Antimicrobial stewardship among Nigerian children: A pilot study of the knowledge, attitude, and practice of prescribers at two tertiary healthcare facilities in Bayelsa State.

Antimicrobial stewardship (AMS), the evidence-based use of antimicrobials, is an effective strategy in controlling antimicrobial resistance (AMR) in humans by reducing the irrational use of antimicrobials. Stewardship in children is less studied. This study assessed the knowledge, attitude, and practice of physicians prescribing antibiotics to children in Bayelsa State, Nigeria to identify gaps in AMS and possible solutions. Following ethical approval, a semi structured questionnaire was distributed among 40 paediatricians and gynaecologists at the two public tertiary healthcare facilities in Bayelsa State (the Niger Delta University Teaching Hospital and the Federal Medical Centre) for self-completion. Responses were expressed as percentages and analyzed using Bloom's cutoffs. The Capability, Opportunity, Motivation, and Behaviour (COM-B) model was employed to identify gaps for intervention in prescribing behavior with gaps in each component identified by aggregate scores <80%. Perceived approaches to improve prescribing among 14 selected options were assessed using 5-point Likert scales and options with scores >90% rated the most acceptable. Questionnaires were administered from August to September 2021. The response rate was 68% (27/40). Participants were paediatricians (81%, 22/27) and gynaecologists (19%, 5/27). Antimicrobial Susceptibility Testing (AST) was not performed before antibiotic selection nine times out of 10 (89%, 24/27). In a third (37%, 10/27) of cases, 2-3 antibiotics were prescribed. The top three antibiotics, in rank order, were: cefuroxime or amoxicillin 41% (11/27); ciprofloxacin or amoxicillin 30% (8/27), and azithromycin (33%, 9/27). Aggregate COM-B scores were: capability, 74%; opportunity, 78%; and motivation, 87%. The most acceptable (100%, 27/27) options to improving antibiotic prescribing were: availability of resistance data, availability of guidelines, readily accessible microbiological data, and easy access to infectious disease physicians. There are gaps in knowledge of AMR and opportunity for rational prescribing. There is need for antimicrobial resistance data to promote pediatric AMS at the surveyed healthcare facilities.


Introduction
Antimicrobial resistance (AMR), an inherent and developed ability of microorganisms to 68 withstand treatment with antimicrobial agents (1), is a global public health emergency (2,3). The 69 factors contributing to the emergence and spread of AMR in humans are related to the misuse, 70 underuse, or abuse of antimicrobial agents in humans, animals, agriculture, and the environment 71 (1,4). Antimicrobial stewardship (AMS), the evidence-based use of antimicrobials, is an 72 effective strategy in controlling AMR in humans by reducing the excessive use of antimicrobials 73 (5). Overall, 50% of global antibiotic use is inappropriate, calling for urgent antimicrobial 74 stewardship to reduce indiscriminate and excessive use of antibiotics (6). 75 Antimicrobial use in children tends to be higher than in adults, especially in developing countries 76 (7). In most Low-and Middle-Income Countries (LMICs), the high burden of infectious diseases, 77 unrestricted access to antimicrobials, and poor awareness promote the indiscriminate use of 78 antimicrobials and drive AMR (8-12). Emerging research show that the use of substandard and 79 falsified (SF), or poor-quality, antimicrobials can also contribute to AMR (13,14).    CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted December 2, 2021. ; https://doi.org/10.1101/2021.11.30.21267070 doi: medRxiv preprint Sample size: A sample size of 20 pediatricians was targeted, generally considered appropriate for 105 a pilot study (23,24). To provide information on prescribing practices around pregnancy and 106 childbirth, the survey was extended to prescribers with obstetrics and gynaecology (O&G) 107 specialization, increasing the sample size to 40.

108
Participants: Participants were purposively selected as pediatricians and obstetrics and gynecology 109 specialists.

110
Study instrument: The survey instrument was a semi-structured questionnaire adapted from 111 previously used and validated questionnaire surveys assessing knowledge, attitude, and practice 112 of prescribers and other health workers in Europe (25,26) and Yemen (27). The questionnaire was 113 arranged in 3 "sections" and contained 40 items, not counting an initial consent question and asking 114 any additional question at the end (Supplement Table 1). The questions were adaptive, such that 115 participants did not have to answer all but answered related questions based on their selection of a 116 previous option. Selected questions were repeated to check for individual response reliability.

117
Section I collected demographic data on the prescriber's specialty and years of experience (2 118 questions). Section II assessed knowledge, attitude and practice, or prescribers' behaviour. It 119 contained 37 questions on current antibiotic prescription patterns and decision-making, knowledge 120 of, and attitude or perceptions, about both AMR and existing antimicrobial stewardship structures.

121
Knowledge was both real (assessed by 1 question containing 7 "key" statements with true or false 122 answers) and assumed, or perceived, (assessed by 3 questions with Yes/No answers or 5-point 123 Likert scales). Section III consisted of one question that further evaluated perceptions among 14 124 options for improving antibiotic prescribing, or AMS. The questionnaire was made available in 125 both paper and electronic forms. The online questionnaire was prefaced by an introductory section 126 that presented the title, aims, and objectives of the study in an information section and consent,  Analysis was performed using descriptive statistics expressed in proportions or percentages.

137
Questionnaires completed by hand were converted to digital forms for data analysis. Dummy 138 variables were created for questions not completely filled in the paper forms. These were usually 139 "false", "unsure", or assigned a "neutral" score. This applied to 5 questions in 2 otherwise 140 completely filled paper questionnaires.

141
To characterize providers' prescribing behaviour and identify gaps for interventions, the  Table 2). Capability was an umbrella term encompassing individual knowledge, 144 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted December 2, 2021. ; https://doi.org/10.1101/2021.11.30.21267070 doi: medRxiv preprint awareness, and perceptions. Opportunity was defined both as factors that lie outside the individual, 145 including awareness or perception of existing facility-level AMS structures, as well as attitudes 146 including confidence in prescribing. Motivation was all the brain processes that energize and direct 147 behaviour and included a self-belief of a key role in the containment of AMR and association 148 between prescribing and the emergence of AMR. All responses on a 5-point Likert scale were re-149 coded such that "strongly agree" and "agree" were assessed as "agree", and "strongly disagree", 150 "neither agree nor disagree", and "disagree" re-coded as "disagree". Incomplete responses on 151 paper were assigned the 'neutral' "neither agree nor disagree". Bloom's cutoffs were then applied 152 to aggregate scores for the relevant responses to identify gaps to target for each behaviour 153 component. With responses assessing participants' knowledge and awareness of AMR, as well as 154 opportunity for AMS, the cutoffs and interpretations were set at: >80% as "good", 60-79% as 155 "moderate", and <60% as "poor" (28). For attitude, these cutoffs were, respectively, analyzed as 156 "positive", "neutral", and "negative" (28). Motivation was assessed as "high" >80%) or "low" 157 (≤80%). Gaps were scores <80% in any one domain. 158 To assess acceptable options for improving AMS, responses were re-coded and rated based on 159 percentages. Responses on the 5-point Likert scale were aggregated with very helpful and helpful 160 classified as "helpful" and neutral, unhelpful and very unhelpful re-coded as "unhelpful". The 161 options recategorized as helpful were then rated as "most acceptable" if they had aggregated scores 162 >90%, "acceptable" with scores of >80-90%, and "maybe acceptable" if rated as helpful by 70-163 80% of respondents.

165
Responses are detailed in Supplement Table 3.  Table 3a).  CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

197
Guidelines: Standard Treatment Guidelines, or a guide specific to prescribing, was used by slightly 198 more than half (55%, 15/27) of participants to guide antibiotic prescribing in their facility.

208
AMR capacity: The knowledge and ability to inform others about AMR and use this to guide 209 appropriate prescribing ranged from 70% (moderate) to 89% (good) (Table 1). Overall, AMR 210 capacity was good (81%).

211
Knowledge: There were individual variations in answering the 7 "key" knowledge statements.

212
Only two of these were correctly answered by all 27 (100%). To varying degrees, the answers to  (Table 2). Aggregate real knowledge was good.

219
One Health: Knowledge of One Health contributors to AMR varied widely. Only 37%, 11/27) 220 correctly agreed that environmental factors such as wastewater contributes to AMR. In contrast, 221 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted December 2, 2021. ; https://doi.org/10.1101/2021.11.30.21267070 doi: medRxiv preprint 70%, 19/27) correctly agreed that the excessive use of antibiotics in livestock and food production 222 contributes to AMR (Table 3). The aggregate One Health knowledge score was poor.

223
Awareness of WHO AWARE classification: There was poor awareness of the WHO Access, 224 Watch, and Reserve (AWaRe) classification of antibiotics. Only 15% (4/27) had heard about the 225 classification.

(which was not certified by peer review)
The copyright holder for this preprint this version posted December 2, 2021. ; https://doi.org/10.1101/2021.11.30.21267070 doi: medRxiv preprint emergence and spread of AMR but can also lead to other diseases later in life (29). It is thought 260 that intestinal dysbiosis resulting from early antibiotic exposures predisposes to paediatric 261 idiopathic arthritis, inflammatory bowel disease, asthma, and diabetes(citation).

262
This study attempted to evaluate the knowledge, attitude, and practices on antimicrobial use, 263 resistance, and stewardship, as well as attitude towards interventions to strengthen AMS among hemisphere is low. In Nigeria, one study found that none out of 12 surveyed tertiary hospitals had 272 a formal AMS structure (33).

273
Prescribing is a complex, time-dependent activity in which the prescriber will consider, amongst 274 other things, the symptoms and severity at presentation, the age, allergy history, drug-drug AWaRe classification is to ensure that 60% of prescribing is from the access group of antibiotics.

290
The findings from the prescription audit earlier (30) showed that this is not yet the case in this  . This innovation has also led to several approved devices that can rapidly 299 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

301
Approaches to tackle AMR in tertiary healthcare facilities should start from the community. The  This study clearly shows that there is a need to improve providers' knowledge on AMR and AMS.

315
Providers also need to be better acquainted with the WHO guidelines among pediatricians and 316 other practitioners who prescribe antibiotics for use in the pediatric population. Continuing 317 medical education for the rational use of antibiotics at regular intervals could motivate the 318 necessary behaviour change among pediatric practitioners. This study is not without its limitations.

319
For one, we did not assess the impact of cost on prescribing behavior, nor did it evaluate 320 practitioners' knowledge of the pharmacodynamics and pharmacokinetics of antibiotics prescribed, 321 as this was beyond the scope of the current research effort.  There is a need for readily accessible resistance data and information to guide rational 330 antimicrobial prescribing in tertiary healthcare settings in the Niger Delta region of Nigeria. A 331 clinical decision support system that integrates local resistance data to guide antibiotic prescribing 332 would be an effective pediatric AMS strategy in this and similar settings. To be effective, this 333 system should be readily accessible to prescribers and regularly updated with resistance data.

334
Shared ownership, fostered through a participatory approach, could help enhance uptake and use.

335
Rapid diagnostic tests to rapidly inform antibiotic choice are a must and would facilitate 336 transitioning from empiric therapy to definitive therapy. Future research should focus on 337 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted December 2, 2021. Financial support 342 We confirm no financial support was received for this project.

473
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted December 2, 2021. ; https://doi.org/10.1101/2021.11.30.21267070 doi: medRxiv preprint Figure 1Map of Nigeria with study location indicated by the rectangle.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted December 2, 2021. ; https://doi.org/10.1101/2021.11.30.21267070 doi: medRxiv preprint   Aggregate score 52% . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted December 2, 2021. ; https://doi.org/10.1101/2021.11.30.21267070 doi: medRxiv preprint Aggregate score 78% . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted December 2, 2021. ; https://doi.org/10.1101/2021.11.30.21267070 doi: medRxiv preprint Aggregate score 87% Note: Aggregate score is the overall average score of participants across questions.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted December 2, 2021. ; https://doi.org/10.1101/2021.11.30.21267070 doi: medRxiv preprint