Article Text
PDF
Abstract
Introduction An increasing number of opioids and other controlled substances are being stolen from healthcare facilities, diverting medications from their intended medical use to be used or sold illicitly. Many incidents of medication loss from Canadian hospitals are reported as unexplained losses. Together, this suggests not only that vulnerabilities for diversion exist within current medication-use processes (MUPs), but that hospitals lack robust mechanisms to accurately track and account for discrepancies and loss in inventory. There is a paucity of primary research investigating vulnerabilities in the security and accounting of medications across hospital processes. The purpose of this study is to map hospital MUPs, systematically identify risks for diversion or unintentional loss and proactively assess opportunities for improvements to medication accounting and security.
Methods and analysis We will conduct human factors-informed clinical observations and a Healthcare Failure Mode and Effect Analysis (HFMEA). We will observe hospital personnel in the intensive care unit, emergency department and inpatient pharmacy in two hospitals in Ontario, Canada. Observations will capture how participants complete tasks, as well as gather contextual information about the environment, technologies and processes. A multidisciplinary team will complete an HFMEA to map process flow diagrams for the MUPs in the observed clinical units, identify and prioritise potential methods of medication loss (failure modes) and describe mechanisms or actions to prevent, detect and trace medication loss.
Ethics and dissemination We received province-wide research ethics approval via Clinical Trials Ontario Streamlined Research Review System, and site-specific approvals from each participating hospital. The results from this study will be presented at conferences and meetings, as well as published in peer-reviewed journals. The findings will be shared with hospitals; professional, regulatory and accreditation organisations; patient safety and healthcare quality organisations and equipment and drug manufacturers.
- hospital medication use process
- diversion
- healthcare safety and quality
- human factors
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
Applying human factors methodologies embraces system complexity and allows diversion to be studied from a systems, as opposed to an individual blame, perspective.
Basing the analysis on data collected through observations enables the study to identify vulnerabilities in processes according to how they are actually performed instead of how they are perceived to occur (work as done vs work as imagined).
Conducting the study in multiple units in two hospitals enables corroboration of results between sites, as well as the comparison of workflows and failure modes across hospitals and as a function of clinical area.
Probability and severity scoring of failure modes (and other components of the hazard analysis) is subjective; however, our study design mitigates this with a multidisciplinary team and independent scoring.
There are widespread system-level and individual-level practice variations within a hospital, and point-in-time observations likely do not capture all possibilities, even as attempts to increase the number and time of observations are employed.
Introduction
The opioid crisis claims lives every day, with opioid misuse causing increasing rates of morbidity and mortality across Canada.1–4 A worrisome parallel trend suggests a growing number of opioids and other controlled substances (CS) going missing or being stolen from Canadian healthcare facilities5–11 and entering the illegal street market.7 8 The theft of medications for personal substance use or trafficking is described as ‘diversion’, as drugs are transferred, or diverted, from legitimate medical to non-medical use.12 Weaknesses in the security and accounting of CS in hospitals enable medications to be lost or diverted.13 It is increasingly recognised that Canadian hospitals lack robust processes and infrastructure to accurately track and resolve discrepancies in their CS inventory. For example, of the 1020 incidents of CS losses and thefts detected and reported by Canadian hospitals to Health Canada in 2016,9 >80% were reported as unexplained losses, meaning that at the time of reporting (ie, within 10 days of discovery), the loss could not be attributed to any particular cause or action. What have not been explored are the vulnerabilities within the hospital medication-use process (MUP, eg, procurement, storage, preparation, prescription, dispensing, administration, reconciliation, waste, return and removal) that increase the potential for diversion to occur. With Canadian hospitals experiencing increasingly formal expectations that they will verify and enhance diversion safeguards to protect patients and healthcare workers,14 15 they require systematic knowledge about where vulnerabilities exist and advice and guidance on how to mitigate these risks.
Impact of hospital medication diversion
The hospital setting is vulnerable to diversion by healthcare workers because of the large quantity of stock and proximity with which many hospital personnel interact with medications. Ease of access and frequent interaction with CS can be considered occupational hazards, increasing the risk of diversion and substance use disorder among healthcare workers.16–18 The opportunity to divert medications can escalate drug-seeking behaviour and lead to overdose and death13 19 or infection from unsterile medications and needles.20–22 There are also professional risks, including suspension or termination of employment, revocation of license to practice, civil malpractice claims and criminal prosecution.23–25
Diversion has been shown to have negative effects beyond its impact on the person who is diverting medications, including on patients, healthcare facilities and the larger community. Patients have been directly harmed by receiving inadequate analgesia or anaesthesia when their medication is diverted,26–28 been provided substandard care when their healthcare worker was impaired29 30 and even contracted viral or bacterial infections due to medications or syringes compromised in the diversion process.20 21 31 Hospitals bear the cost of diverted medications from their stock, follow-up patient care and investigations stemming from diversion, and reporting to authorities.26 32 33 The larger community is impacted by the increase in the supply of medications ending up on the street7 8 and decreased public trust in healthcare professions, institutions and workers.
Gap in understanding vulnerabilities for diversion in hospital MUPs
System-wide gaps in the security and traceability of medication transactions through technologies, processes and environments can result in considerable losses of medications without recourse to audit or trace their whereabouts. As a result, many hospitals may not be aware of the deficiencies in their medication accounting and security processes. Further, the large proportion of unexplained losses suggest that current estimates of medication thefts in Canadian hospitals, diversion or otherwise, underestimate the issue. There is a lack of primary research describing how medications are lost or stolen from hospitals. Diversion literature largely consists of expert commentary and institutional experience,34–37 case reports,38–40 commentary on past incidents21 41 and audit reports.42–45 These methods are retrospective and limited in their ability to identify or adequately characterise the system vulnerabilities that enable diversion. Although it is important to investigate the effects of these incidents and update best practices in response, it is equally, if not more important, to proactively identify potential risks to prevent new and unexpected patterns of diversion. To address this gap, we propose a study designed to map two hospitals’ MUPs and systematically identify vulnerabilities in these processes that increase the risk for diversion. To our knowledge, this is the first study to prospectively and systematically investigate the vulnerabilities compromising the security and accounting of medications across the scope of hospital MUPs, as opposed to confined to a specific task or process, and to suggest mitigation strategies.
Objectives
The objectives of this study are to understand the security and accounting of medications throughout the MUPs in two Ontario hospitals, to identify vulnerabilities and existing safeguards and to proactively identify opportunities for improvement.
Recognising the sensitivity of the topic, we emphasise that our study seeks to understand diversion from a systems perspective, empirically and objectively identifying process failures in the security and accounting of medications rather than characterising, blaming or otherwise criminalising healthcare workers who may be diverting.
Methods and analysis
Overview
The study team comprises five health services researchers with backgrounds in medication safety—three (MDV, MF and PT) with expertise in human factors, one with clinical experience as a hospital pharmacist (DT) and one as a practising physician (MH).
Our study comprises two integrated parts, as one (clinical observations) informs the other (risk analysis). Figure 1 describes the study design, showing the order of the steps from each part. We will conduct clinical observations to understand and contrast MUPs across units and hospitals. Although we are interested in identifying vulnerabilities in the MUP that could allow diversion to occur, we do not expect to observe incidents of diversion. Rather, the purpose of the observations is to map the MUPs. We will use Healthcare Failure Mode and Effect Analysis (HFMEA) to proactively identify and evaluate failure modes in MUPs and identify opportunities for improvement to medication accounting and security. The study observations and analysis will take place from May 2018 to October 2019.
Study design. Integration of clinical observations and HFMEA. HFMEA, Healthcare Failure Mode and Effect Analysis; MUP, medication-use process.
Clinical observations
Setting
Clinical observations will be conducted in three units (intensive care unit, emergency department and inpatient pharmacy) in two large (over 400 acute care beds) full-service hospitals in Toronto, Ontario, Canada. We purposively selected the settings to meet three criteria: one academic and one community hospital site, units with high use and access to CS and sites using different automated dispensing cabinet (ADC) platforms. Table 1 describes the units and lists the processes and personnel that we expect to observe at each. Several process tasks are expected to follow similar procedures/protocols given that both hospitals have central inpatient pharmacies that distribute unit-dosed medications to the floors, have ADCs on the clinical units and operate within the same provincial health system. However, some process tasks are expected to differ between hospitals and clinical units because of differences in technologies (eg, use of different ADCs) and protocols (eg, requirement of a witness for wasting). For example, emergency departments often use paper documentation of medication orders and administration, whereas electronic systems are used to record these events in the intensive care units.
Description of clinical observation sites and medication-use processes
Participants
We will use purposive sampling to recruit participants for the clinical observations. We will include front-line healthcare workers who have a role in or interaction with at least one component of the MUP and who consent to being observed. This includes healthcare workers who directly interact with medications (eg, dispensing and administering medications), as well as hospital personnel who are involved indirectly (eg, encountering partial vials of medication while cleaning patient rooms). We estimate that a sample size of 20 participants is the minimum number of observations required to reach theoretical saturation, whereby additional sessions would not likely yield further insights. Therefore, the estimated number of participants is 160 (20 individuals per unit x 2 hospitals x 4 units). However, the number of healthcare workers recruited for observations is expected to differ somewhat between units because of differences in staffing complement, shift schedules and number of tasks related to the MUP. For example, in the intensive care units, we expect to observe a minimum of 14 nurses, 2 pharmacists, 2 physicians, 1 respiratory therapist and 1 environmental services staff, whereas in the inpatient pharmacies, we expect to observe 18 pharmacy technicians and 2 pharmacists (see table 1 for a description of MUPs and personnel who will be observed in each clinical unit at each site).
Participants will be asked by the study team to sign consent forms before being observed. Participants will be given as much time as they require to review the consent form and have their questions answered by the study team prior to deciding if they wish to participate. The study team will highlight that participation is voluntary and can be stopped at any time for any reason and that clinical performance is not being assessed or evaluated.
Data collection
Two members of the study team (one human factors specialist and one clinician) will jointly observe within each hospital unit for approximately five times a week for 4 weeks. Observations will take place on all days of the week and include all hours of the day. Each observation session will last for 2–8 hours, depending on the participants’ availability, the shift duration and the task(s) being observed. Some tasks are frequent and repetitive so require less time to capture, whereas others occur infrequently or over the course of a longer time period so require longer observation periods. Observers will unobtrusively shadow participants as they carry out their daily activities. The purpose of the observations is to obtain a detailed understanding of participants’ typical tasks and responsibilities, as well as the procedures and equipment related to the MUP. The observations will also characterise problematic issues that are observed (eg, not logging out of the ADC system) or that participants describe to the observer (eg, unwillingness of peers to witness wasting). Observations will capture the MUP for all medications, but with a focus on CS to identify safeguards and vulnerabilities specifically for these medications.
Observers will take free-form notes, collect artefacts of clinical practice (eg, blank pre-printed forms), as well as take photographs of the environment, technology and supplies. The photographs will be used to recall or visualise process steps during the mapping process. Images will also be used to provide context when presenting and describing results. The free-form notes will capture step-by-step how participants complete tasks as well as contextual information, including the physical layout of the unit, the roles and shifts covered by staff, technologies used to document dispensing and locations of medications on the unit. The observer will fully transcribe their free-form notes into Word© and upload them onto a secure SharePoint© site hosted at the study team’s home organisation. Emerging findings will be confirmed with healthcare workers in the units.
Coding of observation data
Data collected during observations will be uploaded into MAXQDA© V.2018.1 data management and analysis software. One human factors specialist will code the observation data using codes for hospital units (intensive care unit, emergency department and inpatient pharmacy), tasks and vulnerabilities or safeguards. A second study team member will review the codes, and any discrepancies will be resolved through discussion. Coding of the observational data in MAXQDA© will create a dataset that is structured so that the study team can search and filter data related to specific MUP tasks, roles, technologies or environments. These are important inputs for conducting the HFMEA, providing not only information on how tasks were performed and by whom but also contextual information for conducting the hazard analysis described below.
Healthcare Failure Mode and Effect Analysis
Overview of HFMEA
HFMEA is a prospective risk analysis that involves mapping detailed process flow diagrams and then systematically identifying and prioritising vulnerabilities via a structured decision-making algorithm.46 HFMEA was developed by the Department of Veterans Affairs National Centre for Patient Safety (NCPS) in 2002.47 It has been successfully applied to several healthcare processes, including the ordering and administration of medications as well as the sterilisation and use of surgical instruments.48–52 HFMEA combines concepts and components from the Failure Mode and Effect Analysis (FMEA), Hazard Analysis and Critical Control Point (HACCP) and root cause analysis (RCA).47 FMEA was originally used in aviation, manufacturing and nuclear industries to evaluate risk of products, and has been used in healthcare to conduct proactive risk analyses on high-risk technologies and processes.53 54
The HFMEA approach was developed to address criticisms of using FMEA in healthcare, particularly with respect to the use of a single risk priority number (RPN) to rank vulnerabilities.55 The RPN in FMEA is calculated by multiplying scores from three ordinal scales: severity, probability and detectability. Multiplying these scores creates an RPN that is mathematically flawed, unstable (small changes in one score can lead to large changes in RPN) and masks important distinctions.55–57 For example, a failure mode with high detectability, high probability, but low severity would be prioritised the same as a failure mode with high detectability, low probability, but high severity despite having different risk implications.55 Given that failure modes with the highest RPN would be considered as hazards with the highest priority, efforts may be misdirected based on a misleading RPN score. HFMEA addresses these concerns by prioritising vulnerabilities using a decision tree analysis. The decision tree analysis considers not only severity and probability scores, but also assesses the criticality of the failures (ie, single point weaknesses) and whether there are controls in place to prevent or detect these failures. The use of ‘yes’ and ‘no’ responses in the HFMEA decision tree to assess the criticality, presence of control measures and detectability of the failure modes46 is less subjective and more easily agreed upon than assigning scores.58
The HFMEA process includes five main steps.47 After the study team defines the topic that will be analysed and assembles a multidisciplinary team, information from the clinical observations will be used to map process flow diagrams for the management and use of medications in the clinical units. Next, we will identify potential methods of medication loss and evaluate their severity, risk and detectability, as well as identify potential areas where mitigation strategies can be implemented. Unique to our study is that the HFMEA will be conducted for the same processes at two sites, enabling us to find similarities and differences in processes, failure modes and controls.
Define the topic
The first step is to define the HFMEA topic, including boundaries to limit its scope. Our HFMEA will examine the hospital MUP, including the procuring, storing, ordering, dispensing, preparing, administering and wasting of medications. The study team will limit the topic to specific units within the hospital (ie, intensive care unit, inpatient pharmacy and emergency department). Any hospital personnel role, technology or object that directly or indirectly interacts with medications will be included. Processes that are external to the hospital unit or roles that are not affiliated with the hospital will be out of scope (eg, administration of medications by paramedics and delivery of medications from distribution centres).
Assemble the team
The second step is to assemble a multidisciplinary team. The HFMEA team will comprise three human factors specialists, two pharmacists, one physician, two nurses and two pharmacy technicians. The membership of the team ensures there is expertise in collecting and analysing observational data and proactive risk analysis, as well as knowledge and experience working in the different hospital settings and performing tasks covering the breadth of the MUP. For particular steps of the HFMEA, team members will vary as per the function of the unit being analysed (eg, pharmacists will brainstorm failure modes in the pharmacy). The team will communicate over email as well as during in-person meetings. A minimum of five in-person meetings for each clinical unit will take place to cover the graphical description of the MUPs; identification and description of failure modes; assignment of severity and probability scores; decision tree analysis and identification of critical failure modes, causes and controls; and actions and outcome measures. These meetings are embedded within the remaining steps described below.
Graphically describe the process
The third step is to develop process flow diagrams and number each task and subtask. Creating process flow diagrams is an important first step in identifying safety risks from different aspects of a work system (eg, individual, technology and administration).59 The HFMEA team will use the data collected during the clinical observations to graphically map the step-by-step MUPs from each clinical unit at each hospital site. Using direct observation of processes, as opposed to mapping processes according to how tasks are supposed to occur, will strengthen the validity of our results.60 The maps will be created by retrieving data coded for specific units and tasks and translating the process steps into a visual process flow diagram using draw.io©. The mapping process will be completed collaboratively between observers and iteratively during the clinical observation period, so that gaps or steps requiring clarification can be gathered in the next observation session. If observers note differences in how participants perform the same process, this variation will be discussed by the team and described in the process flow diagrams, because variations may suggest vulnerabilities in process. Figure 2 shows an example of the task and subtask figure that will be constructed from the process flow diagrams produced in this step of the HFMEA. The team will review the detailed process flow diagrams and one human factors specialist will transcribe each task (eg, dispensing from ADC) and subtask (eg, logging into the ADC, selecting the patient and selecting the desired medications) into Excel©.
Example task and subtask figure for the distribution of medications from the inpatient pharmacy to the clinical unit. The first level of the figure is the pharmacy process, the second level is the flow diagram of tasks and the third level is the numbered subtasks that occur within each task. Subtasks are described separately for the two hospital sites. FM1, FM2 and so on indicate the subtasks where critical failure modes were identified. C1, C2 and so on indicate the subtasks that act as controls at one site for critical failure modes identified at the other site. Numbering of critical failure modes and controls correspond to the descriptions in figure 5. ADC, automated dispensing cabinet; CS, controlled substances.
Conduct a hazard analysis
The fourth step consists of four sub-steps: (A) list and number all potential failure modes (ways a step within a subtask can fail to accomplish its intended purpose) and potential effects if a failure were to occur; (B) score the severity and probability of potential failure modes; (C) use a decision tree analysis to identify critical failure modes and (D) list all causes of critical failure modes.
Two HFMEA team members will independently brainstorm failure modes and effects for each of the subtasks, and any discrepancies will be discussed. If a decision on whether or not to include a failure mode cannot be reached, a third member of the team will reconcile the discrepancy. Failure modes will be organised into a worksheet (figure 3) to facilitate the recording of results.
Two HFMEA team members will independently score failure modes based on their severity and probability, as described by the NCPS (table 2).46 A hazard score is calculated by multiplying the severity and probability scores. The intra-class correlation (ICC) will be calculated for a subset of hazard scores to assess inter-rater reliability. Definitions of scale scores will be discussed and refined until an accepted level of agreement is reached (ICC ≥0.60). The severity and probability of the remaining failure modes will then be scored.
The HFMEA team will use a decision tree to prioritise the failure modes (figure 4). Failure modes with sufficient hazard scores or that are single point weaknesses (ie, failure in this step will invariably result in an adverse event) are considered in the next step of the decision tree. If an effective control measure exists (eg, storing medications in a locked drawer to prevent an individual from opening the drawer and removing medications from it) or the failure mode is so obvious and apparent that a control measure is not warranted, then the failure mode does not proceed through the next steps of the HFMEA. All remaining failure modes are labelled as critical and considered in sub-step D. Figures 2 and 5 together provide an example of the anticipated outputs of the hazard analysis. Figure 2 shows which subtasks are associated with critical failure modes at each site using FM1, FM2 and so on as markers. When one site has a control in place to mitigate a critical failure mode identified in the other site, this is flagged with C1, C2 and so on. Figure 5 provides a description of the corresponding critical failure mode and controls.
The HFMEA team will brainstorm the potential causes of the critical failure modes and record these in the worksheet (figure 3). Completing the hazard analysis will produce a list of critical failure modes and their causes.
Example of HFMEA worksheet. CS, controlled substances; ED, emergency department; HFMEA, Healthcare Failure Mode and Effect Analysis; ICU, intensive care unit; Pharm, inpatient pharmacy.
Probability and severity scoring
Decision tree analysis. Used to conduct step 4C of the HFMEA. HFMEA, Healthcare Failure Mode and Effect Analysis. Adapted from VA National Center for Patient Safety.46
Example results table of critical failure modes and controls for the distribution of medications from the inpatient pharmacy to the clinical unit. The table describes the critical failure modes and controls identified in step 4 of the healthcare failure mode and effect analysis. Numbering of critical failure modes and controls correspond to the markers in figure 2. ‘X’ indicates the hospital sites where the critical failure mode was identified. ‘C’ indicates the site where a control was identified for a critical failure mode at the other site. Numbers in square brackets correspond to the numbered subtasks in figure 2. ADC, automated dispensing cabinet; CS, controlled substance.
Develop action and outcome measures
The fifth step is to determine which failure mode causes can be eliminated or controlled and describe what actions could be used to accomplish this. This step also includes developing measures that can be used to test and analyse the success of a redesigned process. The HFMEA team will use the list of critical failure modes from the hazard analysis to describe each step in the MUP that increases the hospital’s potential risk for medication loss, including those related to both the security and accounting of medications. The team will consider the causes listed for the failure modes and describe mechanisms or actions that can be implemented to prevent, detect and trace incidents of medication loss. Finally, the team will suggest measures that could be used to assess successful implementation of these mechanisms and process improvements.
Patient and public involvement
Hospital personnel have supported this work by facilitating opportunities for observations and analysis of different aspects of the MUP in units that have high CS use and access. Healthcare providers and hospital staff will also be engaged during the HFMEA and will inform the dissemination strategy. Patients and public were not involved in the design of this study.
Ethics and dissemination
Ethics
This study has received province-wide Research Ethics Board (REB) approval via Clinical Trials Ontario Streamlined Research Review System, as well as site-specific approvals from each participating hospital under this framework.
Consent for observations is obtained from the healthcare worker who is being observed. When photographs are taken, no patients or healthcare workers will be photographed, and all person identifiers will be eliminated (eg, patient name/ID will be covered). Hospitals that choose to participate will remain anonymous and will be described using general terms (eg, a community hospital) in publications and presentations.
All signed consent forms, observation free-form notes, artefacts and photographs and database records will be kept secure and confidential. Observational data will be associated with a participant number to reduce the risk of participant identification. All data reported outside of the study team will be in aggregate form, without reference to any specific participant.
The observers are responsible only for collecting data as part of the study and will not perform clinical duties (eg, helping with tasks). However, in the unlikely event that observers suspect an error is about to be made that could compromise patient safety, observers will intervene by asking the participant for clarification, as indicated in the REB.
Dissemination
The audience for our research includes front-line hospital staff and administrators, as well as professional, regulatory and accreditation organisations, patient safety and healthcare quality organisations and equipment and drug manufacturers. The findings from our study will be used by organisations to inform recommendations, guidance and standards.
The results will be shared with hospitals in Ontario and across Canada through collaboration with the Institute for Safe Medication Practices Canada. Findings from this study will be presented at conferences and meetings, as well as in manuscripts submitted for publication. This study will be among the first to proactively capture empirical evidence of how current controls for MUPs in Ontario hospitals may be improved to protect against medication losses.
Limitations
It is challenging for observations to capture how participants actually conduct tasks, because participants may alter their behaviour due to the presence of the study team on the unit (ie, the Hawthorne Effect61). We will mitigate this effect by reassuring participants that results will not be used to evaluate individual performance but will only be used to describe an overall process. To minimise disruption and further normalise our presence, we will be as unobtrusive as possible and conduct several hours of observations at multiple sites with multiple participants.
The validity of the results is strengthened by accurate note-taking by the observers. However, it is possible that some subtasks or contextual features of the environment will be missed. To limit the extent of missing information, observers will receive an orientation to each unit before beginning observations, will ask clarifying questions while observing and will fully transcribe field notes. Two observers will capture MUPs in each unit, enabling corroboration and identification of tasks requiring further observation. Consistent study team members will observe, transcribe and analyse the data.
The subjective nature of identifying potential failure modes and scoring the probability and severity of their effects may compromise the reliability of the results.62 But, by using data collected through observations to conduct the HFMEA, the subjectivity of the hazard analysis is lessened by basing the work on observed behaviours as opposed to perceived actions based on accepted practices.58 63 64 To further limit threats to reliability, brainstorming failure modes, scoring probability and severity, and completing the decision tree will be conducted independently by two consistent members of the HFMEA team, with a third member reconciling differences when required.
Conclusion
It is expected that the clinical observations and HFMEA will lead to an understanding of the current workflows and failure modes affecting the MUPs in one community and one academic hospital. Results of this analysis will allow for a comparison of workflows, failure modes and controls between hospitals and as a function of clinical area (eg, emergency department vs intensive care unit). Identification of critical failure modes and controls will demonstrate where vulnerabilities exist for diversion or unintentional loss, and how they can be mitigated, including those related to the physical security as well as the documentation and accounting of CS.
Acknowledgments
The authors would like to thank Devika Jain for supporting the development of the study protocol and the submission for research ethics board approval.
References
Footnotes
Contributors PT, MF, MDV, DT and MH were involved in the conceptualisation of the study. MDV was responsible for drafting the protocol manuscript. PT, MF, DT and MH reviewed and revised the manuscript for intellectual content. All authors reviewed and approved the final version of the manuscript.
Funding This work was supported by BD – Canada (Becton Dickinson Canada Inc) [grant number ROR2017-04260JH-NYGH].
Competing interests MDV, MF and PT have received honoraria from BD - Canada for presenting at BD-sponsored events. BD - Canada was not involved in study design or writing of this article, and will not be involved in the collection, analysis or interpretation of data; in the writing of the report or in the decision to submit the article for publication.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement As this is a protocol submission, we do not have additional unpublished data to share.
Correction notice This article has been corrected since it was published. The article title has been revised.
Patient consent for publication Not required.