Article Text
Abstract
Introduction Overprescription of antibiotics poses a significant threat to healthcare globally as it contributes to the issue of antibiotic resistance. While antibiotics should be predominately prescribed for bacterial infections, they are often inappropriately given for uncomplicated upper respiratory tract infections (URTIs) and related conditions, such as the common cold. This study will involve a qualitative systematic review of physician-reported barriers to using evidence-based antibiotic prescription guidelines in primary care settings and synthesise the findings using a theoretical basis.
Methods and analysis We will conduct a systematic review of qualitative studies that assess physicians’ reported barriers to following evidence-based antibiotic prescription guidelines in primary care settings for URTIs. We plan to search the following databases with no date or language restrictions: MEDLINE, Web of Science, CINAHL, Embase, the Cochrane Library and PsycInfo. Qualitative studies that explore the barriers and enablers to following antibiotic prescription guidelines for URTIs for primary care physicians will be included. We will analyse our findings using the Theoretical Domains Framework (TDF), which is a theoretically designed resource based on numerous behaviour change theories grouped into 14 domains. Using the TDF approach, we will be able to identify the determinants of our behaviour of interest (ie, following antibiotic prescription guidelines for URTIs) and categorise them into the 14 TDF domains. This will provide the necessary information to develop future evidence-based interventions that will target the identified issues and apply the most effective behaviour change techniques to affect change. This protocol follows the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols guidelines.
Ethics and dissemination Ethical approval is not required. Findings will be published in a peer-reviewed journal and presented at conferences.
- PRIMARY CARE
- Respiratory infections
- QUALITATIVE RESEARCH
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
STRENGTHS AND LIMITATIONS OF THIS STUDY
We have prepared this protocol according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis Protocols guidelines.
The study will follow a rigorous methodology to identify, appraise and synthesise our findings.
We will analyse our findings using frameworks from behaviour sciences with a strong evidence base for their application in designing and evaluating behaviour change interventions.
This review will be limited by excluding grey literature, which could increase the chance of publication bias.
The analysis may be affected by the research team’s experience and personal biases (eg, data analysis will be completed by non-clinicians who do not prescribe antibiotics or have expertise in antibiotic prescription); however, findings will be reviewed in detail by a coauthor (KA-B) who is a practising family physician.
Introduction
Upper respiratory tract infections (URTIs) are one of the most common diagnoses patients receive in primary care. It has been estimated that in 2015, there were over 17 billion instances of URTIs globally.1 2 URTIs are infections that cause irritation and swelling of the upper airways. They often involve the nose, sinuses, pharynx, larynx and large airways.3 Symptoms can include a sore throat, cough, runny nose, nasal congestion, headache, low-grade fever and malaise, lasting up to 3 weeks.3 Treatment for URTIs typically includes recommendations for rest, lots of fluids and over-the-counter cold medicines to help with symptoms. Since URTIs are virtually always caused by viral pathogens, not bacteria, they do not respond to antibiotics.4 5 Guidelines have been very clear about not prescribing antibiotics for URTIs and only prescribing for several related conditions such as pharyngitis, bronchitis, sinusitis and otitis media when there are clear indications (see online supplemental appendix 1 for the guidelines and recommendations for antibiotic prescribing for URTIs and related conditions).4 However, prescribing antibiotics for these conditions has been an issue globally for decades and does not appear to be decreasing. Rates of inappropriate antibiotic prescribing for URTIs have been reported from anywhere between 15.4% and 60% in outpatient settings.6–8
Supplemental material
Overprescribing antibiotics poses a significant threat to healthcare globally as it contributes to the issue of antibiotic resistance.5 9 10 Family physicians are one of the primary health providers who prescribe antibiotics for URTIs, and numerous interventions have been developed to improve the quality and quantity of their antibiotic prescribing (eg, patient and provider education, decision support, point-of-care testing and delayed prescribing). While there has been some success for some of these interventions (eg, point-of-care testing, communication training and delayed prescribing), effects sizes are generally small, and no intervention has been able to move the needle so to speak.11–13
What are the barriers to reducing antibiotics for URTIs?
The most recent reviews in this area were completed by Rezal et al in 2015 and Germeni et al in 2018. Rezal et al included physicians of any specialisation while Germeni et al included any healthcare provider who prescribed antibiotics.14 15 Reported factors largely fell into one of three categories: physician-related factors (eg, previous clinical experience, continuous medical education, misconceptions about evidence-based prescribing, perceived patient expectations, diagnostic uncertainty, confidence regarding following appropriate prescribing behaviours and desire for a quick fix), patient-related factors (eg, patient’s signs and symptoms at the time of the prescription), and healthcare system/resource-related factors (eg, time restrictions, patient load, cost savings and financial incentives).14 These results are consistent with other reviews that have been completed in this area.16 Thus, antibiotic prescribing is a complex process with several factors influencing physicians’ prescribing behaviours. However, this review only included studies from 1990 to 2014, and since then, several new studies from multiple countries have rendered this information outdated.
Furthermore, this review did not complete a theoretically driven analysis of their data. While their synthesis approach summarised the results, quality and limitations of their included studies to provide useful knowledge on the topic, a theory-driven analysis using a behaviour change theoretical framework can provide more useful information for designing interventions to target the behaviour of inappropriate prescribing of antibiotics for URTIs. Interventions systematically developed based on a theory-informed assessment of the barriers to adopting a behaviour have a better chance of including strategies to change that behaviour effectively.17 18
One comprehensive behaviour change theoretical framework is the Theoretical Domains Framework (TDF). The TDF is an established framework developed by a collaboration of behavioural scientists and implementation researchers to provide a theory-based approach to identifying determinants of behaviour and is a synthesis of 36 behaviour change theories grouped into 14 domains (see online supplemental appendix 2 for the TDF domains and their definitions19 20). Examples of the domains include knowledge, skills, beliefs about capabilities and social influences. The TDF was developed in conjunction with the behaviour change techniques (BCTs) taxonomy. The BCT taxonomy is a list of 93 techniques that can be used for changing behaviour, and these techniques have been linked to the 14 TDF domains (see online supplemental appendix 3 for a list of all BCTs19 20). The TDF and BCT taxonomy are intended to be used together to help design theory-informed behaviour change interventions. As such, Michie et al20 provide guidance on selecting the most appropriate BCT (based on the best available evidence, theory and expert consensus) to target an identified TDF barrier or enabler. Using these resources will enable future researchers to design interventions targeting known barriers matched with appropriate BCTs. This approach has been used widely in the literature to understand the barriers and enablers to change behaviour and to underpin the design and evaluation of behaviour change interventions.21–26 Pinder et al27 published a report nearly a decade ago using a similar methodology to understand antibiotic prescribing. However, this report was not a peer-reviewed systematic review and was largely limited to studies relevant to the UK.28
Supplemental material
Supplemental material
Our study aims to systematically review barriers and enablers to physicians’ antibiotic prescribing behaviours following a framework analysis using the TDF. Much of the literature focuses on quantifying antibiotic prescribing, but a deeper understanding of why overprescribing persists is required to reduce antibiotic prescriptions. Completing this qualitative systematic review will help us better understand why physicians continue to not follow antibiotic guidelines for URTIs and inform intervention design to address these barriers and ultimately improve antibiotic prescribing practices in primary care.
Research question
What are family physician-reported barriers and enablers to following evidence-based antibiotic prescription recommendations for URTIs and related conditions in primary care settings?
Objectives
The objectives for the proposed study are as follows:
To conduct a qualitative systematic review of family physicians’ perspectives and experiences regarding barriers and enablers to following evidence-based antibiotic prescription recommendations for URTIs and related conditions in primary care settings.
To analyse the data collected from the systematic review using the TDF.
Methods and analysis
All methods were designed following the JBI Manual for Evidence Synthesis for Systematic Reviews of Qualitative Evidence.29 This protocol was prepared following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols (see online supplemental appendix 4) reporting guidelines, with additional guidance from Enhancing Transparency in Reporting the Synthesis of Qualitative Research (ENTREQ) checklist to ensure we maintain reporting standards in our protocol (see online supplemental appendix 5 for the checklist30). The study’s full report will follow the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA31) and ENTREQ recommendations for reporting and transparency in qualitative research and systematic reviews to ensure we are maintaining reporting standards for the dissemination of our findings.
Supplemental material
Supplemental material
Patient and public involvement
This study is part of a larger Canadian Institutes of Health Research-funded grant entitled ‘De-implementing low value care: A research program of the Choosing Wisely Canada Implementation Research Network’, which involves multi-jurisdictional research across Newfoundland, Ontario and Alberta that focuses on how to support the adoption of Choosing Wisely Canada recommendations. The research team for this grant already has an established patient partner council which was consulted for the identification and prioritisation of this project. In addition, the study results will be reviewed with the patient partner council. They will be invited to participate in the interpretation of the results and asked to help produce all post-publication knowledge translation products, such as a plain-language summary and an infographic.
Eligibility criteria
We have followed the adapted PICO framework to define the question components and eligibility criteria recommended for qualitative reviews, which include the terms Population, Phenomenon of Interest and Context. Our population of interest is family physicians. Our phenomenon of interest is the discussion of the barriers and enablers to following guidelines for antibiotic prescribing for URTIs and related conditions (ie, any barriers to following evidence-based antibiotic guidelines for URTIs or any enabler to not following these guidelines). We have used the most recent Choosing Wisely Canada recommendations for URTI antibiotic prescribing published in ‘The Cold Standard’ to inform our definition of URTI and related conditions (see online supplemental appendix 1 for details).32 Our context of interest is patients of any age seeking care for URTIs (and related conditions) in primary care settings. See table 1 below for a complete list of inclusion and exclusion criteria for these terms and criteria for eligible study designs, publication types and publication languages.
Search strategy
An experienced librarian at the Health Sciences Library of Memorial University of Newfoundland has developed a comprehensive search strategy that adheres to PRESS (Peer Review of Electronic Search Strategies) guidelines.33 Keywords included antibiotic resistance, antibiotics, general practice, acute respiratory tract infections, upper respiratory tract infections, qualitative and family physicians. Databases that will be searched include: MEDLINE, Web of Science, CINAHL, Embase, the Cochrane Library and PsycInfo (see online supplemental appendix 6 for copies of our search strategies). All databases will be searched from database inception to the search date, with no date or language restrictions. To ensure our search is robust, we will conduct reference list screening and citation tracking of all included studies. We have also identified three previous relevant systematic reviews that we will include in our reference screening and citation tracking.14–16 Finally, we will contact key content experts in the area to check if any relevant studies they know of have been missed.
Supplemental material
Selection process
All titles identified by the initial search will be added to Covidence systematic review software (available from covidence.org), and duplicates will be removed. Two reviewers will screen article titles and abstracts of all studies identified following a screening template with predefined eligibility criteria (see online supplemental appendix 7 for screening template). The screening template will be pilot tested on 20 articles prior to completing the screening of all identified articles. Any conflicts that arise will be resolved by a consensus. If a consensus cannot be achieved, a third investigator will be consulted. Two reviewers will also complete the full-text review following the screening template to select the final articles to be included in the study. A third reviewer will be available to mediate disagreements if a consensus cannot be reached between the reviewers. The screening process will be documented using the PRISMA flow diagram (see online supplemental appendix 4).34
Supplemental material
Assessment of methodological quality
To our knowledge, there is currently no risk of bias assessment tool for qualitative studies. Therefore, we will follow the methodology outlined by Hall et al,21 which has combined elements from the Critical Appraisal Skills Program35 methodological section B on methods and the four methodological domains from the Consolidated Criteria for Reporting Qualitative Research (COREQ36) guidelines (recruitment, data collection, researcher–participant relationship and analysis)35 (see online supplemental appendix 8 for the checklist). Two reviewers will apply this tool to each included study and score each question in the checklist as ‘yes’, ‘no’ or ‘can’t tell’. Based on these tools and the process used by Hall et al,21 studies will be given an overall score which will determine if the study will be ranked as having good, moderate or low methodological rigour. Any disagreements will be resolved via a consensus, and a third reviewer will be consulted if necessary.
Supplemental material
Assessment of reporting quality
The COREQ checklist will be used to assess reporting quality.36
Data extraction process
Two researchers will extract all data using data extraction templates (see online supplemental appendix 9). Any discrepancies will be resolved via a consensus. The data extraction templates will be piloted on two studies to ensure they capture all the necessary information. Information to be extracted includes study characteristics, including study year, country, setting, sample size, research aim and data collection methods. Additionally, the results of the included studies will be extracted in terms of the themes of the main findings. We will contact the authors if data are missing or unclear (eg, inclusion criteria such as infection types and specialisation of physicians).
Supplemental material
Strategy for data synthesis
Target behaviour
The target behaviour for this analysis is based on Choosing Wisely Canada’s evidence-based prescribing guidelines for family physicians—specifically, that antibiotics should not be prescribed for URTIs.
TDF synthesis
To synthesise the extracted data for this review, we will use the TDF to create a framework for content analysis, deductively assigning the results of included studies to one or more TDF domains. Since the domains of the TDF are quite comprehensive in terms of categorising barriers and enablers that influence behaviour, we do not anticipate identifying any information that does not fit within these domains. However, if we collect data that cannot be coded to any of the TDF domains, we will code the information to a category called ‘other’, which will be analysed inductively to uncover additional categories that may be used to organise the thematic findings of the studies included in this review. Under the direction of a TDF expert, two researchers will be trained to code extracted data to the TDF domains. We will use data from a previous similar review (on a different topic but using the TDF coding scheme) to practise coding. From this work, the researchers will create a codebook specific to the current study that will act as a guideline and reference to ensure accuracy and consistency. The codebook will contain the coding strategy and a table of coded text, which will define clear methods for making decisions on which domain is appropriate and how to deal with disagreements.
Using NVivo qualitative data analysis software, two researchers will independently code the complete results of the included studies (ie, authors’ descriptions of the results, identified themes and subthemes and illustrative participant quotes provided in the results section (or results tables) of included studies). Once both reviewers have independently coded all data to the TDF domains, a summary of the coding results will be reviewed with the research team for discussion and agreement on coding interpretations. Finally, we will analyse the extracted data to determine the number of contributing studies for each theme and to describe the relevant study information to prepare the data for the confidence assessment.
Assessment of confidence in findings
To assess the extent to which a finding is representative of the phenomenon of interest (ie, how representative the results of a study are of the true barriers and enablers to following URTI antibiotic guidelines), we will follow the GRADE-CERQual (Confidence in the Evidence from Reviews of Qualitative research) approach for assessing the confidence of evidence of reviews of qualitative research. This will allow us to determine how much confidence we can place in the results of the studies we included.37 This approach focuses on four components: the methodological limitations, the relevance of studies to the review question, the coherence of the review findings and the adequacy of data contributing to the review findings. There are four confidence levels: high, moderate, low and very low. Confidence levels are downgraded from high based on the four components. See online supplemental appendix 10 for the detailed criteria we will use for assessing confidence in our findings following the GRADE-CERQual approach based on the criteria from Hall et al.21 Currently, to our knowledge, there is no standardised assessment tool for publication bias in qualitative research; therefore, this study will not be assessing included studies for this meta-bias.
Supplemental material
Ethics and dissemination
Since this review is drawing on already published data, we do not need to undergo formal ethics approval. We plan to publish our findings in an open-access peer-reviewed journal (following ENTREQ guidelines for reporting) and conference presentations. Additionally, we plan to disseminate our findings in non-traditional methods such as infographics or short research videos to improve the visibility of our research findings.
Ethics statements
Patient consent for publication
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors AH and KB conceptualised and designed this review. AH and KB drafted the protocol. KB, AH, KR and AP developed the search strategy. AH, HE, KA-B, AP and AMP provided feedback on the manuscript for content and methodology. KB, along with AP and AH, will perform study selection and data extraction. AH is the guarantor of this review.
Funding This study is supported by the Canadian Institutes of Health Research (grant number 398 527).
Competing interests None declared.
Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.