 167 sorted in the order of highest to lowest consumption, vitamins/supplements occupied the first place in 168 two studies, but the fourth place in another study (Table 2). 169 The second group (studies that assessed the use of medication for a specific symptom) included one 170 study performed on the general population of Saudi Arabia, which found that 35.1% of the surveyed 171 individuals used self-medication for fever [36] (Table 2).

obtained, and adverse effects
Four of the eight included studies mentioned the reasons for self-medication. However, each study 175 assessed the reasons differently. Thus, one study referred only to the symptoms that motivated 176 participants to self-medicate, and the other three collected a plethora of reasons, including the fear of 177 stigmatization, fear of quarantine, affordability, the convenience of self-medication, and patients 178 believing that the symptoms were not severe (Table 3). 179 Only one study [34] specified the source of the patients' self-medications. This study was performed 180 on adults from Nigeria. The most common sources reported were the pharmacy (73.9%) and the patent 181 medicine vendor (23.6%). 182 Only two studies reported whether adverse events occurred in those who self-medicate. One study 183 performed on Undergraduate students of health-related careers found that 11.7% of those who self-184 medicate presented with side effects of self-medication in the past three months [32]. Nonetheless, this 185 study did not specify the adverse effects that occurred. The second study, performed in adults from 186 Nigeria, reported body rash (23.1%) followed by worsened condition (17.3%), yellowish eyes (7.7%), 187 swollen face (3.8%), and vomiting of blood and severe diarrhea (5.8%) [34]. 188  (2021) Fear of stigmatization or discrimination (79.5%), fear of quarantine or self-isolation (77.3%), fear of infection or contact with a suspected or known COVID-19 patient (76.3%), delay in receiving treatment at health facilities" (55.6%), influence of friends" (55.2%), social media (54.3%), nonavailability of medications for COVID-19 treatment in the health facilities (53%), emergency illness (49.1%), delay in getting hospital services (28.1%), distance to the health facility (23%), proximity of the pharmacy (21%), nonavailability of medicine in a health facility (19.3%), and health facility charges (15.3%) Pharmacy (73.9%), patent medicine vendor (23.6%), hospital (7.6%), hawkers (4.5%), faith-based outlets, and herbalists (2.1%, each) Body rash (23.1%) followed by worsened condition (17.3%), yellowish eyes (7.7%), swollen face (3.8%), and vomiting of blood and severe diarrhea (5.8%). Dare (2021) Self-medication is affordable (37%) and convenient (32%), lack the means to get to the health NR NR facility/hospital (15%), fear of being diagnosed COVID-19 positive (9%), fear of visiting health facility or hospital (7%) Miñan-Tapia (2020) Believing that the symptoms were not severe enough to go to a doctor (64.3%), referring to family/friends that are nonmedical health professionals (34.9%), and owing to economic reasons or use of over-thecounter medications (34.9%) NR 11.7% participants reported adverse events (not detailed) NR: not reported 191 Note: The percentage was only based on the people who self-medicated in each study. 192

193
To our knowledge, this is the first systematic review to assess the self-medication prevalence for 194 COVID-19 prevention or management. It assessed the most important databases and sources for grey 195 literature search and preprints repositories; without language limitations. Thus, our results should 196 reflect the state of knowledge until the search date. 197 Self-medication is a global phenomenon that may involve health risks at both the individual and 198 community levels [19,37]. Previous studies have found that self-medication is a common practice. 199 Three systematic reviews conducted in Iran [23], Ethiopia [25], and India [24]; reported a prevalence 200 of 53%, 44%, and 53.6%, respectively; for a recall period that ranged from a single day to 6 months. 201 In our review, we only found eight studies. Although the included studies were pe rformed in six 202 countries, the lack of studies in some regions such as North America, the Middle East, North Africa, 203 and Oceania is evident. It is expected that the intercountry differences in drug promotion, regulations, 204 and the possibility of accessing some medications without a prescription can influence the self-205 medication patterns [38,39]. Moreover, seven of the eight included studies were performed in low-and 206 middle-income countries. While self-medication may be higher in such countries, due to the conditions 207 and structure of the health system [40], it is not possible to make solid comparisons due to the 208 heterogeneity in the self-medication assessment across studies. 209 Some of the included studies assessed over-the-counter medications such as acetaminophen or 210 nonsteroidal anti-inflammatory drugs, which are used for the symptom management of 211 along with several other diseases. However, the included studies have also found an heterogeneous 212 prevalence of self-medication with medications that have not proven to benefit in the prevention or 213 management of COVID-19 (such as antibiotics, chloroquine or hydroxychloroquine, vitamins or 214 supplements, ivermectin, and antiasthmatics), as stated by the current COVID-19 guidelines by WHO 215 and IDSA [41,42]. 216 In fact, the adverse effects of some of these medications are of great concern, such as antibiotic 217 resistance due to extensive antibiotic use, bleeding caused by aspirin use, inhibition of the immune 218 system caused by corticosteroids, or arrhythmia caused by hydroxychloroquine [4]. Some healthcare 219 organizations have issued statements on self-medication. For instance, the WHO recognizes that a 220 "successful" (i.e., beneficial) self-medication can be achieved in many countries only by improving 221 individuals' knowledge and education level in such a way that the potential damages of this practice 222 can be avoided [16]. Similarly, the International Pharmaceutical Federation, in tandem with the World 223 Self-Medication Industry [43], and the World Medical Association [44], emphasizes the responsible 224 use of non-prescription medications. 225 Similar to that found in previous systematic reviews [20,21], the included studies in our review tended 226 to use different questions to assess self-medication, or they did not specify the question used. 227 Moreover, studies have established different recall periods, and even those that assessed self-228 medication since the beginning of the pandemic would have different time frames depending on the 229 month in which they were performed. This lack of clear information prevents direct comparisons 230 between studies and the subsequent meta-analysis of their results [45], adding a potential source of 231 bias that affects the internal validity of the results [45,46]. In addition, it is possible that studies that 232 evaluated self-medication through a pre-specified list of medications reported biased prevalences, 233 since these medications may not have included all the most commonly used medications [47]. 234 The included studies had an important risk of bias, mainly in the "representativeness of the sample" 235 domain of the NOS, the worst-rated item in the risk of bias assessment. This prevents the adequate 236 extrapolation of the results [48], which should be taken into account when interpreting the results of 237 this review. 238 adequately reported studies are needed, as well as the use of standardized self-medication definitions 240 across studies. Moreover, since the COVID-19 status is changing rapidly, future studies are required 241 to assess if there is any variation in self-medication trends between COVID-19 waves and after 242 introducing the massive vaccination. 243 The findings of this review suggest that there is an important though heterogeneous prevalence of self-244 medication to prevent or manage COVID-19. This includes some medications that have not shown any 245 benefit so far and may therefore expose people to unnecessary adverse events. The results vary widely 246 across studies, suggesting that each context (region, country) has its own self-medication patterns and 247 impact, calling on local health authorities to promote research and interventions to reduce potential 248 self-medication adverse consequences. 249

250
We identified eight studies that assessed the use of self-medication for the prevention or management 251 of COVID-19. Only one detailed the question used to assess self-medication. The recall period was 252 heterogeneous across studies. With respect to the seven studies that assessed self-medication without 253 focusing on a specific symptom: five performed in the general population (self-medication prevalence 254 ranged between <4% -88.3%) and three in specific populations. In these studies, the most used 255 medications varied widely. More well-designed and adequately reported studies are warranted in this 256 regard. 257 S1 Checklist. PRISMA 2009 checklist. 397 S1