Awareness of Antibiotics and Antibiotic Resistance in a Rural District of Ha Nam Province, Vietnam: A Cross-Sectional Survey

Low awareness of antibiotics and antibiotic resistance may lead to inappropriate antibiotic use and contribute to the problem of antibiotic resistance. This study explored levels and determinants of antibiotic awareness in a rural community in northern Vietnam, through a cross-sectional survey of 324 households in one commune of Ha Nam Province. Awareness and knowledge of antibiotics and antibiotic resistance and determinants were evaluated using structured questionnaires. Most respondents (232/323 (71.8%)) had heard of antibiotics, but fewer could name any antibiotic (68/323 (21.1%)) or had heard of antibiotic resistance (57/322 (17.7%)). In adjusted regression models, antibiotic awareness was lower among those who lived further from health facilities (Odds Ratio (OR): 0.08; 95% Confidence Interval (CI): 0.04–0.19) but higher among those who used interpersonal sources for health information (OR: 4.06; 95% CI: 1.32–12.46). Antibiotic resistance awareness was lower among those who used private providers or pharmacies as their usual health facility (OR: 0.14; 95% CI: 0.05–0.44) but higher among those with medical insurance (OR: 3.70; 95% CI: 1.06–12.96) and those with high media use frequency (OR: 9.54; 95% CI: 2.39–38.07). Awareness of Antimicrobial Resistance (AMR) was also higher among those who sought health information from official sources (OR: 3.88; 95% CI: 1.01–14.86) or had overall high levels of health information seeking (OR: 12.85; 95% CI: 1.63–101.1). In conclusion, communication interventions need to target frequently used media platforms, such as television, as well as key health information providers, such as health workers, as channels for increasing knowledge and changing community antibiotic use behaviour.


Introduction
The discovery of antibiotics a century ago, followed by a golden age of antibiotic discovery between the 1940s and 1960s, revolutionised medical care; however, the evolution of resistance now threatens modern medicine [1,2]. Antibiotics are instrumental in treating primary bacterial infections as well as reducing infections that occur due to procedures such as surgery, kidney dialysis, organ transplantation, and cancer treatment. As bacterial resistance to antibiotics increases, community-acquired infections and particularly infections following routine medical and surgical procedures are becoming more difficult to treat [3,4]. Antibiotic resistance has been associated with increased morbidity and mortality, a longer hospital stay, and higher hospital costs [5]. An estimated 1.27 million deaths occurred due to antibiotic resistance in 2019 [6]. Without additional interventions, it is estimated that by 2050, there will be 10 million deaths annually due to antibiotic resistance, surpassing projected deaths due to cancer, and [7] the economic impact of antibiotic resistance will be USD100 trillion. Antibiotic resistance disproportionally affects Low-and Middle-Income Countries (LMICs), with higher burdens of infectious disease, poorly regulated antibiotic supply, weak hospital infection prevention and control, and limited access to expensive treatment alternatives and third-line antibiotics [8]. Southeast Asia is considered one of the hotspots for antibiotic resistance selection, emergence, and transmission, and Vietnam has among the highest proportions of resistant pathogens in Asia [6,9].
Consumer practices and behaviours are key factors accelerating antibiotic resistance globally. Non-prescription antibiotic use is high in Vietnam, with 90% of antibiotic sales in pharmacies being made without a prescription; this is driven strongly by consumer demand [10]. Consumer demand for antibiotics is underpinned by limited knowledge of antibiotics and antibiotic resistance. A study conducted in the United States of America (USA) in 1999 found that 27% of the sampled population believed that taking antibiotics when they had a cold would make them recover more quickly, whereas a World Health Organisation (WHO) survey conducted in Vietnam in 2015 indicated that 62% of the respondents believed that antibiotics could cure a cold [11,12].
Interventions that target community antibiotic use awareness have been shown to have a positive impact, although much of this evidence is drawn from high-income countries. For instance, in the USA, community educational campaigns that aimed at raising antibiotic awareness among parents and clinicians resulted in a larger increase in knowledge and awareness regarding appropriate antibiotic use than in control areas [13]. A systematic review conducted in 2017 that included studies from the USA and Europe, recommended that components of multifaceted communication interventions that target both the general public and clinicians could reduce antibiotic use in high-income countries [14].
Few studies on antibiotic awareness have been conducted in LMICs, particularly in rural communities. A 2018 study on antibiotic awareness in five provinces of the central highlands of Vietnam reported that 67.4% of participants had heard of antibiotics, whereas 55.8% had heard of antibiotic resistance [15]. However, adequate knowledge of antibiotics and antibiotic resistance was limited, with only 18.8% being aware that antibiotic use could lead to antibiotic resistance. Another national study conducted in 2019 identified higher education and higher income as being associated with higher antibiotic knowledge; however, this sample was not representative of the general population [16]. A better understanding of community knowledge is needed to inform the development of community-targeted interventions in LMICs, including Vietnam [17].
The main objective of this study was to assess the levels of awareness and knowledge of antibiotics and antibiotic resistance in a rural community in northern Vietnam and to investigate the determinants of awareness and knowledge, in order to inform the development of interventions.

Results
After stratified random sampling, 389 households were selected for data collection, of which 324 households participated, giving an overall response rate of 83.3% and a margin of error of 5.4% for prevalence estimates. Of those that did not participate, three household respondents refused, whereas the other 62 households had no available respondents found during the study period ( Figure 1).
Individual respondent demographic characteristics are summarized in Table 1. There were more female respondents (84.6%) than male respondents. The younger population (18-29 years) were the least represented (16.4%), whereas those above 50 years old were the most represented (43.8%). Most respondents had attended school at some level (81.3%). About two-thirds of respondents (65.8%) were farmers, whereas 28.5% had other types of employment, and the remainder were unemployed. gin of error of 5.4% for prevalence estimates. Of those that did not participate, three household respondents refused, whereas the other 62 households had no available respondents found during the study period ( Figure 1). Individual respondent demographic characteristics are summarized in Table 1. There were more female respondents (84.6%) than male respondents. The younger population (18-29 years) were the least represented (16.4%), whereas those above 50 years old were the most represented (43.8%). Most respondents had attended school at some level (81.3%). About two-thirds of respondents (65.8%) were farmers, whereas 28.5% had other types of employment, and the remainder were unemployed.  Access to different media platforms is shown in Figure 2. Television (TV) was the most accessed (97.8%), whereas print media was the least accessed (14.8%). Health workers (97.2%) and television (95.4%) were the most consulted sources o health information, whereas newspapers (13.3%) and books (15.2%) were the least con sulted (Figure 3). Though the commune was small (approximately 9 km 2 and not mor than 5 km at the widest part) and relatively flat, almost half of respondents (48.6%) sai  Health workers (97.2%) and television (95.4%) were the most consulted sources of health information, whereas newspapers (13.3%) and books (15.2%) were the least consulted ( Figure 3). Though the commune was small (approximately 9 km 2 and not more than 5 km at the widest part) and relatively flat, almost half of respondents (48.6%) said they lived more than 10 min travel time to the commune health centre. These respondents were more likely to be low frequency media users (p < 0.001) and low health information seeking types (p < 0.001). Health workers (97.2%) and television (95.4%) were the most consulted sources of health information, whereas newspapers (13.3%) and books (15.2%) were the least consulted ( Figure 3). Though the commune was small (approximately 9 km 2 and not more than 5 km at the widest part) and relatively flat, almost half of respondents (48.6%) said they lived more than 10 min travel time to the commune health centre. These respondents were more likely to be low frequency media users (p < 0.001) and low health information seeking types (p < 0.001). Nearly three quarters of primary respondents had heard of antibiotics (232/323 (71.8%)), but fewer (68/323 (21.1%)) could name any of the 11 antibiotics on our list spontaneously when asked to give the names of any antibiotics they knew. The most well- Nearly three quarters of primary respondents had heard of antibiotics (232/323 (71.8%)), but fewer (68/323 (21.1%)) could name any of the 11 antibiotics on our list spontaneously when asked to give the names of any antibiotics they knew. The most well-known antibiotic was ampicillin/amoxicillin, with only (4.2%) having never heard of it, whereas 98.4%, 94.1%, and 92.3% had never heard of colistin, ciprofloxacin, and Augmentin, respectively, even after probing ( Figure 4). About one-fifth (57/322 (17.7%)) of respondents had heard of antibiotic resistance. known antibiotic was ampicillin/amoxicillin, with only (4.2%) having never heard of it, whereas 98.4%, 94.1%, and 92.3% had never heard of colistin, ciprofloxacin, and Augmentin, respectively, even after probing ( Figure 4). About one-fifth (57/322 (17.7%)) of respondents had heard of antibiotic resistance. In unadjusted models (Table 2), male respondents had lower awareness of antibiotics than female respondents (Odds Ratio (OR): 0.30; 95% Confidence Interval (CI): 0.13-0.73), those who were not working had lower awareness than those who were employed (OR: In unadjusted models (Table 2), male respondents had lower awareness of antibiotics than female respondents (Odds Ratio (OR): 0.30; 95% Confidence Interval (CI): 0.13-0.73), those who were not working had lower awareness than those who were employed (OR: 0.14; 95% CI: 0.03-0.60), those who used traditional practitioners had lower awareness than those who used government health facilities (OR: 0.23; 95% CI: 0.10-0.57), and those who had to travel more than 10 min to the commune health centre had lower awareness than those who had to travel less than 10 min (OR: 0.06; 95% CI: 0.02-0.15). Those who had attended school had higher awareness than those who had not or whose educational level was unknown (OR: 9.99; 95% CI: 3.90-25.60) ( Table 2). After adjusting for age, sex, and education, respondents who travelled more than 10 min to their commune health centre were less likely to be aware of antibiotics than those who travelled less than 10 min (adjusted OR (aOR): 0.08; 95% CI: 0.04-0.19), and those who sought health information from health workers (aOR: 172.78; 95% CI: 13.49-2213.05), the women's union (aOR: 5.08; 95% CI: 1.71-15.09), and relatives (aOR: 4.08: 95% CI: 1.42-11.74) were more likely to be aware of antibiotics than those who did not. In line with this, respondents who used interpersonal sources of health information were more likely to be aware of antibiotics compared with those with low overall health information seeking behaviour (aOR 4.06, 95% CI 1.32-12.46). Crude and adjusted linear models using the continuous antibiotic knowledge score based on the number of antibiotics recognised, showed qualitatively similar results. Exceptions were that those in the richest household tertile were familiar with more antibiotics than those in poorer households, as were those who used private providers, pharmacies, and drugstores. Those who used television, newspapers, health-workers, and pharmacists as regular health information sources were more familiar with antibiotics than those who did not use these sources. Media use frequency was not associated with antibiotic awareness or knowledge of antibiotics in adjusted models.
Determinants of antibiotic resistance awareness are presented in Table 3. In crude models, older respondents, farmers, and those who were not working were less likely to have heard of antibiotic resistance. After adjusting for age, sex, and education, use of private healthcare, pharmacies, or drugstores was associated with less awareness of antibiotic resistance than the use of government facilities (aOR: 0.14; 95% CI: 0.05-0.44), and in line with this, having a national medical insurance card (to be used at government facilities) was associated with higher awareness of antibiotic resistance (aOR: 3.70; 95% CI: 1.06-12.96). Those with high levels of media use also had higher levels of awareness (aOR:9.54; 95% CI: 2.39-38.07), as did those with high levels of health information seeking (aOR: 12.85; 95% CI: 1.63-101.10), and those that sought health information from official sources (aOR: 3.88; 95% CI: 1.01-14.86) had higher awareness of antibiotic resistance than those with low levels of health information seeking. Specifically, those who sought health information from radio, newspapers, community radio, and social media had higher awareness of antibiotic resistance. In linear regression models based on scores from questions about antibiotic resistance, risk factor patterns were qualitatively similar, with the exception that male respondents had lower scores on antibiotic resistance knowledge and higher household wealth was associated with higher knowledge. A high frequency of media use and higher levels of health information seeking were also associated with higher knowledge, but official health information sources were not associated with higher knowledge. Those who sought health information from television, radio, newspapers, community radio, health workers, the women's union, relatives, and social media had higher antibiotic resistance knowledge scores.

Discussion
This study shows that the overall awareness of antibiotics in this rural community in Vietnam is high (71.8%). However, the spontaneous recall or recognition of names of antibiotics from our list was much lower (21.6%). This inconsistency raises the possibility that respondents' understanding of the word "antibiotic" when reporting awareness was not in line with its actual meaning, as has been observed in other studies. For example, a study in Malaysia in 2010 found that 33% of the respondents confused antibiotics with painkillers such as paracetamol and aspirin [18]. A similar finding was reported in a community-based antibiotic access and use study conducted in six LMICs, including Vietnam [19]. Furthermore, differences between the determinants of antibiotic awareness and antibiotic knowledge scores suggests that having heard of antibiotics in general and more in-depth knowledge in terms of being able to name any of the common antibiotics might be learnt through different sources of information.
This study also established that antibiotic resistance awareness is low (18.2%) in this community and driven by similar demographic and media access factors as antibiotic awareness and knowledge. The level of antibiotic resistance awareness is lower than in a recent study in Vietnam, which reported an awareness of antibiotic resistance of 55.8% in a mixed urban¬-rural population; the difference may be due to our study targeting a rural population [15]. A similar antibiotic awareness study in Thailand conducted between 2017 and 2018 reported a higher antibiotic awareness of up to 95.7% and antibiotic resistance awareness of 74.8%. However, the methodology of measuring this awareness was different; the Thailand study relied on the recognition of pictures of antibiotics [20]. Another possible reason for the high antibiotic awareness in Thailand could be the robust systematic antimicrobial resistance campaigns that preceded the survey from 2012 to 2016 [21]. These studies and their distinct results demonstrate the diversity of approaches to measuring antibiotic awareness and knowledge and the subsequent challenges in comparing different study findings.
Factors shown to be associated with awareness of antibiotics in this study, such as demographic characteristics, type of primary facility, location, and use of media, have been reported before [22,23]. Our study found that those who use private clinics, pharmacies, or drug stores as their usual healthcare provider were familiar with more antibiotics than those who used government facilities, but were less aware of antibiotic resistance and scored lower on the antibiotic resistance knowledge questions. This finding aligns with the commercial orientation of most retail pharmacies and drug stores, particularly in LMICs, where antibiotics can easily be obtained without a prescription. Drug sellers may fail to provide adequate advice about a drug when making a sale, whereas buyers may request antibiotics based on prior experiences or advice from friends or relatives [24]. Thus patients may be more familiar with the names of the products they wish to buy, or that drug sellers wish to sell to them, but have less comprehensive information about the negative effects of these drugs. However, as this was a cross-sectional study, we were unable to establish the direction of causality between awareness and knowledge of antibiotics and antibiotic resistance and healthcare seeking.
In our study, television was the most frequently accessed form of media and also the source used most often for health information. Health workers and community radio were also frequently used sources of health information. We found an association between media use and increased awareness of antibiotics and antibiotic resistance. Other media platforms such as the internet, print media, radio, and SMS are also known to be common sources that individuals use to access information regarding antibiotics. A study in Poland found that the majority of respondents relied on the internet as the source of antibiotic knowledge [25]. Limited research has been performed in LMICs on the role of internet use in raising antibiotic awareness. However, the growing internet use in LMICs, including Vietnam, offers an opportunity for leveraging this platform for raising antibiotic awareness [26]. Our study only explored sources of general health information and which of these media sources were the trusted or primary source of information about antibiotics and antibiotic resistance, and the content of their messages remains to be investigated in this population. Research investigating media representations of AMR in the UK and China suggests that media content needs to be reoriented to communicate actions consumers could take to tackle AMR [27].
Although the study area was small and relatively flat, we found that living more than 10 min of travel time from the usual health facility was associated with lower antibiotic awareness. Further investigation into the relationship between distance from health facility and access to health information revealed that those who lived more than 10 min travel time away were low frequency media users and low health information seeking types. This finding, along with the finding that health workers were important sources of both antibiotic and antibiotic resistance knowledge, highlights the importance of health facilities as sources of health information [28].
This study is among the few studies in LMICs that have evaluated the antibiotic awareness within the general public and particularly rural communities. Moreover, this study incorporated diverse measurement approaches to triangulate antibiotic awareness. The response rate is also relatively high and therefore the risk of selection bias is reduced.
This study has several limitations. The results represent one rural community in northern Vietnam and may not be generalizable to other settings in Vietnam. The cross-sectional design limits any conclusions regarding the direction of causal effects. Furthermore, we cannot rule out the possibility of various information biases. For instance, even though data collectors were trained, the interviewers were nurses from the local health centre and may have interpreted the responses using their health knowledge. Respondents may have had difficulty recalling some information related to health-seeking behaviours, and this might have introduced recall bias. Respondents may also have introduced bias by reporting what they thought were desirable/socially-acceptable behaviours. Antibiotics may be known by a mixture of generic and brand names, and the list of antibiotic names we provided might not have been the most commonly used in the area. We are conducting an ethnographic study to explore understanding of antibiotics and the language used to describe them that will aid interpretation of these results and inform future research in this area. Lastly, there might be other unmeasured factors that influence awareness of antibiotics and antibiotic resistance or confound the observed associations.

Study Design
The study was part of a large quantitative cross-sectional household survey conducted with 324 households in a rural commune of Ha Nam Province, northern Vietnam, between 16 July 2018 and 9 April 2019. The commune had a population of 9746 people. The study design has been described in detail elsewhere [29].

Sample Size
A random sample of households was selected from a household list obtained from the commune health centre. Sampling was stratified so that households with children below five years were oversampled, as one of the study's primary aims was to collect samples and conduct antimicrobial susceptibility testing for Streptococcus pneumoniae, which is carried at higher levels among children under five years old. A sample size of 340 would allow us to estimate prevalence with a 5.2% margin of error with 95% confidence. We sampled 390 households, allowing for a 15% drop-out rate. Nurses from the commune health centre were trained to administer structured questionnaires (in Vietnamese) and to collect information on healthcare seeking behaviour, awareness of antibiotics and antibiotic resistance, and general socio-demographic characteristics. For every household, one adult caregiver responded to the questionnaire. Eligibility depended on giving written informed consent to the study and study procedures, and there were no exclusion criteria for this study.

Definitions of Variables
The definitions of dependent and independent variables are provided in Table 4. Awareness of antibiotics and antibiotic resistance were measured by asking primary respondents whether they had heard of antibiotics or antibiotic resistance. [15] "No" and "don't know" responses were grouped into one negative response (those who were not confident they had heard of antibiotics) and compared with "yes" responses (those who were confident they had heard of antibiotics) to create a binary variable. The respondents who had heard of antibiotics were asked whether they could spontaneously name any antibiotics, then a list of common antibiotics (including penicillin, doxycycline, tetracycline, erythromycin, streptomycin, Augmentin (amoxicillin + clavulanate), cephalexin, cotrimoxazole, ciprofloxacin, ampicillin/amoxicillin, and colistin) was read out and respondents were asked to say whether they had ever heard of them. We used binary variables for awareness of antibiotics and antibiotic resistance and also generated an antibiotic knowledge score using responses to the spontaneous naming and recognition of antibiotics and an antibiotic resistance knowledge score based on answers to questions about consequences of and ways of preventing antibiotic resistance. What could the consequences of getting an antibiotic resistant infection be? (Multiple choice answers: Be sick for longer; May have to visit doctor more or be treated in hospital; May need more expensive medicine that may cause side-effects; Other) Can you think of any ways of reducing the problem of antibiotic resistance? (Don't take antibiotics when they are not needed (e.g., for colds and sore throats); Don't demand antibiotics from health-workers or drug suppliers; Make sure antibiotics are good quality and within expiry date; Complete the course as recommended by a health-worker; Don't use antibiotics prescribed for someone else; Make sure you use the right antibiotic for the right infection; Make sure you take antibiotics as soon as you feel sick; Make sure you take a very strong antibiotic to kill the infection; Take several different antibiotics to make sure the infection is killed; Don't use antibiotics in animal feed as a growth promoter; Washing hands after contact with a live animal, slaughtering animals, or preparing meat; Washing hands after contact with someone or something that has been touched by a person who has an antibiotic-resistant infection) Score summing each correct answer Education Whether one had attended any school beyond nursery level Any school level-primary to tertiary Never attended school-less than primary school level-or unknown

Data Analysis
Data analysis utilized STATA software version 17.0. Variable categories of similar characteristics were re-grouped into fewer categories to reduce data sparsity. Descriptive statistics were performed to determine the distribution of individual respondent characteristics. Point prevalences were derived for dependent variables, awareness of antibiotics and awareness of antibiotic resistance. Bivariable and multivariable logistic regression models were performed to show associations between media use, health information sources, distance to a health facility, usual health facility, and having a medical insurance card and awareness of antibiotics and antibiotic resistance. Variables such as age, sex, education, occupation, and household wealth were evaluated to be potential confounding factors and were adjusted for in the multivariable logistic models. Bivariable and multivariable linear regression models were run with the continuous antibiotic knowledge and antibiotic resistance scores.

Conclusions
Self-reported awareness of antibiotics was high in this study, though knowledge about types of antibiotics and awareness of antibiotic resistance was much lower. The determinants of antibiotic and antibiotic resistance awareness and knowledge included socio-demographic factors (especially age and occupation) and access to healthcare (usual facility type, distance from health facility). Health workers and interpersonal health information sources were associated with more awareness of antibiotics. High use of media (particularly TV, radio, community radio/loudspeaker, newspaper, and social media) was associated with more awareness of AMR. The complex relationships between knowledge, media access, health information, and behaviour, makes identifying appropriate intervention strategies challenging. Targeting multiple media channels and health information sources, particularly those with the highest access and use, such as television and health providers, is likely to have the highest reach for communication campaigns. Multi-modal campaigns may also be beneficial.  Informed Consent Statement: Written informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The datasets generated and analysed during the current study are not publicly available as this was not included in the consent process, but anonymised datasets can be made available from the corresponding author on reasonable request.