Awareness about antibiotic resistance in a self-medication user group from Eastern Romania: a pilot study

Ethics statement

The study was approved by University of Medicine and Pharmacy Gr. T. Popa Iasi Ethics Committee. Written informed consent was considered not to be necessary, since the study evaluated only the knowledge, attitudes, perceptions and practices of the participants.

Self-medication user group (SMUG), study site and sampling

This cross-sectional self-filled questionnaire based study was conducted in the winter months, from December 2016 through January 2017. Self-medication user group (SMUG) from our study are defined as individuals who addressed to the pharmacy in order to buy any medication without having a prescription. The study was designed to sample randomly at least 215 subjects over 18 years old from the SMUG group. Four private pharmacies from Galati, one of the largest cities in the East of Romania, similar in terms of size and patient load per month, located in different areas of the city were selected (from a list of all centers, we called randomly and selected the first four which accepted to enter the study). We asked every other third person that entered the pharmacy if it came with the purpose of buying medication without having a prescription; if the answer was positive we randomly choose her by coin tossing (heads = request her; tails = not request her) to participate in the study. Subjects with medical training were excluded. The subjects who accepted to participate were informed about the purpose of the study and assured of the confidentiality of the results. After obtaining their verbal consent for the participation in the study, they received the self-administered questionnaire. The questionnaires did not contain personally identifiable information, but only a code number as identifier. Data were collected by a team of PhD students, trained to answer to the participant’s queries about the questionnaire, if any. A reference group, represented by 100 medical residents in their specialty training (MRG), was randomly selected to answer to the same questionnaire being considered to have a higher level of knowledge and awareness about ABU and ABR by comparison to the general population.

Sample size

We considered that the prevalence of non-prescription antibiotic use in Romania is 17%, based on the data from 2016 (European Commission, 2016). Raosoft^® software was used to compute the minimum required sample size. If the non-prescription antibiotic use prevalence is 17%, to obtain a margin of error of 0.05%, with a 95% confidence interval, the sample size needed would be minimum 215 participants. To adjust for possible missing in filling of the data from questionnaire, we selected 250 participants for the SMUG group. For the MRG group, based on the calculation that the prevalence of non-responders in a class survey is about less than 5%, to obtain a margin of error of 0.05%, with a 95% confidence interval, the minimum sample size needed would be 73 participants, so we selected 100.

Development of the questionnaire

A self-administered questionnaire written in Romanian was compiled after a detailed review of relevant literature (McCullough et al., 2015; Shehadeh et al., 2012), taking into account that both the general population and medical residents in training were supposed to answer to the same questions. The questionnaire had 37 questions with closed or open answers that evaluated the practices, knowledge, perceptions and attitudes of the population under study. Scores were created by summing the scores for respective items such that a higher score indicated more positive perceptions and attitudes about ABR. For the Perceptions about ABR, M = 6.89, SD = 2.55, and Cronbach’s Alpha (an estimate of internal consistency and scale reliability) was 0.70. For Attitudes about the ABR, M = 7.01, SD = 2.51, and Cronbach’s Alpha was 0.69. The questionnaire was tested on a group of 20 final year medical students and on 20 subjects selected from the patients coming to a pharmacy to identify any ambiguity in the questions and validated it on two other similar groups. Data collected during this validation pilot study were excluded from the final analysis.

The questionnaire had six sections:

• Participant’s characteristics were collected by seven questions: age, gender, place of residence (rural/urban), level of education (undergraduate/graduate), occupation (employed/unemployed), together with the self-evaluation of responder’s health status and the attitude in the case of a health problem.

• Practice about ABU was evaluated by seven yes/no type questions. The responders were asked to indicate prior ABU at any time in life and separately, in the month previous to the date of the study. They were also asked to specify if they intend to follow the indications of the physician regarding the recommended antibiotic treatment, about their intention to demand an antibiotic treatment in the case of a common cold or dental infection, if they are interested in obtaining information about the recommended antibiotic.

• Basic knowledge about antibiotic use was evaluated by two yes/no/“I don’t know” type questions, four true/false type questions and two multiple choice type questions with only one correct answer. These questions evaluated the knowledge about antibiotic safe use and about the use of antibiotics only in bacterial and not in viral infections.

• Perception about the phenomenon of ABR was evaluated by five yes/no/“I don’t know” type questions about ABR as a problem at national and hospital level or for the future of the medical practice.

• Basic knowledge about ABR was evaluated by one yes/no/“I don’t know” type question about ABR as a consequence of missing one dose of antibiotic, and four true/false type questions about the definition of ABR and about the factors contributing to ABR.

• Attitudes about the measures used by the physician to prevent ABR were evaluated by five questions based on the 5-point Likert scale (strongly agree/agree/neutral/disagree/strongly disagree) about the agreement with some of the measures used by the physician to prevent ABR.

Data analysis

Data were coded and analyzed using SPSS^® v.11 software for Windows (SPSS Inc., Chicago, IL, USA). The Likert’s items were combined and reduced into two categories: “strongly agree, agree” =Agree and “neutral, disagree, strongly disagree” = Disagree; for the other type of answers “yes” = Yes and “no, I don’t know”= No. Descriptive statistics were used to summarize the numerical variables (mean and standard deviation) and categorical variables (expressed as frequencies and percentages). Categorical variables were assessed by chi-square test or Fisher’s exact test when appropriate. Continuous variables were assessed for statistical significance using the Kruskal–Wallis One Way Analysis of Variance on Ranks. A multivariate logistic regression analysis with antibiotic self-medication as the dependent variable was performed to evaluate the relationship with the attitudes, perceptions and the knowledge level of the responders. Those variables achieving in the bivariate analysis a p value ≤0.1 were included in the multivariate logistic regression models, using stepwise selection of variables, to examine for the effects of each independent variable on the different outcomes of interest. The criterion for entering and exiting the variables in the model was, respectively, being of p value ≤0.1 and p value ≤0.05. Two multivariate logistic regression models have been constructed: self-medication with antibiotics at any time in life (Model 1) and self-medication with antibiotics in the last month (Model 2). The variables represented by the answers to the following questions: Q1–3 and Q10–37 were included in all models. The results of the multivariate logistic regression models are presented as adjusted odds ratios (ORs) along with their 95% confidence intervals (CIs) and the percent concordant. A probability level of p ≤ 0.05 was considered as denoting statistical significance.

Participants’ characteristics

By the end of the data collection phase, 218 fully completed questionnaires out of 250 of the SMUG group were collected, and 32 incomplete questionnaires were removed and not included in the data analysis. All 100 questionnaires from the MRG group were fully filled. The response rate was 87.2% in the SMUG group and 100% in the MRG group. The mean age in the SMUG group was 40 ± 12 years and in the MRG group was 29 ± 3 years. In the SMUG group, 82% were employed and 86% were urban residents. The SMUG group was composed of 69% females, while in the MRG group 75% were females. In the SMUG group, 72% called the family doctor, 14% called a specialized doctor as the first person in the case of a health problem, while 8% called the pharmacist and 6% self-medicated. In the MRG group, 51% called the family doctor, 26% called a specialized doctor as the first person in the case of a health problem, while 23% self-medicated. No further analysis of non-respondents was conducted.

Antibiotic use (ABU): practices and basic knowledge

Practices and basic knowledge about ABU are shown in the Table 1.

Concerning the practices about ABU, the responders from the two study groups reported similar self-medication practices for ABU. The subjects reported self medication practices with a high frequency at any time in life, but with very low frequency from the month previous to the date of the study. Although almost all responders agreed to follow the physician’s recommendations regarding antibiotic treatment (95% of the SMUG group and 99% of the MRG group), only 45% of the SMUG group and 44% of the MRG group were willing to return to check with the physician after the antibiotic treatment. Many individuals from the SMUG group had inappropriate practices about ABU, based on the intention to ask for an antibiotic treatment in the case of a common cold (only 80% answered correctly “No”) or in the case of a dental infection (only 83% answered correctly “Yes”), respectively to ask for information about the prescribed antibiotic (80%). More than 94% of the MRG group had very good ABU practices evaluated by the answers to these questions.

Basic knowledge about the use of antibiotics was of medium level in the SMUG group. Confusion regarding some questions about the appropriate use of antibiotics (Q15, Q16, Q17, Q18, and Q22) is high among the SMUG group. Inaccurate knowledge about the effectiveness of antibiotics on bacteria or on viruses was higher among the SMUG group (only 60% responded correctly), although most of the respondents (80% in the SMUG group) knew the usual duration of antibiotic treatment for pneumonia. The usual duration of the treatment for pneumonia and the necessity of an antibiotic treatment were not clearly known but by a few of the SMUG group and this appropriate knowledge existed to nearly all of the MRG.

ABR: perceptions, basic knowledge and attitudes

Awareness about ABR is lower among the subjects from the SMUG group, as it is presented in the Table 2.

A big confusion in the perception of ABR was noticed among the responders from both study groups. Two questions, one regarding the contribution of any antibiotic to ABR (Q23) and another one regarding the preference for taking less antibiotics than those prescribed (Q25), have quite similar frequencies of responders in both groups and the chi-square test showed statistically significant differences between the study groups. ABR was not perceived as a future problem for medical practice by 67% of the SMUG group and by 84% of the MRG group. On the contrary, almost all in the MRG group were aware that ABR is a national problem, but they perceived ABR more as a hospital linked problem rather than a national problem.

Basic knowledge about ABR was weaker in the SMUG group compared to the MRG group. The statement that a missed dose of antibiotic during the treatment could contribute to ABR it was known only by 38% of the SMUG group and by 84% of the MRG group. From the SMUG group 72% defined correctly ABR, although 87% and 93% knew that ABR could result because of the overuse or inappropriate use of antibiotics outside the doctor’s indications, respectively.

Inappropriate attitudes about the measures to prevent ABR were noticed in a small number of responders from the SMUG group. The great majority of the responders from SMUG group agreed with the measures to prevent ABR. The decision for prescribing or not of antibiotic in a febrile patient unless severity criteria are fulfilled was not agreed by 7% of the MRG group.

Basic knowledge scores

The scores were computed as follows: 0–8 for knowledge about ABU and 0–5 for knowledge about ABR. The accurate knowledge about antibiotics was evaluated based on the maximum cumulated scores of the questions regarding ABU and ABR and on the correct answers to the two key questions: one about the duration of antibiotic treatment for bacterial pneumonia (Q21) and one about the indications for antibiotic prescription in other pathological situations (Q22).

According to our cumulative scoring, the mean accurate score for the SMUG group was 2.09 ± 1.1, statistically significantly lower compared to the MRG group (3.58 ± 0.6; p < 0.05).

Using Q22 as control question for Q13, we notice that the matching for the correct answer to both questions (a subject should answer to both questions correctly, if having correct knowledge and correct practice) was only 58% in the SMUG group by comparison to 81% in the MRG group. This could mean that it is possible that 26% in the SMUG group by comparison to 8% in the MRG group have mistaken inflammation for infection.

Multivariate logistic regression analysis of potentially predictive factors associated with self-medication practices

The multivariate logistic regression analysis showed that self-medication with antibiotics at any time in life or in the month before the study could be potentially predicted by the independent variables represented by the coded answers to the following questions: (Table 3).

In the Analysis of Maximum Likelihood Estimates, the positive or negative influence the independent variables selected in the models have on the dependent variables (self-medication) are indicated by the sign of the estimated coefficients. The modeled probability is self-medication = 1, the event. For the coding of the variables, see the Supplemental Information.

Model 1: In the SMUG group the answers to two questions regarding the practices and knowledge about ABU (Q13: “Demand an antibiotic treatment for a dental infection.” and Q20: “Antibiotics are efficient for the treatment of both bacterial and viral infections.”) could be negative potential predictors (negative sign for estimated coefficient) for self-medication at any time in life, with the correct answer: Q13 = yes and Q20 = no. Because of less knowledge in the SMUG group (possibly inflammation mistaken for infection at Q13 and no effect of antibiotics in viral infections and/or viral infections mistaken for bacterial infections at Q20) by comparison to the MRG group, the frequency of correct answers is less (Table 1).

Model 2: In the MRG group, a question referring to the perception about ABR (Q23: “Do you believe that the antibiotics you will receive might contribute to ABR?”) could be a negative potential predictor for self-medication at any time in life with antibiotics, with the correct answer: Q23 = yes.

Self-medication with antibiotics in the month previous to the date of the study in SMUG group could potentially be predicted with three questions: one about ABU practice (Q14: “Interested to ask information about the prescribed antibiotic.”) as a negative potential predictor, with the correct answer: Q14 = yes; one about ABR perception (Q26: “ABR is a national problem.”) as a positive potential predictor, with the correct answer: Q26 = yes, and one referring to ABR knowledge (Q28 “Missing an antibiotic dose during the treatment contributes to ABR.”) as a negative potential predictor, with the correct answer: Q28 = yes. The positive potential predictor at Q26, which came up in the model for the SMUG group, was thought to be due to the insufficient informational campaign at the national level concerning self-medication and antibiotic resistance, which was reflected in the low frequency of correct answers in the SMUG group by comparison to the MRG group (Table 2).

In the MRG group, a question referring to the knowledge about ABR (Q32: “ABR can result after inappropriate use of antibiotics outside the doctor’s indications.”) could be a potential positive predictor for self-medication with antibiotics in the month previous to the date of the study, with the correct answer: Q32 = yes.