Tertiary hospitals physician’s knowledge and perceptions towards antibiotic use and antibiotic resistance in Cameroon

We conducted a cross-sectional survey from May 2018 to February 2019, targeting all medical doctors practicing in major tertiary hospitals of Yaoundé, Cameroon, namely General Hospital (YGH), Gynaeco-Obstetric and Paediatrics Hospital (YGOPH), Central Hospital (YCH), and University Teaching Hospital (YUCH). The tertiary hospitals together employ approximatively 260 medical doctors including specialists, registrars, and general practitioners.

A convenience sampling was used. An email invitation generated from Research Electronic Data Capture (REDCap) tools [16] hosted at University of Cape Town-Department of Medicine, was sent to every doctor working in these hospitals. The purpose of the study was clearly explained in the aforementioned invitation email. The email was automatically resent after 10 days if there was no answer or if the questionnaire was not yet completed. This follow up period was subsequently reduced to 5 days after an initial low response rate. The invitation was resent up to five times.

The self-administered questionnaire comprised 54 questions and ten clinical cases structured in four sections:

The questions have been made simple and unambiguous. The questionnaire was anonymous and was in English and French, the two official languages of Cameroon. The choice of the clinical cases and micro-organisms were based on the prevalence/frequency of the diseases in the selected hospitals irrespective of the wards’ specificities and are similar to the questions used in previous studies on knowledge, attitudes and practices on antibiotics use and antibiotics resistance [17].

Electronic informed consent was obtained from each participant. All procedures were approved by both the Human Research Ethics Committee of the University of Cape Town (HREC REF: 203/2018), South Africa, and the Ethics Committee for Human Health Research in the Centre Region, Cameroon (REF: 0548/CRERSHC/2018). The study was performed in accordance with the Declaration of Helsinki.

Descriptive statistics were used to characterize the total sample. Continuous variables were described using mean with standard deviation or median with interquartile range depending on the distribution of data. Proportions were calculated for categorical variables. Composite scoring of perceptions and attitudes was based on the five-point Likert-type response scales: strongly agree/agree/neutral/disagree/strongly disagree, strongly influence/influence/neutral/do not influence/do not influence at all, often/sometimes/rarely/never/not familiar, and very helpful/helpful/neutral/not helpful/not helpful at all [18]. Likert-type ratings (single survey questions) were summarised as proportions by collapsing the two categories on either side of the neutral category and creating an ordinal variable with three categories. The reported categories were therefore a positive response, a neutral response, or a negative response (e.g., agree/neutral/disagree). For analysis of knowledge, each correct response was scored 1 while a wrong or doubtful response 0. The mean knowledge score was then calculated. The association between the knowledge scores and composite scores of perceptions and attitudes expressed as Likert-type categories or baseline socio-demographic characteristics were compared using either one-way ANOVA test for variance or two-sample t-test respectively. Chi-square or Fisher’s Exact tests were used for assessing the differences between socio-demographic variables and perceptions/attitudes questions expressed as Likert-type categories. A p-value < 0.05 was considered as statistically significant.

Of the 206 doctors invited to participate to the study, 98 (48%) responded and 66% (65/98) completed the questionnaire in its entirety. The median age was 31 years (interquartile range (IQR) 29–34) and 55% were female. The majority of respondents were residents (60%) (Table 1).

The overall mean score on the knowledge questionnaire was 56% (± 14), with over a quarter (26%, 17/65) of participants scoring less than 50%. Participants scored well regarding the use of antibiotics for common infections—community acquired pneumonia (CAP) 71%, urinary tract infection (85%)—but scored poorly on antibiotic de-escalation (14%, 9/65) and duration of antibiotic prophylaxis (25%) (Table 2). Regarding the management of ESBL-E. coli infection, the choice of antibiotic was appropriate for 42% (27/65) of respondents.

When evaluating factors influencing doctor’s knowledge, we found no correlation between the mean knowledge score and a previous training on antibiotic use and resistance. The knowledge score was not influenced by the doctor’s category, specialty, age, experience, or practice location (hospital, ward). Moreover, perceptions on antibiotic use and ABR were not associated with knowledge. Participants who do not mainly rely on textbooks as resources for antibiotics prescribing scored better than those who do (58% ± 13 vs 40% ± 12; p = 0.01) (Table 3). Doctors who were not influenced by patient pressure for antibiotic prescribing had a better knowledge score than those influenced by patients or those who were neutral (67% ± 12 vs 53% ± 13 or 62% ± 13; p = 0.01).

When assessing common resources used by participants for antibiotics prescribing, we found that the most used were, in decreasing order of frequency, textbooks (93%, 68/73), professional/learned societies’ guidelines (93%, 68/73), WHO (World Health Organization) guidelines (81%, 59/73) and smart phones apps (60%, 44/73). Our results revealed that Infectious Diseases (ID) specialists are used as resources for antibiotic prescribing by only 58% (42/73) of doctors surveyed. Most respondents rely on other colleagues (73%, 53/73) and consultants (54%, 39/72) were the least utilised resource when prescribing antibiotics.

During the last 12 months, respondents attended a lecture or workshop or course addressing the following topics: diagnosis of infection (49%, 34/70), dose and duration of antibiotic therapy (43%, 30/70), indications for antibiotic use (34%, 24/70), spectrum of activity of commonly used antibiotics (23%, 16/70), indications for intravenous antibiotic use (21%, 15/70) and principles of Infection Prevention and Control (19%, 13/70). Nearly all participants agreed that medical journals (97%, 69/71) and conferences (93%, 66/71) are useful resources for learning on antibiotic use and resistance. Pharmaceutical representatives were considered useful resources by 44% (31/71) of respondents. It was almost unanimously felt (97%, 73/75) that development of local guidelines would be useful in improving antibiotic use and mitigating ABR. Furthermore, almost all respondents (99%, 74/75) were of the opinion that more education on antibiotic use and resistance would be helpful.

Most respondents, 93% (74/80) agreed that ABR is a significant problem in Cameroon, but only 40% (32/80) believed that it is a problem in their hospital wards. Interestingly, 54% (42/78) of doctors disagreed that poor hand hygiene is a cause for spread of antibiotic resistant bacteria (Fig. 1).

Despite acknowledgement by 96% (77/80) of participants that antibiotics are overused in Cameroon, only a third believed that antibiotics are overused in their ward (Fig. 1). Two-thirds (65%, 49/75) of prescribers were confident that they use antibiotics optimally (Fig. 2). This confidence was not associated with the category of the prescriber, age, experience or previous education on antibiotic use, perceptions of antibiotic use and ABR, or level of knowledge.

Almost all respondents (97%, 73/75) agreed that inappropriate antibiotic use can harm patients (Fig. 2).

The leading factors influencing antibiotic prescribing were guidelines from learned societies and WHO (100%), quality of drug and cost effectiveness ratio (99% and 96% respectively), immunosuppression and/or critical clinical status of the patient (90%) (Table 4). Interestingly, patient expectations did not influence antibiotic prescription for 71% of doctors.