Antimicrobial Use and Management of Childhood Diarrhea at Community Drug Retail Outlets in Eastern Ethiopia: A Matched Questionnaire-Based and Simulated Patient-Case Study

Background

One of the major breakthroughs of 20th century was the discovery of antibiotics which are the magic bullets for several infections.^1–3 These agents have saved millions of lives worldwide and transformed the health of the world one step forward.⁴ However, emerging and gradually evolving nature of antimicrobial resistance (AMR) has threatened the effectiveness of these agents.^1–3 The AMR is recognized as a serious public health challenge and a global bottleneck in this era.^1,5 As a result, the World Health Organization (WHO) recommended member states to prepare national action plans for containment of AMR and preservation of antimicrobials.⁶ On the contrary, there is a dearth of antimicrobial agents in the development pipeline and is alarming the world about the post-antibiotic era which may be realized in the near future unless appropriate measures are taken in advance.^7–9

Major reasons for the increase of AMR are multi-factorial and include unregulated drug availability, absence of antimicrobial stewardship programs, inadequate surveillance, and widespread attitude to antimicrobial misuse, and self-medication.^2,10–12 Patient-related factors are also major drivers of misuse for antimicrobials.^13,14 The dispensing practices of community pharmacy professionals should play a central role for ensuring the safe use of antimicrobials in the community.^15–17

The use of antimicrobials is the a major driving factor for AMR.¹⁸ Oral antimicrobials are inappropriately dispensed without prescription for diseases such as malaria, diarrhea, common cold, pneumonia, and mild childhood infections.^10,19 It is an irrational practice which contributes to the emergence and spread of AMR. Globally, it has been estimated that more than 50% of the antimicrobials are sold over the counter despite the fact that such dispensing practice is illegal in several countries.^20–24 Healthcare professionals have a vital role in preserving antimicrobial agents and containing AMR. In this regard, the spread of AMR will be augmented if they lack up-to-date information, cannot identify the causative agent, or dispense antimicrobials without prescription.²⁵ WHO also estimated that two-thirds of all antimicrobials are sold without prescription through unregulated private drug retail outlets.²⁶

Therefore, assessing dispensing practice of professionals in community drug retail outlets (CDROs) of Eastern Ethiopia was found vital to devise appropriate interventional strategies. In addition, it is also meant to generate pertinent baseline data on the relationship between knowledge and practice as well as the agreement between two measurements for the same case. Analyzing the knowledge and actual practice of dispensing antimicrobial agents in CDROs was found mandatory to take evidence-based measures. Therefore, this study was aimed to assess the dispensing practice of professionals in CDROs supported with their management of childhood acute watery diarrhea in Eastern Ethiopia.

Methods

Study Settings and Design

The study was conducted in CDROs of selected towns in Eastern Ethiopia. The CDROs refer to premises registered and licensed by an appropriate government agency for the provision of pharmaceutical services to the public. The study was conducted from 1 August to 30 September 2020. A community based cross-sectional study was conducted with a mixed data collection tools (structured questionnaire matched with simulated patient method). In this case, acute watery diarrhea was presented to CDROs in both tools. A simulated patient was trained to visit CDROs to enact a scenario that tests a specific behavior of the Pharmacy professional. The simulated patient went to CDROs and presented his symptoms to the pharmacy professional. He told to the professionals that he has suffered from watery diarrhea since yesterday morning and defecated six times a day without associated mucus and symptom of fever.

Study Participants

Pharmacy professionals (pharmacists, pharmacy technicians or druggists) who were working in the CDROs of selected towns of eastern Ethiopia during the study period were our study population. For questionnaire-based study, pharmacy professionals who contacted the simulated-patient were intentionally included for the study. This was made to assess the knowledge of professional about antimicrobial use and AMR as well as determine the degree of agreement between the two measurements for the specified therapeutic area. Community pharmacy professionals who did not serve or contact the simulated-patient (the first study) in selected CDROs were excluded from questionnaire-based study.

Sample Size Determination and Sampling Procedure

A total of 100 CDROs were randomly selected from CDROs of selected towns in eastern Ethiopia. In which, 90 CDROs were randomly selected from Dire Dawa city administration, Harar and Jigjiga towns (30 CDROs from each) and the remaining 10 CDROs were from the towns of East Hararghe. Then, 100 participants (one professional in each CDROs) were selected for one-time mixed method (matched questionnaire-based and simulated patient methods) study, totaled to 200 responses, for ease of assessing the practitioner response and their actual performance in naturalistic settings.

Data Quality Control

A combination of these methods (simulated patient and questionnaire) was pretested on 5% of the sample size and modifications were made accordingly to enhance the feasibility and consistency. In addition, the quality of the study was maintained by explaining the purpose of the questionnaire to the respondents at the beginning of interview. To avoid recall bias, the simulated-patient himself filled the checklist based on the professional recommendations regarding the case management upon exit from each CDRO.

Data Collection Tools and Methods

A structured and pretested questionnaire, prepared in English language, was administered through face-to-face interview to community pharmacy professionals who were working in CDROs. To minimize bias, a simulated-patient visit was initially implemented in each CDRO to collect data about the same case. The simulated patient case employed for this study was adapted from standard treatment guideline of childhood acute watery diarrhea for general hospitals in Ethiopia²⁷ and Infectious Diseases Society of America (IDSA) clinical practice guideline.²⁸

In this regard, the simulated patient method has become popular in health services research to observe the behavior of healthcare practitioners in a natural setting. This method has primarily been used to assess the actual performance of practitioners in the primary care setting. However, the method does not directly capture knowledge. As we also aimed to examine the knowledge of practitioners in a specific therapeutic area, the simulated patient method alone may not be suitable as it captures the actual behavior of practitioners rather than their knowledge (of course, knowledge may inform behavior). A combination of methods (mixed questionnaire-based and case-based tools) was found to be appropriate.²⁹

The data collection tool for the simulated-patient was translated to Amharic, Afan Oromo and Somali languages to consider simulated-patient’s primary language preference in each setting. A pharmacy professional working in each CDRO and who served the simulated patient was interviewed. To match the two methods on the same professional, the data collector with questionnaire immediately entered into the CDRO following the exit of the simulated patient. Recommendations of a professional to a simulated patient was discarded when he/she declined to participate in the interview. Details of patient-related information, all medicines dispensed or recommended, queries about possible drug interactions and allergies, and non-pharmacological advices given (including referral) were recorded by the simulated patient himself after he left each CDRO.

Study Variables

Knowledge of the pharmacy professional about antimicrobial use and AMR as well as dispensing practice of the pharmacy professionals in the management of childhood diarrhea were assessed as key outcome variables. In this study, age, sex, marital status, level of education, years of experience, income level of professional, ownership of premise, clinical conditions and drug related information were considered as potential explanatory variables.

Definition of Operational Terms

Data validity (accuracy) of the questionnaire was assessed against the standard simulated patient method on the same clinical condition as follows:

• Sensitivity: proportion of participants with the condition (appropriate practice) and who are correctly identified by the measurement = true positive/(true positive + false negative).
• Specificity: proportion of participants without the condition (no or inappropriate practice) and who are correctly identified by the measurement = true negative/(false positive + true negative).

Cohen’s kappa: the degree of agreement of the results of two data collection methods on the same case can be measured by Cohen’s kappa (κ), which is given by:

Perfect agreement is evident when Cohen’s kappa equals 1; a value of Cohen’s kappa equal to zero suggests that the agreement is no better than that which would be obtained by chance alone. It considers agreement between two measurements on the same samples (repeatability).³⁰

Appropriateness: Appropriateness of the antimicrobial dispensing practice for the case was determined based on the national and international guidelines:

• In most people with acute watery diarrhea and without recent international travel, empiric antimicrobial therapy is not recommended. An exception may be made in people who are immunocompromised or young infants who are ill appearing. Empiric treatment should be avoided in people with persistent watery diarrhea lasting 14 days or more. Asymptomatic contacts of people with acute or persistent watery diarrhea should not be offered empiric or preventive therapy, but should be advised to follow appropriate infection prevention and control measures.^27,28

Data Processing and Analysis

The data from questionnaire and the simulated patient methods were coded and entered into Epi-data version 3.0 and exported to SPSS, version 21.0 for analysis. Descriptive statistics was employed to summarize sociodemographic characteristics and knowledge of professionals about antimicrobial use and AMR related issues. Sensitivity and specificity of questionnaire was assessed against simulated patient method. Cohen’s Kappa was also analyzed to measure the agreement between the questionnaire-based and simulated patient-based methods. Note: Cohen’s Kappa value indicates level of agreement. So its values < 0 as no agreement; 0–0.20 as none to slight agreement; 0.21–0.39 as fair agreement; 0.41–0.60 as moderate agreement; 0.61–0.80 as substantial agreement; above 0.80 as excellent agreement.³¹

Binary logistic regression analysis was conducted to determine the presence of association between inappropriate antimicrobial dispensing practice for management of diarrhea and explanatory variables. The significance of association was measured using odds ratio with 95% confidence interval and P-value < 0.05. So, explanatory variables which showed potential for association in bivariable logistic regression analyses at P ≤ 0.25 were considered to enter into multivariable logistic regression analysis to adjust for confounders. The results were presented using tables.

Results

Sociodemographic Characteristics of Study Participants

Out of 100 CDROs included in the study, majority of the practicing professionals were aged 25–34 years with the median age of 29 years. Again, the average monthly income of the professionals was 6652 Ethiopian birr (ETB). Majority of the professionals were male (65%), were employed by the owner (61%), were pharmacy technicians (56%) and had work experiences of 2 or more years (Table 1).

Table 1 Sociodemographic Characteristics of the Pharmacy Professionals Working in Selected Community Drug Retail Outlets (CDROs), Eastern Ethiopia, 2020

Knowledge on Antimicrobial Use and AMR

Accordingly, nearly all participants (92%) dispensed antimicrobial agents in response to prescription, 90% of them did it a day before the day of data collection. Ninety nine percent of them advised clients on how to use and handle the antimicrobial agents, more than three-fifths (63%) of them were advising clients not to prematurely discontinue the prescribed antimicrobial agents. Majority of the professionals agreed on how microorganisms develop resistance to antimicrobial agents (81%), increasing number of infections that became resistant to treatment by antimicrobials (96%), challenges of treating resistant bacteria (92%), and impacts of AMR on all community members at large (83%). However, nearly one-third of the professionals had wrong belief and advised clients to use antimicrobials given to family or friends (34%) or those agents that benefited them in the past (39%) and about a quarter of them (27%) thought AMR as not being a problem in the setting. Taking the mean score of correct responses (mean = 5.85) as cut off value for AMR questions, 61% of the professionals who scored above the mean were considered knowledgeable about AMR (Table 2).

Table 2 Knowledge of the Pharmacy Professionals About Antimicrobial Dispensing and AMR in Selected CDROs, Eastern Ethiopia, 2020

Interviewer and Retailer Interaction About Management of Diarrhea

Majority of the retailers responded that they should inquire age of client (94%), onset and duration of diarrhea (94%), presence of vomiting (92%), presence of fever (91%), frequency of diarrhea (90%), presence of blood or mucus (90%), and appearance of diarrhea (86%). Most of them also agreed on the requirement of asking clients about presence of dehydration (73%), medication history (66%) and nutrition condition (61%). Overall, about three-fifths of the retailers (61%) responded sufficiently about the requirements of demographics and patient history in relation to diarrhea symptoms (Table 3).

Table 3 Interviewer and Retailer Interactions Towards the Requirements in the Management of Childhood Diarrhea (Questionnaire Based-Study), Among CDROs, Eastern Ethiopia, 2020

Simulated-Patient and Retailer Interaction

Majority of the retailers requested the simulated-patient’s age (85%), appearance of diarrhea (80%), presence of blood or mucus (76%), and onset and duration of diarrhea (62%). About half of them probed to identify frequency of diarrhea (54%) and presence of fever (51%). However, low proportion of the retailers assessed vomiting symptoms (40%), medication history (11%), dehydration symptom (7%) and nutrition condition which would have been equally important for effective management of diarrhea.

Despite the absence of scientific evidence supporting infectious etiology, about two-thirds of the retailers dispensed antimicrobial agents (67%). Majority of the drug regimen components were addressed in less than 50% settings (Table 4).

Table 4 Simulated-Patient and Retailer Interaction About the Management of Childhood Diarrhea, Eastern Ethiopia, 2020

Agreement Between Questionnaire-Based and Simulated-Patient Case Studies

Generally, the average sensitivity and specificity of the questionnaire-based study against the standard (simulated-patient study) for management of childhood acute watery diarrhea was found to be 64.90% and 36.24%, respectively. The kappa values indicated that the agreement between responses with two methods was from none to slight level in almost all essential components of the management of diarrhea. There was also no agreement for many variables whereas a moderate agreement was observed with regard to recommendation of professionals for other specific agents. Generally, from a total of 2886 scores, only 745 scores were reliable (agreed between raters). The overall percent agreement between the two data collection methods to the same case was 26%, indicating weak agreement between the two measurements and low specificity of questionnaire-based study. About 67% of practices in response to the childhood diarrhea symptom presented to pharmacy professionals was found inappropriate (Table 5).

Table 5 Agreement Between Questionnaire-Based and Simulated-Patient Case Studies for Dispensing Practice of Pharmacy Professionals in Response to Childhood Diarrhea Symptom, Eastern Ethiopia, 2020

Bivariable logistic regression analysis revealed that only retailers’ qualification, favoring sharing of antibiotics, thinking AMR as a risk to users only, knowledge and satisfactory interaction with clients in responding to symptoms were associated with inappropriate dispensing of antimicrobial agents. However, after adjusting for covariates, only retailers with insufficient knowledge on AMR were significantly associated with the inappropriate dispensing of antimicrobial agents for the simulated patient (AOR, 2.9; 95% CI: 1.23–6.89) (Table 6).

Table 6 Factors Associated with Inappropriate Dispensing of Antimicrobial Agents in Response to Childhood Acute Watery Diarrhea Symptom

Discussion

About 61% of community drug retailers had sufficient knowledge on antimicrobial use and AMR. Yet, a considerable number of them had insufficient knowledge about AMR. Indeed, knowledge and formulary management are key aspects to effective antimicrobial stewardships.^32,33 To have sufficient knowledge about antimicrobial use and emergence and spread of resistance, continuous professional development of retailers is mandatory.³⁴ However, majority of community retailers reported limited training opportunities for their continuous professional development.³² In line with this gap, dispensing antibiotics without prescription is becoming an increasing practice.³⁵ In a study conducted in China, the community pharmacy professionals’ knowledge about antimicrobials was good, but their practices regarding antimicrobial stewardship were poor. Hence, tailored training shall be provided to the community pharmacy staff on antimicrobial use, and the need and implementation of antimicrobial stewardship programs.³⁶

Nevertheless, this indiscriminate use of antibiotics has contributed to the development of resistance to these essential medicines.³⁷ Some of the key justifications related to irrational antibiotic dispensing practice could be linked to self-medication, lack of regulations and supervision, and readily available antibiotics.³⁷ Regulations on dispensing antibiotics without prescription should be enforced. Policy makers need to implement antimicrobial stewardship and encourage the use of public educational interventions in community pharmacy.³⁸ Several community retailers agree that undertaking patient examinations and questioning the rationale for antibiotic prescriptions would allow them to contribute to antimicrobial stewardship activities.³⁹ Moreover, key other gap is related with a difference in perception and response of retailers on how AMR occurs.⁴⁰

Practice response of retailers in 40% case scenarios of uncomplicated acute childhood diarrhea presented by a trained simulated-patient was accompanied by recommendation of oral rehydration solution (ORS). On top of this, about two-thirds (67%) of the retailers recommended antibiotic medicines to patient cases without any evidence of bacterial ethology nor did clinical and laboratory evaluation. In knowledge response of retailers to the hypothetical case presented, however, ORS and antimicrobial medicines were recommended for 67% and 69%, respectively. For the case, there was no symptomatic hint to suspect infection. Some illogical reasons linked with such inappropriate practice could be patient pressure, inclination to business, and lack of tight supervision.³³ Always, retailers felt pressured to respond to clients’ demands.⁴¹ Besides, antibiotic treatment was not significantly associated with better treatment outcomes for such diarrheal cases.⁴² Despite a similar proportion of retailers who acknowledge AMR and appropriate antibiotic use in this study, very small proportion of them aligned appropriate matching of their knowledge and dispensing practice. In line with this, previous study indicated that history taking by CDROs for diarrheal case is poor with key patient assessment questions for decision-making missed. Referral decision and the use of ORS is a rare practice. Majority of the dispensaries issued antimicrobials illegally.⁴³

The number of scores matched for Cohen’s interrater agreement was 745 out of 2886 scores. That means, the percent agreement between questionnaire-based and simulated-patient based study was 26%. Particularly, the low specificity value had an overwhelming effect on the kappa-values of the two measurements. This indicates professionals who responded “yes” in the questionnaire-based study (false positives) did not actually practice the appropriate management of childhood diarrhea in simulated-patient study. To obtain substantial or excellent agreement between methods on the same case, there must be a maximum combined sensitivity and specificity. This is also a ringing bell for those using questionnaire-based studies for assessing the actual performance of practitioners in specific therapeutic area.

Observing global research reports, approximately 47.2% community pharmacists dispensed antibiotics to simulated clients without a prescription in the country of Belize.⁴⁴ In Pakistan, of the 386 visits for acute diarrhea, nonprescription antibiotic dispensing occurred in 67.1% visits. Inappropriate dispensing practices were prevalent largely at the drugstores, and antibiotics were effortlessly obtainable without prescription.⁴⁵ Consistent to this, illegal supply of antibiotics without a prescription was reported in about 50% of simulated-patients with acute diarrhea presented to retailers in Sri Lanka.⁴⁶ Again, 15 out of 50 (30%) for simulated-patient visits to community retailers in Gondar town, Ethiopia, led to metronidazole dispensing.⁴⁷ In the same town, 23.9% of retailers dispensed mebendazole to simulated-patients with diarrheal illnesses.⁴⁸ In India, 40.24% antibiotics were dispensed for management of simulated-patients with childhood diarrhea.⁴⁹ More importantly, antibiotics were dispensed to simulated-patients in 62.4% and 52.2% of the cases presented with acute diarrhea in Nigeria and Thailand, respectively.^50,51 Likewise, antibiotic over-prescription occurred in 52.5% of Kenyan children with diarrhea.⁵² Moreover, the proportion of antibiotic dispensing without prescription most commonly for diarrhea and upper respiratory infections account for above 60.0%.⁵³ Along with this varying extent of practice, antibiotic dispensing without prescription is alarmingly an inappropriate practice.⁵⁴ Indeed, irrational antibiotic use in early childhood is considered as a major distractor of gut microbiota, which is linked with reduced floral species diversity, altered metabolic activity, and the selection of resistance.^54,55 Besides, this early childhood exposure to antibiotics can also lead to several gastrointestinal, immunologic, and neurocognitive problems.⁵⁵ Key behaviors justified for antibiotic dispensing without prescription are lack of clinical training and awareness gaps of the community.⁴⁸ The antibiotic dispensing behavior can also be influenced by pressure from patients if customer satisfaction is the most important factor motivating their work.⁵⁶

Retailers with insufficient knowledge on issues around antimicrobials and AMR were 2.9 times more likely to dispense antimicrobials inappropriately to simulated-patient with case of diarrhea. Consistent to this, knowledge of dispensers about antibiotics and AMR was linked with reduced likelihood of illegal dispensing for acute diarrhea (AOR: 0.76, 95% CI: 0.58–0.99; P=0.048).⁵⁷ Non-prescription dispensing of antibiotic agents for diarrhea was one of the most frequently presented case scenario in CDROs⁵⁸ Knowledge of what AMR is and how it occurs are quite rational to influence the practice of retailers.⁵⁹ A study conducted in Jordan emphasized that educational level is a significant predictor of good knowledge about antibiotic use and AMR.⁶⁰ In line with this, continuing education of dispensers, monitoring, and supervision are suggested aspects beneficial to combat resistance.⁶¹ Again, periodic refresher training is crucial in helping dispensers deal with clients who demand antibiotics.⁶²

This study tried to address the knowledge of retailers about antimicrobial use and AMR as well as assess the actual practice of managing diarrhea with a combination of simulated patient and questionnaire-based tools. Indeed, the finding highlights a profound discrepancy in actual practice and knowledge responses towards the management of acute diarrhea as observed from low agreement score. Again, a higher proportion of CDROs was taken from major towns that may not represent practice of smaller towns in eastern Ethiopia. Therefore, interpretation of findings should be made in the context of such limitation.

Conclusion

The study revealed that a considerable proportion of retailers had insufficient knowledge about antimicrobial use and AMR. Nearly a quarter of their response agreed with their actual practice towards the management of acute childhood diarrhea. The study also revealed that there was low sensitivity and specificity of questionnaire against the standard simulated-patient, reinforcing the notion that the low kappa-value is largely influenced by poor specificity of the questionnaire in assessing the actual performance. Insufficient knowledge of retailers was significantly associated with their inappropriate dispensing of antimicrobial medicines. To fill this gap, designing and implementing antimicrobial stewardship programs, applying tight regulation of antimicrobial use in CDROs, conducting targeted and overarching refresher trainings, and holding public awareness campaigns regarding rational antimicrobial use and AMR are highly recommended.