FormalPara Key Summary Points

What is already known?

Inappropriate use of antibiotics in humans happens mostly in community settings where qualified medical practitioners are absent.

Lack of adherence to the prescriber’s instruction is one of the major causes of developing antimicrobial resistance.

What are the new findings?

Multifaceted supply-side and demand-side pressure increases the unnecessary use of antibiotics in rural Bangladesh.

Lack of knowledge of antibiotic use by unqualified providers triggers antimicrobial resistance (AMR).

What do the new findings imply?

Irrational antibiotic consumption is likely to decrease by placing qualified healthcare providers in the rural areas of Bangladesh.

Awareness campaigns through coordinated efforts from government and pharmaceutical companies can increase knowledge of antibiotics in pharmacy shopkeepers, unqualified village medical practitioners, and patients, thereby minimizing inappropriate use of antibiotics.

The implementation of antibiotic guidelines and monitoring and some penalties for violating the guidelines may prevent unqualified healthcare providers from unnecessarily prescribing antibiotics.

Introduction

Antimicrobial resistance (AMR) is on the top list of global threats to public health declared by the World Health Organization (WHO) [1]. Although the development of AMR is a natural phenomenon, anthropogenic factors such as misuse and overuse of antimicrobials are significant contributors to this crisis. Knowledge of rational medicine use is insufficient in patients and the general population, irrespective of their educational background and geographical locations [2,3,4]. More alarmingly, irrational antibiotic use by physicians is a significant threat to public health. Guan et al. analyzed 58,512 antibiotic prescriptions from China’s county hospitals, where more than 60% of prescriptions disagreed with standard guidelines [5]. A 2019 study from rural China revealed unnecessary antibiotic use in more than 77% of patient visits [6]. Rahbi et al. studied the knowledge, attitude, and practice (KAP) among physicians in a hospital in Oman and revealed the knowledge gap as a significant contributor to AMR spread [7].

Research on the prevalence and causes of improper antibiotic use can play pivotal roles in reducing the spread of AMR. Scientific articles on this issue are published mainly in the USA, followed by China, India, and the UK, whereas the data from low- and middle-income countries are scarce [8]. In a 2022 survey on rural Bangladeshi people, respondents were found to have moderate knowledge but an aversive attitude toward proper antibiotic use [9]. In a cross-sectional survey conducted in Bangladesh, most respondents (83.2%) were unaware that antibiotic use is inappropriate for most common colds [10]. A recent review has demonstrated that the availability of antibiotics without prescriptions significantly contributes to AMR in Bangladesh [11].

Approximately 62% of Bangladeshi people live in rural areas [12]. A 2007 report says that there are more unqualified village medical practitioners (8.5%) than qualified (MBBS or above) physicians (3.7%) among healthcare providers in Bangladesh [13]. In rural areas, people need more access to registered medical practitioners. Therefore, patients primarily seek medical consultation from unqualified village medical practitioners and pharmacy shopkeepers. The journey of a typical rural patient for medical services is illustrated in Fig. 1. Hence, we aimed to study the impact of knowledge, attitude, and practice (KAP) on antibiotic use and the development of AMR by unqualified village medical practitioners and pharmacy shopkeepers.

Fig. 1
figure 1

The journey of a typical rural patient for medical services

Full size image

Methods

Study Design

A population-based cross-sectional survey was conducted to assess the KAP of participants to further correlate with AMR. The study period was June–December 2021. The study was performed in rural community settings in the Sylhet and Jashore districts of Bangladesh.

Definitions

“Pharmacy Shopkeepers” are “Grade-C Pharmacists” or “Untrained Sales Persons” working in a medicine shop, either listed or unlisted by the Directorate General of Drug Administration (DGDA) of Bangladesh (https://dgda.portal.gov.bd/). According to the Pharmacy Council of Bangladesh (https://pcbapps.com.bd), “Grade-C Pharmacists” are “Short-trained” professionals who can sell medicines. However, in practice, “Pharmacy Shopkeepers” occasionally provide medical consultations. We classified “Grade-C Pharmacists” as “Trained Pharmacy Shopkeepers.” “Unqualified Village Medical practitioners” are “Trained” or “Untrained” medical practitioners who are not legally eligible to prescribe antibiotics. Medical practitioners were classified as “Trained” when they claimed to complete one or more short-term training courses such as Licentiate of Medical Assistants and Family Welfare (LMAF) offered by the Ministry of Health and Family Welfare, Bangladesh and Diploma in Medical Faculty (DMF), Diploma in Medical Assistant (DMA), and Diploma in Medical Science (DMS) offered by the Bangladesh Medical Studies and Research Institute (BMSRI), Bangladesh. In this study, “Providers” indicate both “Pharmacy Shopkeepers” and “Unqualified Village Medical practitioners.” Of note, all village medical practitioners were unqualified providers and not permitted to use the title “Doctor” or “Dr.” according to the existing Bangladeshi law [14].

Sampling

The sample size was calculated using Raosoft (http://www.raosoft.com/samplesize.html). The minimum recommended sample size was 385, considering a 5% margin of error, 95% confidence interval, and 50% response distribution [15]. We randomly approached 500 participants, of whom 396 responded: 149 were “Pharmacy Shopkeepers” and 247 were “Unqualified Village Medical practitioners.” All participants were male in this study.

Sample Recruitment

The eligibility criteria were individuals working in either a drug store or a private clinic, providing healthcare in the form of prescribing, recommending, or dispensing medications, aged ≥ 18 years, and willing to participate in the study. Exclusion criteria were not willing to participate and registered physicians.

Patient and Public Involvement

Patients or the public were not involved in the design, conduct, reporting, or dissemination plans of our research.

Study Questionnaire

A questionnaire was developed on the basis of previous studies [16,17,18] with minor modifications. Initially, we conducted a pilot study and finalized the questionnaire comprising 43 questions (Supplementary Material). Answers were dichotomous (yes/no), Likert-scale (4–5-point) entries, or Likert-type (3–5-point) entries. Each answer received a maximum score of 5 (in agreement with the respective KAP measure). There were seven questions on antibiotic use and AMR, with a score range of 7–35, in each of the following domains: knowledge, attitude, and practice (Supplementary Material, Table S1). We then calculated scores as percentages, where 35 points would make 100%. For each domain, the score was categorized using Bloom’s original cutoff scale [19]. For knowledge and practice, ≥ 80% was considered good, ≥ 60% moderate, and < 60% poor. Similarly, the attitude was categorized as positive (≥ 80%), neutral (≥ 60%), and unfavorable (< 60%).

Primary and Secondary Outcome Measures

Primary outcome variables were knowledge, attitude, and practice of antibiotic use and AMR. Secondary outcomes were the measuring determinants of knowledge, attitude, and practice.

Data Collection and Analysis

Two researchers collected data through structured personal interviews with respondents in rural areas of the Sylhet and Jashore districts of Bangladesh. An interview typically took 30–50 min. Data were recorded on-site using the printed questionnaire and transferred to a spreadsheet electronically. We calculated Pearson’s correlation (r) coefficient to assess the linear correlation between two continuous variables.

Multiple linear regression allows researchers to identify the most important predictors of a health outcome and assess the magnitude and direction of their impact. This approach has been widely used in healthcare research to examine the relationship between various factors and health outcomes [20, 21]. The multiple linear regression analysis was conducted using the ordinary least squares (OLS) method. The OLS method estimates the coefficients of the independent variables that best explain the variability in the dependent variable. The model object was created by specifying the dependent variable and independent variables, and the fit method was applied to estimate the coefficients of the independent variables. The regression results included the R2 value, which measures the goodness of fit of the model, the coefficients of the independent variables, and their associated standard errors, t values, and p values.

A significant difference between the means of the two independent samples was evaluated using the two-tailed t test. A significant difference was assumed when the p value was ≤ 0.05. Statistical analysis and visualization were performed using the open-source applications Statsmodels [22], SciPy [23], Pingouin [24], Pandas [25], and Matplotlib [26] in Jupyter Notebook [27].

Ethics Statement

The study was approved by the Institutional Review Board (IRB)/Ethics Review Committee (ERC) of North South University, Dhaka, Bangladesh. Written informed consent was obtained from each respondent.

Results

Demography

To investigate KAPs, initially, we approached 500 participants and, finally, selected 396 who completed the questionnaire. In Bangladesh, most unqualified village medical practitioners and pharmacy shopkeepers are men as a result of socioeconomic influences. As expected, we did not find any female respondents in our survey. The demographic attributes of the respondents are summarized in Table 1. The age range of the participants was 18–70 years, with a mean age of 37 (± 12) years and a median of 35 years. Among the 396 respondents, 247 (62%) were identified as “Unqualified Village Medical practitioners” and the remaining 149 (38%) as “Pharmacy Shopkeepers.” Two-thirds (66%, 162/247) of the unqualified village medical practitioners reported having at least one medical training. One-fourth (25%, 34/149) of the pharmacy shopkeepers obtained pharmacy training (Table 1). Most participants we surveyed completed at least secondary education. In contrast, only 10% (38/396) had a bachelor’s degree, and 5% (19/396) had a master’s degree (Table 1). Participants’ length of service in their current occupation varied greatly, from less than 1 year to 45 years. Notably, the mean service length was 13.09 (± 9.77) years with a median of 10 years (Table 1).

Table 1 Demographic information of the participants
Full size table

Diseases

We asked the participants about the top three illnesses reported by the patients they serve. Responses are summarized as a word cloud in Fig. S1 (Supplementary Material). In our survey, patients most frequently sought medical services for fever. Patients also visited unqualified providers to treat two related conditions: cold and cough. Gastrointestinal disorders were prevalent, including dysentery, diarrhea, vomiting, and gastritis. Diabetes, hypertension, and headache were prominent among the non-communicable diseases.

Knowledge of Rational Antibiotic Use

When we asked the participants (Q16), “Do you think you have enough sources of information about antibiotics when you need them?”, 62% (246/396) responded “Yes,” while the remaining 38% (150/396) answered “No.” Interestingly, occupation made a stark contrast: most unqualified village medical practitioners (82%, 202/247) claimed to have the required information on antibiotics, whereas only 30% (44/149) of pharmacy shopkeepers responded positively. This may be attributed to the knowledge of antibiotic use the unqualified village medical practitioners gained from their short-term training.

As expected, most participants agreed that antibiotics help treat bacterial infections (Fig. 2A, Q19). On the contrary, most participants have the wrong idea that antibiotics are helpful for viral infections. Only 19.48% of pharmacy shopkeepers and 29.96% of unqualified village medical practitioners disagreed (“Disagree” or “Strongly disagree”) with this statement (Fig. 2B, Q20). Even more alarming, a higher proportion of unqualified providers believe that antibiotics are indicated in pain and inflammation (Fig. 2C, Q21). Nevertheless, most respondents correctly identified that antibiotic resistance refers to the bacterial loss of antibiotic sensitivity (Fig. 2D, Q22). The number of participants corresponding to the proportions depicted in Fig. 2A–D is presented in Table S2 (Supplementary Material).

Fig. 2
figure 2

Knowledge of participants on antibiotic use and resistance. A–D Response summary for four questions from the knowledge domain of the KAP matrix. A Q19, “Antibiotics are useful for bacterial infections (i.e., typhoid)”; B Q20, “Antibiotics are useful for viral infections (i.e., flu)”; C Q21, “Antibiotics are indicated to reduce the symptoms of pain and inflammation”; D Q22, “Antibiotic resistance is a phenomenon for which a bacterium loses its sensitivity to an antibiotic.” E–H Knowledge score of the participants. E Correlation of knowledge scores and age. F Correlation of knowledge scores and the service length. G Knowledge scores by occupation. H Knowledge scores by training

Full size image

Knowledge scores are summarized in Fig. 2E–H. Point estimates and p values for t tests for comparison of knowledge scores by occupation and training status are summarized in Table S3 (Supplementary Material). The scores ranged from 34.29% to 88.57%, with a mean of 59.59%. Overall, the participants had poor knowledge of antibiotic use and AMR. Our data revealed that unqualified providers accumulated knowledge over time through experience. As expected, knowledge scores were positively and significantly correlated with the participant’s age (Fig. 2E) and service length (Fig. 2F). Although participants had poor knowledge scores (< 60%), unqualified village medical practitioners were more knowledgeable about antibiotic use than pharmacy shopkeepers in our survey (Fig. 2G). The mean values were 62.59% (CI 61.52–63.66%) and 54.73% (CI 53.25–56.21%) for unqualified village medical practitioners and pharmacy shopkeepers, respectively. The boxplot in Fig. 2G shows even distributions of knowledge scores around the median values, indicating low variations in knowledge for half of the participants. Irrespective of the occupation, short medical or pharmacy training was a predictor of having slightly higher knowledge (Fig. 2G). The mean knowledge scores were 61.74% (CI 60.64–63%) and 57.42% (CI 56–58.84%) for trained and untrained providers, respectively.

Attitude About Rational Antibiotic Use

Most providers in this survey demonstrated a positive attitude that prescribing antibiotics is harmful to patients if not indicated (Supplementary Material, Fig. S2, Q28). We found the two professional groups (unqualified village medical practitioners and pharmacy shopkeepers) possessing similar attitudes. Two-thirds of the unqualified village medical practitioners recognized, 26% (64/247) strongly agreed, and 49% (122/247) agreed on their roles in fighting against antibiotic resistance, and the remaining unqualified village medical practitioners were either neutral or against this idea (Supplementary Material, Fig. S2, Q30). For this question, we obtained a similar response pattern from the pharmacy shopkeepers (Supplementary Material, Fig. S2, Q30). Among the unqualified village medical practitioners, 25% (62/247) strongly agreed and 55% (137/247) agreed that antibiotic resistance is a global problem (Supplementary Material, Fig. S2, Q31). The agreement was slightly less among the pharmacy shopkeepers: 17% (26/149) strongly agreed and 51% (76/149) agreed. A substantial proportion of the unqualified village medical practitioners (43%, 107/247) and the pharmacy shopkeepers (58%, 86/149) did not consider antibiotic resistance a problem in their regular practices (Fig. S2, Q32).

Total attitude scores ranged from 28.57% to 100.00% (mean 78.46%, std 12.28%), and participants’ attitudes became more positive with age (Fig. 3A) and the service length (Fig. 3B). Attitude scores varied more among unqualified village medical practitioners than pharmacy shopkeepers. Nevertheless, unqualified village medical practitioners scored significantly higher (median 82.86%) than pharmacy shopkeepers (median 74.29%) (Fig. 3C). Interestingly, professional training generated a positive shift in the attitude of providers towards rational antibiotic use (Fig. 3D). Median scores were 74.29% and 85.71% among untrained and trained participants, respectively. Point estimates and p values for t tests for comparison of attitude and practice scores by occupation and training status are summarized in Table S3 (Supplementary Material).

Fig. 3
figure 3

Attitude and practice scores of the participants. A–D Summary of attitude scores. A Correlation of attitude scores and age. B Correlation of attitude scores and the service length. C Attitude scores by occupation. D Attitude scores by training. E–H Summary of practice scores

Full size image

Practice of Rational Antibiotic Use

In response to the question “How frequently do you prescribe or provide an antibiotic for a cold or sore throat?” more than 50% of the respondents among unqualified village medical practitioners and pharmacy shopkeepers answered “none” (Supplementary Material, Fig. S3A, Q35). However, a significant portion of the practitioners (unqualified village medical practitioners, 44%, 108/247; pharmacy shopkeepers, 28%, 41/149) admitted providing an antibiotic to some of their patients for treating a cold or sore throat, whereas 11% (17/149) pharmacy shopkeepers and 5% (12/247) unqualified village medical practitioners give an antibiotic to most patients.

As expected in reasonable practices, most unqualified providers disapproved of completing the full antibiotic dosage regimen of the patients (Supplementary Material, Fig. S3B, Q36). We obtained mixed responses on patient counseling by the providers regarding adherence to the proper use of antibiotics (Supplementary Material, Fig. S3C, Q37). Notably, 8% (20/247) of unqualified village medical practitioners and 19% (28/149) of pharmacy shopkeepers never brief the patients about appropriate antibiotic usage (Supplementary Material, Fig. S3C, Q37). Most respondents denied providing antibiotics to patients when antibiotics are unnecessary (Supplementary Material, Fig. S3D, Q43). However, a significant number of providers (unqualified village medical practitioners, 14%, 35/247; pharmacy shopkeepers, 12%, 18/149) admitted (“Strongly Agree” or “Agree”) to provide antibiotics to patients even when not indicated (Supplementary Material, Fig. S3D, Q43).

For all participants, the mean score for the practice of rational antibiotic use was moderate (mean 70.39%, std 10.14%). Total scores varied between 40% and 97.14%, with 25th, 50th, and 75th percentile values of 65%, 71.43%, and 77.14%, respectively. Neither the practitioners’ age nor service length correlated with the total practice score (Fig. 3E, F). The minimum and median scores were identical for unqualified village medical practitioners and pharmacy shopkeepers (Fig. 3G). However, the 25th percentile value was slightly higher for unqualified village medical practitioners (65.71%) than for pharmacy shopkeepers (60%). On the other hand, individuals with professional training scored significantly higher in practicing proper antibiotic use (Fig. 3H).

Combined KAP Scores

The combined KAP scores of the participants ranged from 40.95% to 87.62% (mean 69.49%, std 7.64%). KAP scores exhibited weak but statistically significant positive correlations with age and the service length of the participants (Fig. 4A, B). The KAP score was slightly higher for unqualified village medical practitioners (median 72.38%) than for pharmacy shopkeepers (median 66.67%), and the difference was statistically significant (Fig. 4C). In addition, trained providers had significantly higher combined KAP scores than untrained providers. The median scores were 73.33% and 66.67%, and the 75th percentile scores were 77.14% and 70.48% for trained and untrained service providers, respectively (Fig. 4D). Participants with a master’s degree scored higher (mean 75.29%, std 4.46%) than participants with a bachelor’s degree (mean 69.32%, std 7.36%) or higher secondary certificate (mean 68.93%, std 7.92%), secondary school certificate (mean 69.67%, std 7.80%), or below secondary education (mean 67.27%, std 4.62%). Point estimates and p values for t tests for comparison of KAP scores by occupation and training status are summarized in Table S3 (Supplementary Material).

Fig. 4
figure 4

Combined KAP scores. A Correlation of practice scores and age. B Correlation of practice scores and the service length. C Practice scores by occupation. D Practice scores by training

Full size image

Figure 5 (and Table S4) shows correlations among the participants’ KAP, knowledge, attitude, and practice. Intriguingly, practice showed no correlation with knowledge but a moderate correlation with attitude. On the other hand, knowledge and attitude had only a negligible positive correlation (Fig. 5, Table S4). The combined KAP exhibited high positive correlations with attitude and practice and a moderate correlation with knowledge (Fig. 5, Table S4).

Fig. 5
figure 5

Correlation among knowledge, attitude, and practice scores

Full size image

For knowledge scores, a negatively skewed distribution and dissonance with practice scores (Fig. 5) reflect that unqualified providers had low knowledge, quantitatively and qualitatively, of rational antibiotic use. On the other hand, attitude scores demonstrate a multimodal distribution (Fig. 5), revealing at least several distinct patterns of attitudes of the participants. Nonetheless, most participants follow a very identical approach to practicing rational antibiotic use, as portrayed by the narrow unimodal histogram of practice scores (Fig. 5).

Multiple Regression Analysis

The multiple linear regression analysis was conducted using the OLS method, which estimates the coefficients of the independent variables that best explain the variability in the dependent variable. The OLS regression analysis is a powerful tool for identifying the significant predictors of a dependent variable and understanding the relationship between the independent and dependent variables. The summary of the model is presented in Table 2. The independent variables in the model, namely age, service length, number of patients served per day, bachelor’s degree, pharmacy training, and medical training, accounted for 23% of the variation in the dependent variable, KAP score. This is reflected in the R2 value of 0.23. The adjusted R2 value of 0.22 suggests that the model does not overfit the data. The F-statistic of 19.44 and its associated p value of 7.43E−20 indicate that the model is statistically significant.

Table 2 Multiple linear regression analysis
Full size table

Among the independent variables, the number of patients served per day, bachelor’s degree, pharmacy training, and medical training have statistically significant coefficients at the 0.05 level. The coefficient of bachelor’s degree, pharmacy training, and medical training is positive, suggesting that having a bachelor’s degree, pharmacy training, and medical training were associated with higher KAP scores. However, the age and service length coefficients were not statistically significant, implying that these variables do not significantly impact KAP score. Overall, the model suggests that patients served and professional training are more significant predictors of KAP score than age and service length are.

Patient Factors

In our survey, we asked the participants about the attitudes and practices of patients about inappropriate antibiotic use, potentially accelerating AMR. Approximately 50% of participants (N = 396) opined (strongly agree, 18%; agree, 33%; neutral, 15%; disagree, 29%; strongly disagree, 5%) that patients’ demand is responsible for excess antibiotic use. Many patients visiting a clinic or medical shop seek antibiotics without prescriptions (all patients, 1%; most patients, 34%; some patients, 61%; none, 4%). In the case of pediatric patients, most parents demand antibiotics from the medical providers (strongly agree, 13%; agree, 52%; neutral, 19%; disagree, 16%; strongly disagree, 1%) for a possible quick recovery. On the contrary, only an insignificant proportion of patients (all patients, 0%; most patients, 9%; some patients, 66%; none, 25%) buy the entire antibiotic course.

Interpretation of Qualitative Studies

We conducted structured interviews with the participants to extract their opinions from the quantitative findings. Previous “successful” episodes with similar symptoms prompted them to demand antibiotics at nearby pharmacy shops making an over-the-counter purchase. Pharmacy shopkeepers said self-medication for conditions like diarrhea and colds with antibiotics is common. This study identified inappropriate antibiotic dispensing by pharmacy shopkeepers and self-medication as severe issues in rural areas because villagers are mainly reluctant to visit qualified medical practitioners to seek advice with fees. Moreover, access to medical practitioners is troublesome as community-friendly primary care facilities need to be developed more.

From the prescribers’ perspective, lack of diagnostic facilities, such as the inability to perform a routine blood test, and worries of turning into a more complicated condition, such as pneumonia or bronchitis, were the most frequently reported reasons for antibiotic prescriptions. Pressure and financial incentives from pharmaceutical companies often compel unqualified medical practitioners to prescribe unnecessary antibiotics. Similarly, pressure to maintain an excellent doctor–patient relationship for long-term business was also reported to satisfy patients’ requests for antibiotic prescriptions. Drug administration’s loose monitoring and lack of forceful regulations on antibiotic prescription and sale indirectly support many unethical activities in the market, neglecting AMR.

Discussion

KAPs of healthcare providers play crucial roles in the emergence and spread of AMR, a multifaceted threat to public health. Irrational prescribing and irresponsible dispensing, two seemingly preventable risk factors, negatively impact patients’ proper antibiotic use and aggravate the ongoing danger of antibiotic resistance. Rural Bangladesh primarily has low socioeconomic status, inadequate healthcare facilities, and unqualified medical and pharmaceutical care providers. We investigated the KAPs of rural providers related to rational antibiotic use and AMR using a questionnaire-based cross-sectional survey among unqualified village medical practitioners and pharmacy shopkeepers.

Through interviewing a total of 396 participants, we observed that this service area is entirely dominated by men, which may limit access to care by women, the population group representing almost half (49.5%) of the total Bangladeshi population (https://data.worldbank.org/), a situation demanding attention by national policymakers.

Two-thirds of the unqualified village medical practitioners claimed to have completed medical training (short course, 3–6 months); the remaining 34.41% still needed to have training. On the other hand, all pharmacy shopkeepers in rural areas are regarded as “Pharmacists” by the general population. However, most (75.17%) of the participating shopkeepers had no pharmacy training or diploma. The rest of the shopkeepers only had a short certificate course in pharmacy. Indeed, 49.75% of the participants were untrained, unqualified providers, potentially limiting proper antibiotic use in the community. Against this backdrop, a higher level of formal education (having a bachelor’s or master’s degree) could compensate, to some extent, for the lack of professional training. However, only a tiny percentage of the participants had a university degree.

At the rural community level, diagnostic facilities are limited. Thus, providers use their experience and judgment to deduce a clinical diagnosis. Since patients seek care for common diseases, including fever, cough, cold, and gastrointestinal disorders, often erroneously identified as bacterial infections, an incorrect diagnosis followed by improper antibiotic use is very likely. If relevant antibiotic use guidelines, such as Standard Treatment Guidelines (STG) on Antibiotic Use in Common Infectious Diseases of Bangladesh [28], are perfectly in place, unqualified providers could be guided to avoid irrational antibiotic use. Ironically, none of the participants identified any such national guidelines.

Viral infections such as colds and flu are often self-limiting, and antibiotics are not indicated [29, 30]. The STG on Antibiotic Use in Common Infectious Diseases of Bangladesh [28] also recommends against antibiotic use in acute febrile conditions for most patients. Paradoxically, approximately 70% of the unqualified village medical practitioners and 80% of pharmacy shopkeepers were either neutral or in favor of antibiotic use in viral infections. As anticipated, knowledge of antibiotic use and AMR was slightly above average among the providers, although experience and training improved the scores somewhat.

Knowledge inadequacy is consequently reflected in irrational practices by unqualified providers. The mean practice score was approximately 70%, with no noticeable impacts on experience, occupation, or training. Instead, we observed inappropriate and alarming trends in antibiotic prescription by rural providers. According to existing laws in Bangladesh, providers without a recognized medical degree (MBBS or BDS) are not allowed to prescribe antibiotics [14, 31, 32]. In addition, dispensing antibiotics by a pharmacist without the prescription of a registered physician is strictly prohibited. Considering prevailing laws and treatment guidelines, the ideal answer should be “None” or “Never” for the survey question “How frequently do you prescribe or provide an antibiotic for a cold or sore throat?”

In contrast, approximately 50% of unqualified village medical practitioners and 60% of pharmacy shopkeepers admitted prescribing antibiotics to patients. In our survey, participants’ attitude scores were slightly higher than knowledge or practice scores. In addition, the combined KAP scores showed a high positive correlation with attitude. However, the attitude was only moderately correlated with practice.

We can infer that participants’ poor knowledge and attitude regarding antibiotic use expose patients to unnecessary antibiotics channeled through irrational prescribing and dispensing practices. A good confidence level and a low knowledge level among providers were also identified in a systematic review of studies conducted in several developing countries [33]. In a 2019 study by Nair et al., 88% of registered physicians and 85% of informal practitioners reported prescribing antibiotics for common viral illnesses [16].

Patients’ attitudes and socioeconomic status also contribute to the spread of AMR [34]. Recent studies indicate that knowledge of antibiotic use and resistance could be better among Bangladeshi people. In a 2020 study, 60% of parents of low-income urban parents reported not completing the entire antibiotic course [35], potentially favoring antibiotic misuse in children and the spread of AMR. Makhdum and colleagues revealed moderate knowledge and poor antibiotic use practices among rural Bangladeshi people [9]. In our study, 70% of unqualified village medical practitioners denied the influence of patients’ attitudes on their prescriptions. Thus, blaming only the patients for the spread of AMR might not be justified [36].

Irrational antibiotic use and AMR are multifactorial problems involving many stakeholders at multiple levels. Targeting prescribers, patients, and dispensers in isolation is less likely to bring positive outcomes; instead, integrated action plans are needed to promote antibiotic stewardship. One approach to solving the problem would be improving knowledge among rural prescribers and patients through education and training programs. In a survey on health providers in Iran, most participants suggested that education programs can help reduce AMR [37]. However, another study found that institutional training provided to medical students might not effectively raise sufficient awareness [38]. In a 2021 study [39], experts suggested that improving awareness and practices, primarily among households and pharmacy shopkeepers, are expected to bring the highest impact, secondary to targeting behaviors of physicians and pharmaceutical companies.

People receive healthcare services and purchase antibiotics for common illnesses from unqualified providers in rural communities, where diagnostic facilities and evidence-based prescribing are nonexistent and empirical treatments prevail. Thus, disseminating national STGs on antibiotic use among unqualified village medical practitioners is expected to lead to refrain from prescribing antibiotics when not indicated. As per the laws, antibiotics should only be prescribed by registered physicians. Similarly, antibiotics should be dispensed only with prescriptions, and registered pharmacists should be available to provide antibiotics and counseling to the patients. Digital tools and platforms have become increasingly important in our daily life. These tools include online training modules, mobile apps, electronic decision-making algorithms, and electronic health records. By integrating these tools into healthcare practice, practitioners and patients can access up-to-date information on best practices for antimicrobial use. This can enable informed decision-making regarding the use of antibiotics, ultimately leading to a reduction in AMR. Therefore, creating awareness among all stakeholders and enhancing vigilance and law enforcement by concerned government authorities would be integral to the prudent use of antibiotics in rural Bangladesh.

Limitations

A few limitations could confound the interpretation of results from this study. First, the survey was limited to a small geographic location and, thus, might not adequately represent all rural areas of Bangladesh. Second, participants’ responses were not verified for accuracy. For example, we needed to confirm whether a participant completed the reported training. We could not include qualified medical practitioners in this study as they are usually unavailable in the rural areas we selected.

Conclusion

To the best of our knowledge, this is the first study to assess KAPs among unqualified village medical practitioners and pharmacy shopkeepers in Bangladesh. Participants’ age, service length, and training positively correlated with scores. In all domains of KAP, unqualified village medical practitioners performed slightly better than pharmacy shopkeepers. We found discordance between knowledge and practice but a moderate positive correlation between attitude and practice. All participants’ mean knowledge and practice scores were poor and moderate, respectively, although the mean attitude score was positive. The government should take necessary measures to increase awareness of proper antibiotic use and AMR among patients, prescribers, and dispensers, implement treatment guidelines for common illnesses, and ensure that all parties adhere to the laws and regulations.