Surveys of knowledge and awareness of antibiotic use and antimicrobial resistance in general population: A systematic review

Background Currently, various tools exist to evaluate knowledge and awareness of antibiotic use and antimicrobial resistance (AMR) and are applied by various organizations. Previous systematic reviews have focused mainly on study findings such as levels of knowledge and AMR awareness. However, the survey procedures and data instruments used ought to be scrutinized as well, since they are important contributors to credible results. This review aims to assess the study methods and procedures of existing population-based surveys and explore key components which determine the general population’s levels of knowledge and awareness of antibiotic use and AMR. Methods We searched existing literature for population -based surveys which sought knowledge and awareness of antibiotic use or AMR in the general population. Databases searched included Ovid, MEDLINE and EMBASE, PsycINFO and Scopus, domestic journals and gray literature sources. Population-based cross-sectional studies published in English or Thai from January 2000 to December 2018 were included in the review. Quality assessment was conducted using the ‘Appraisal Tool for Cross-Sectional Studies’ (AXIS). Results All 22 studies included in the analysis had clear objectives focusing on assessing people’s levels of knowledge, awareness, attitudes and behavior relating to antibiotic use and awareness of AMR. These studies had employed appropriate methodologies for population-based cross-sectional surveys relative to research questions. More than half of studies (14 out of 22) had scientifically soundly designed methodologies which captured the representativeness of the population; whereas the remaining studies had unclear sample size estimations, inappropriate sample frames and selection biases. Half of the studies had tested the validity and reliability of the questionnaire. The common questions used by these surveys were categorized into four themes: behavior related to antibiotic use, knowledge and awareness of antibiotic use, knowledge and awareness of AMR and others such as receiving information about antibiotic use and AMR or cross-cutting issues like self-medication. Conclusion This review identified four key features of good practices in antibiotic use and awareness surveys: a) clear survey objective; b) scientifically sound sampling techniques ensuring representativeness; c) strategies for recruitment of samples and survey administration methods; and d) credible measurement to prevent non-sampling biases. During questionnaire design, the health systems context in terms of access to health services and antibiotics should be taken into account. In conclusion, to maximize the use of surveys, the application of findings in surveys and associated factors related to antibiotic use and AMR should primarily generate public health interventions and target specific groups to make progress in solving AMR problems.


Results
All 22 studies included in the analysis had clear objectives focusing on assessing people's levels of knowledge, awareness, attitudes and behavior relating to antibiotic use and awareness of AMR. These studies had employed appropriate methodologies for population-based cross-sectional surveys relative to research questions. More than half of studies (14 out of 22) had scientifically soundly designed methodologies which captured the representativeness of the population; whereas the remaining studies had unclear sample size estimations, PLOS  Introduction Global efforts to assess public knowledge and awareness of antibiotic use and antimicrobial resistance (AMR) are underway. In 2015, the World Health Organization (WHO) developed a questionnaire survey to assess current public knowledge and awareness and behaviors related to antibiotic use in six WHO regions [1]. Similarly, multi-country surveys exist in Europe which use a common protocol, questionnaire and interview methodology [2,3,4]. These population-based surveys are part of the monitoring and evaluation framework proposed by the WHO Global Action Plan on Antimicrobial Resistance (AMR). In recent years the quest to halt AMR has been materialized in many countries. Thailand is amongst the exemplary countries that made substantial effort to counter AMR. One of the five goals of Thailand's National Strategic Plan on Antimicrobial Resistance (2017-2021) is to increase public knowledge and awareness of antibiotic use and AMR by 20% before 2021 [5]. The Thai working group on Health Systems and Policy Research on AMR has developed an AMR module and embedded it into the existing biennial Health Welfare Survey (HWS) conducted by National Statistical Office. The aims are to assess among Thai adults the volume of antibiotic use, levels of knowledge on antibiotic use and AMR, exposure to information related to antibiotic use and AMR, and awareness of the use of antibiotics in farm animals. The HWS in 2017 has provided a baseline level of knowledge and AMR awareness in adult populations as required by the National Strategic Plan for monitoring progress against the target [6].
Embedding an AMR module in national surveys has various advantages, such as the possibility for long-term monitoring and opportunities to assess factors associated with knowledge and awareness; it also saves costs compared with conducting independent surveys. Moreover, the merit of survey provides better understanding on behavioral pattern on antibiotic use (either misuse or overuse) in the population, which is one of the key contributing factors to the emergence of AMR. Survey information can serve as a basis to demonstrate an association between knowledge/awareness/practices on antibiotic use and AMR. Thus, it is necessary to establish suitable antibiotic use surveys with associated factors such as knowledge and awareness, in order to tackle with the rising of AMR trends.
However, the national survey on AMR almost always face some difficulties and limitations such as difficulties in the analysis to claim causal relationships, the presence of limited number of independent parameters and information bias when respondents do not understand the questions. This module was adapted from international survey tools such as the Eurobarometer survey in 2009, 2013 and 2016 [2,3,4] and the WHO tool [1], which also present challenges around generalizability and measurement bias, especially in different country contexts. Also, recently published systematic reviews relating to knowledge and awareness of antibiotic use and AMR focused on the results of the surveys [7,8] rather than presenting a review and recommendation of the survey methods and tools, which ultimately influenced the credibility of results.
In attempts to fill existing knowledge gaps pertaining to survey instruments, this review aims to assess the procedures of population-based surveys that ensures representativeness and minimizes biases. It also explores common contents in the questions used by these populationbased surveys and categorizes them into thematic areas. The review findings are useful for countries seeking to develop methods and tools to monitor population knowledge and awareness of antibiotic use and AMR in response to the Global Action Plan on AMR.

Material and methods
This review was registered with PROSPERO database (CRD42019123385) to review protocol: search strategy, inclusion and exclusion criteria, quality assessment and data extraction.

Search strategy
Search terms were developed along three domains: a) antibiotics or antimicrobial resistance; b) knowledge or awareness; and c) survey or questionnaire. Four international databases (Ovid MEDLINE, Ovid EMBASE, PsycINFO and Scopus) were searched using the search terms as detailed in Table 1. The search terms for international publications were applied to title, abstract, keyword, and full text. Three domestic journals (Health System Research Institute Journal, Thai Journal Citation Index Center, and Thai Journal Online) and gray literatures were manually searched.

Eligible criteria
The inclusion criteria comprised publications on population-based cross-sectional surveys in the general population which had investigated either knowledge or awareness of antibiotic use or AMR. Publications in English or Thai from international and domestic peer reviewed journals, and gray literature sources which were published between January 2000 and December 2018 were included. Studies on specific population groups, clinical research or studies which could not be electronically retrieved were excluded from the review.

Study selection
Four researchers (HK, SC, TI and WK) were responsible for abstract screening and full paper review for eligibility. Two researchers screened titles and abstracts to see if they met eligible criteria. The abstracts were included by a consensus between the two researchers and a third opinion was sought if they disagreed. The same process was conducted for the full paper review to select the papers related to research question in term of household-based surveys.

Data extraction
Data extraction was conducted into three sets: a) characteristics of studies: author, year of publication, objective, country, study design, sample size, eligible criteria, administration and tool development; b) themes emerging from common questions asked in the surveys to determine level of knowledge and awareness of antibiotic use and AMR or any relevant issues; and c) key findings in the studies.

Results
An electronic search comprising the four international databases and hand search of three domestic databases and international and domestic gray literature sources yielded a total of 2,761 records (2,740 from the databases and 21 from other sources). After duplicate removal, there were 2,663 papers for abstract screening. 2,537 papers were excluded as they were not relevant leaving 120 records to be searched for full texts. Thirteen full papers that were not electronically available and thirteen duplicates were excluded. 94 full papers were reviewed for eligibility and 72 publications were excluded for not being relevant or pertinent to the review objectives. Finally, 22 studies met the eligible criteria and were included for analysis. A PRISMA flow describing the study recruitment process of this systematic review is shown in Fig 1.
Regarding study design, 3 out of 22 studies were interventional studies which assessed the outcomes of campaign and educational interventions on the proper use of antibiotics [13,17,28]. The remaining 19 publications were observational studies. Only one study [15] was conducted by using mixed methods while the other 21 studies applied quantitative methods.

Quality assessment of 22 studies
The results of the quality assessment of 22 eligible studies by using the AXIS tool are shown in Tables 3 and 4.
Our analysis found that all publications had clear study objectives which focused on assessing the levels of knowledge and awareness or attitudes and behavior related to antibiotic use and AMR and associated factors.
Half of the studies [11,12,15,16,18,21,22,24,26,27,29] tested both the validity and reliability of the questionnaire and provided the statistical significance of key variables. Only one study [23] insufficiently described the method which was difficult to repeat.
Regarding the reporting of survey results, almost all studies presented adequate basic data and all results for the described methods, except three studies [1,4,23] which did not show basic data. The results in two studies [25,28] were not internally consistent and one study [1] could not be identified due to showing only percentage data. Vallin M. et al (2016) was the only study which addressed and categorized non-responders and which reported profiles of non-respondents to the survey. Widayati A. et al (2012) did not categorize data of non-  • 93% of the respondents knew the term "antibiotic".
• Penicillin was correctly identified as common antibiotic but 36% of the respondents incorrectly identified Benadryl (diphenhydramine), a common over-the-counter drug for cough and cold formulation, was antibiotic.
• Beta-lactams were the most frequently used antibiotics in the previous year, and 20% of antibiotics users used multiple antibiotics.
• A quarter of the respondents had asked a doctor for antibiotic prescriptions.
• 29% of the respondents said that antibiotics are drugs for treating bacterial infections.
• Respondents who have completed tertiary education (university) was significantly associated with correct knowledge of the safety of antibiotics. • 27% of the respondents believed that taking antibiotics when they suffered from a common cold made them feel better more quickly.
• 32% of the respondents believed that taking antibiotics helped preventing more serious illnesses.
• 48% of the respondents expected a prescription for antibiotics when they were ill.
• 58% of respondents were not aware of health dangers associated with taking antibiotics.
(Continued ) • 57% of the respondents used leftover antibiotics or took someone else's advice.
• 34% of the respondents were not aware of the dangers of antibiotics.
• Males, younger age, and those with low and medium income and lower level of education showed poorer practice and lower level of knowledge on antibiotics and awareness of the health dangers associated with antibiotics.
(Continued ) • About 8% of the respondents had ever acquired non-prescribed antibiotics.
• About 7% of the respondents had ever kept the leftover antibiotics while around 70% of the respondents had always finished the full course of antibiotics.
• About 75-77% of the respondents agreed that the purchase of antibiotics without prescription and incomplete courses of antibiotics will lead to undesirable consequences.
• About 39% of the respondents agreed that they could help the prevention of resistance.
(Continued ) • 64% of the respondents incorrectly believed that viruses such as colds and flu can be treated with antibiotics.
• 87% agreeing that people should use antibiotics only when prescribed.
• 32% of the respondents think that they should stop taking antibiotics when they feel better.
• 72% of the respondents correctly believed that many infections are becoming increasingly resistant to treatment by antibiotics.
(Continued ) • About 79% of the respondents reported that antibiotics used to treat bacterial infection while about 53% of those reported that antibiotics used to treat viral infections.
• About 62% of the respondents were aware of antibiotic resistance in relation to the overuse of antibiotics.
• About 35% of the respondents reported that when they got cold, they would take antibiotics to help them get better more quickly.
• Education levels, healthcare-related occupation and family's occupation related to healthcare were significantly associated with knowledge of antibiotics.
• Healthcare-related occupation, marital status and income were significantly associated with better attitude.
• Knowledge score was positively associated with attitude score. and sore throat (14%).
• 43% of the respondents knew that antibiotics are ineffective against viruses and 56% of those knew that antibiotics are also ineffective against colds and flu.
• 84% of the respondents knew that unnecessary use of antibiotics makes them become ineffective.
• 82% of the respondents think they should stop taking antibiotics once they begun a course of treatment.
• 33% of the respondents remembered that they received information about the unnecessary use of antibiotics in the last 12 months; mostly from health professionals.
Regarding conflict of interests and ethical reviews, five studies did not clearly declare funding sources which might influence authors' interpretation of the results [1,4,12,23,27]. Seventeen studies indicated they had ethical approval or consent of the participants. Another five studies [1,4,6,12,13] did not provide information on ethical clearance or whether they attained consent of the survey participants. Three studies [6,13,21] declared that their studies were exempted from ethical review.
According to AXIS quality assessment, the tool does not provide a numeric scale for assessment, but it is flexible enough for users to judge the quality of the paper overall. However, authors in this systematic review classified all papers into three types based on methodology, results and discussions: 1) fully qualified; 2) partly qualified; and 3) unqualified. Fully qualified means the studies are qualified in all parts; there are no studies which reach this qualification. Partly qualified considers the studies that have qualifications in some parts; there are six studies in this group [11,15,16,18,21,27]. Four studies [15,16,18,21] are not fully qualified in methodology while the other two studies [11,27] are not fully qualified in discussion. Kim  Sixteen studies are unqualified because they are missing important parts of quality assessment; for example, some studies had inappropriate selection processes influencing representativeness and some studies did not have validity and reliability tests of measurements.

Thematic concerns of questions in the questionnaire survey
Of the 22 studies, 13 [6,11,16,17,18,20,21,22,24,26,27,28,29] adapted a questionnaire from prior studies, and the questionnaire for household-based cross-sectional surveys in general population from the Eurobarometer survey (2013,2014,2016 Note: Y = Yes, N = No, CT = Cannot Tell � Exemption for ethical approval �� Only consent of respondents attained ��� This study did not categorize non-responders but it compared the characteristics of respondents who were familiar and were not non-familiar with antibiotics. https://doi.org/10.1371/journal.pone.0227973.t004 Systematic review of surveys of knowledge and awareness of antibiotic use and AMR in general population Four themes emerged from the analysis of the contents of the questionnaire: a) behavior related to antibiotic use; b) knowledge and awareness of antibiotic use; c) knowledge and awareness of AMR and d) other issues such as receiving information and advice about proper use of antibiotics, or AMR campaign message and cross-cutting issues such as self-medication. See Table 5.
With regard to behavior-related antibiotic use, we identified four sub-themes covering: 1) frequency of using antibiotics in the recall period such as one month, six months or a year; 2) source of antibiotics; 3) clinical indications or conditions for which antibiotics are used; and 4) instruction and advice from drug sellers or pharmacists on the proper use of antibiotics. Table 5. Common questions used to determine level of knowledge and awareness of antibiotic use and AMR.

Subthemes Common questions/statements
Behavior related to antibiotic use Frequency of using antibiotics Have you taken any antibiotics in the last one month or 12 months?
Source of antibiotics How do you obtain the antibiotics?
Indication/reason of antibiotic use What was the reason for last taking the antibiotics that you used?
For knowledge and awareness of antibiotic use, three sub-themes emerged: 1) antibiotic names; 2) general knowledge; and 3) awareness of using antibiotics in common cold and flu symptoms. Questions were asked about respondents' recognition of antibiotics, for example whether penicillin or tetracycline were antibiotics or not. In terms of general knowledge, questions were asked about the mechanism of action of antibiotics, such as its action towards bacteria or viruses, its side-effects and allergies, and inappropriate antibiotic use. Finally, concerning awareness of antibiotic use, questions were designed to explore opinions about antibiotic use for common cold and flu symptoms.
On knowledge and awareness of AMR, various questions in the survey tools were categorized into three subthemes: 1) definition of AMR; 2) general knowledge about AMR; and 3) awareness of AMR. The general knowledge questions focused on misuse, overuse, sub-optimal use and inappropriate use of antibiotics, which could lead to AMR and the spread of resistant bacteria. Concerning awareness of AMR, various questions explored people's concerns about AMR, which had the potential to affect themselves, their families and countries.
Self-medication with antibiotics is a cross-cutting issue in all the three thematic areas. Questions explored the necessity of completing the full course of antibiotics and proper management of the leftover antibiotics.
Additional questions explored exposure to public information relating to proper use of antibiotics and AMR. These included media channels and sources of information such as health professionals, and the impact of this information on people's behavior in relation to antibiotic use. For doctor-patient relationships, the questions related to trust and communication between people and healthcare providers.

Discussion
The systematic review observed several important features in design and methodology of included studies that would be useful for developing a tool to determine levels of knowledge and awareness of antibiotic use and AMR.
Setting objectives is vital to guide study design and all included studies had clear objectives focusing on assessing levels of knowledge, awareness or attitudes and behavior related to antibiotic use and awareness of AMR and associated factors. A cross-sectional survey is appropriate for the assessment of population knowledge about and awareness of proper use of antibiotics under the resource constraints. It measures exposure and outcomes at the same time and can find possible associations between exposure and outcomes [30]. Cross-sectional surveys are less costly and less time-consuming than longitudinal studies [30]. However, the casual relationships are better identified through longitudinal studies where temporal relationship can be addressed [31]. A recent systematic review on public knowledge and beliefs about AMR has shown that synthesis of qualitative and quantitative studies provided more in-depth understanding of people's knowledge and beliefs about AMR than using quantitative data alone [8]. In this review, the number of quantitative studies was three times higher than qualitative studies and mixed methods. Due to the strengths and limitations of each method, quantitative studies, especially cross-sectional surveys, are more appropriate for population-based surveys while qualitative methods are useful for in-depth explanation in small-scale researchbased assessments.
Although various methods can be used for sampling and recruitment, the key strengths of household-based cross-sectional surveys is the representativeness of the population. Although the sizes of samples are usually limited by the budget available for very large surveys, a representative sampling frame is essential for generalization of the survey findings to the population [32]. Inappropriate sampling frames were seen in the studies conducted in Lebanon and Syria [20,26]). In the Lebanon study, which aimed to assess knowledge, attitudes and practice of antibiotic use in the Lebanese population, the sampling frame was the population in the capital city which did not therefore represent the whole population. In the Syrian study, which aimed to provide an insight of the current knowledge and practices regarding antibiotic use among individuals living in the Kalamoon, Syrian Arab Republic, the sampling frame was of households in main streets of two cities which therefore missed some samples for representing the whole population.
We acknowledge that while random sampling is ideal as it properly represents the population, it is time-and resource-consuming. Stratified random sampling and cluster random sampling can be applied to household-surveys as these methods can also achieve representativeness and reduce selection bias. Cluster random sampling is also less costly and feasible; it is a common method used by many studies [33].
Recruiting samples such as adult members or those who have clear understanding of the language used in surveys is critical for ensuring high-quality responses in many surveys. However, specific sampling methods may introduce selection biases, which should be considered before setting inclusion and exclusion criteria.
The high level of non-response rate such as refusals, unreachable households or incomplete data compromises the validity of survey results and conclusions [10].
Although two studies had low response rates, only Vallin M. et al (2016) mentioned this consideration in the discussion section. Various measures can be applied to minimize nonresponse errors such as making appointments for follow-up interviews for those who were absent on the interview days or using combined user-friendly survey instruments such as faceto-face interviews, telephone, mail or online self-administered surveys [32]. Even where there is high response rate, the non-responders profiles such as those are very high or very lower users of antibiotics; this non-respondent bias can affect the validity of findings about the prevalence of antibiotic use in the population. Almost all studies did not address and describe the profiles of non-responders in their studies which therefore affected the credibility of results. Parameters about non-respondents should be recorded during the field survey and analyzed to verify if the non-responders are similar or dis-similar to the responders.
Two broad methods of questionnaire administration are identified in this review: a) interview survey, either face-to-face or the use of telephone by trained interviewers; and b) selfadministration either through postal or internet methods. Using multiple survey methods, when no single method is adequate to address research objective, can minimize the low response rate, prevent coverage, measurement and non-response errors [32]. Each method may have its advantages and disadvantages. For example, self-administered surveys present challenges of interpreting questions as it is "one-way communication" which can introduce measurement error. Face-to-face interviews can prevent measurement bias.
Many studies addressed limitations about coverage errors and measurement errors. In term of coverage errors, Andre´M. et al (2010) addressed the fact that 6% of the Swedish population aged 16-75 years did not have a fixed telephone line in 2006. Parimi N. et al (2002) also mentioned that 10% of the households in Trinidad and Tobago did not have telephone service and that 15% of the Telecommunication Services customers have unlisted telephone numbers. However, the limitations of questionnaire administration depend on the context specific to each country.
In relation to measurement errors, Parimi N. et al (2002), Eng JV. et al. (2003) and Barah F. and Goncalves V. (2010) raised concerns about the level of understanding as regard to the questionnaire such as the term 'antibiotics' or explanation about illness and treatments. Therefore, some studies reduced these errors by setting criteria to recruit respondents who understood the term 'antibiotics' or to those who had used it before. Nevertheless, selection bias should be taken into consideration when studies select based on these specific groups. Interestingly, findings from Widayati A. et al (2012) showed that the characteristics data from groups of responders who were and were not familiar with antibiotics, were not significantly different. It means using this inclusion criterion was useful for ensuring the validity of the questionnaire.
Tailor-made design in line with country contexts is important. For example, in countries with a high prevalence of "polypharmacy" which means using multiple drugs to treat a single ailment or condition at the same time, a careful design is needed to ensure correct interpretation of respondents' understanding and their ability to distinguish antibiotics from vitamins and analgesic they use. Antibiotics are one of the most common items in polypharmacy, which can cause serious adverse drug events or drug interactions [34]. Additionally, to assess the effectiveness of antibiotic awareness campaigns, the surveys should align with the campaign's contents as seen in these two studies: Mazińska B. et al (2017) and Haenssgen JM. et al (2018).
With regard to the validity and reliability of measurement, half of these 22 studies did not report testing validity and reliability before finalizing the questionnaire. Although some questionnaires were developed by other studies, the validity and reliability test are still essential because of the difference in population, health systems, culture and terminology for which adaptation to local contexts would be required.
Key findings from these studies showed the differences in prevalence regarding antibiotic use, levels of knowledge of antibiotics and awareness of AMR, frequency and sources of receiving information about antibiotic and its use and AMR. However, there is no study expounding on the outcomes of surveys, including further implications such as impacts on AMR trends. All studies tried to identify the gaps in low levels of knowledge and awareness in terms of contents and characteristics of population in these groups. Furthermore, almost all studies assessed the association between demographics, for example, sex, age, education levels, wealth status with levels of knowledge and awareness, practices, regarding to antibiotic use and AMR except Ahmad H. et al (2013). According to findings, education levels were proven as consensual factors associated with knowledge and awareness while other factors differed depended on each study. Some studies found the significant association between key variables such as exposure to information or campaigns with levels of knowledge and awareness [21], and level of knowledge with level of awareness [16,22,27]. Nevertheless, the relation between levels of knowledge and awareness and antibiotic use remained unclear [11,13] and there were no studies linking the findings with AMR trends. In discussion section, all studies recommended the enhancing of knowledge and awareness from key findings. Most common sources of information are from health workers so they should be key actors in promoting appropriate antibiotic use [14].
Aligned with global action plan on AMR, surveillance on levels of knowledge and awareness can contribute to the design of interventions which can change the population's behavior on antibiotic use which could potentially lead to a reduction in AMR [35]. Nonetheless, the majority of the reviewed studies encountered limitations in demonstrating an association between knowledge/awareness/practices and the emergence of AMR in the community. Only two studies described the association between knowledge and attitudes, and practices of antibiotic use-showing the association between these factors [11,13]. If a novel antibiotic survey is to be implemented in order to reduce inappropriate antibiotic use, standardized questions on knowledge, awareness and practices on antibiotic use should be focused.
To change pattern of inappropriate uses in the population, it is necessary to have strategies or policies developed based on survey evidences. Experts and academics in the field should discuss and reach consensus on the required questions in the AMR survey module. The critical point is the linkage between levels of knowledge and awareness to behaviors which are influenced by various factors such as access to healthcare or social and cultural aspects [36]. The surveys can be a surrogate measure used to probe into possible causes of the problem. Survey evidence can be used for public advocacy. However, among these studies, only five studies linked their surveys with communication campaigns and evaluate policies [4,6,13,17,28].
The strength of this study is the focus on evaluation of questionnaire tools in householdbased cross-sectional surveys. The systematic review contributes to new knowledge about the monitoring of knowledge and awareness of antibiotic use and AMR in two key areas. Firstly, it provides quality assessment of these cross-sectional surveys, which is important for tool development and data collection. We find that AXIS is a useful tool which provides qualitative assessment for the review of survey methodologies [9,10]. Secondly, the previous reviews focused on results of studies; this study fills review gaps by looking at the main contents of the questions that were asked by these 22 survey instruments and associated factors related to knowledge, awareness and practice of antibiotic use and AMR.
However, there remain some limitations. For instance, firstly, despite the authors widened the search strategy as large as possible; it is very likely that some studies had been left out, particularly the gray literature in the archives of domestic universities or research institutes. Secondly, this review was unable to capture the linkage between AMR tools and the actual knowledge and behavior of antibiotic use in the wider population. This issue cannot be addressed by the review; primary data collection through either quantitative survey or qualitative interview is needed. In addition, a more complex review design (for example, realist review) [37] are likely to be beneficial to answering this question. Future systematic reviews that explore the tools in specific subpopulations, such as health professionals, patients, and general populations, are of huge value in the AMR field.

Conclusion
In response to AMR threats, countries need to assess their population's knowledge and awareness of antibiotic use and AMR. Valid household-based assessments require clear survey objectives, valid and reliable tools for measurement, representativeness for generalizing the survey findings to the population and minimize sampling and non-sampling biases.
The survey design needs to take into account local contexts and terminologies related to medicines, antibiotics and disease conditions used by the communities, and recruit qualified respondents who can provide accurate responses representing the population. Common questions in existing household-based surveys cover four thematic areas: behavior related to antibiotic use, knowledge and awareness of antibiotic use, knowledge and awareness of AMR and others such as receiving information about antibiotic use and AMR or cross-cutting issues like self-medication.
Countries can learn from previous survey instruments applied by other and avoid mistakes. Accurate survey tools contribute to valid evidence which can be used to inform policies for specific interventions to improve population knowledge and awareness on antibiotics and AMR. The country-specific health system context of access to health services and antibiotics should be taken into account in the design of the survey questionnaire. Identifying levels of knowledge and awareness of antibiotic use and AMR is crucial. Eventually the utmost goals of such surveys would be to enhance the application of this knowledge to target specific target groups as well as to generate public health interventions related to antibiotic use and mitigating AMR.