A primer for clinical researchers in the emergency department: Part IV: Multicentre research

Authors


  • Ed Oakley, MBBS, FACEM, Emergency Physician, Associate Professor; David McD Taylor, MD, MPH, DRCOG, FACEM, Emergency Physician, Professor; Tim Coats, MBBS, FRCS, FCEM, Emergency Physician, Professor; Andrew Davidson, MBBS, MD, FANZCA, Anesthetist, Associate Professor; Amanda Fry, RN, Research Coordinator; Franz E Babl, MD, MPH, FRACP, FAAP, FACEP, Paediatric Emergency Physician, Associate Professor.

Correspondence: Associate Professor Franz E Babl, Emergency Department, Royal Children's Hospital, Parkville, Vic. 3055, Australia. Email: franz.babl@rch.org.au

Abstract

In this series we address key topics for clinicians who conduct research as part of their work in the ED. Multicentre research is conducted to enrol larger numbers of participants and improve the validity and generalisability of the findings. Little has been published outlining the practical mechanics of multicentre research in the ED. This article presents strategies for setting up, conducting and participating in multicentre studies in the Australian and international settings. We address the two major categories of multicentre research, collaborative group research based on a group of researchers from a moderate-sized number of EDs and large-scale investigator-led research where a central group of investigators leads a large pool of research sites. Although we focus on clinical trials, the basic principles also apply to multicentre observational studies.

Introduction

Clinical practice is informed by a variety of types of research. In the past, clinical research in emergency medicine has been largely confined to single-centre studies. Well designed and performed single-centre studies are valuable and have changed practice; however, there are some research questions that are best answered with multicentre studies. Multicentre research provides, in particular, access to a larger number of participants and the results are likely to be more applicable to a variety of settings.

Although overlapping to a large degree, there are two major categories of multicentre research in the ED setting: first, collaborative group research, which is based on researchers who often know each other and is undertaken at a moderate number of research sites; second, large-scale investigator-led research, which involves a central group of investigators who lead a large pool of research sites. Although high-quality research can be performed in the single-centre setting or through one of the multicentre approaches, it might be more appropriate to use one or the other of the multicentre research categories for particular research questions. In this article, we discuss the advantages and challenges of multicentre research, explore the differences between the two major categories of multicentre research and suggest strategies to organise, conduct and facilitate their successful completion within the larger infrastructure that is required. Although we focus on clinical trials, defined as research in which a therapeutic, preventive or diagnostic intervention is tested,[1] the same basic principles set out here hold true for other valuable forms of studies, such as multicentre observational studies or audits. Important issues, such as the regulatory and ethical background of clinical research, development of the research question, the outcome measures and protocol development, have been addressed in Parts I and II of this series.[2, 3]

Advantages and challenges of multicentre research

The key reason to pursue a multicentre approach is the ability to recruit larger numbers of participants in a shorter time frame. This allows studies to be adequately powered to detect clinically relevant differences between study groups.[3-5] This is important in conditions where the desired outcome is infrequent but significant (such as death and disability) rather than more frequent but less clinically important surrogate outcomes. Large numbers are also needed if the likely differences between the groups are small. Further, in conditions where subgroups of patients might respond or behave differently, the larger sample size might allow adequate power to explore such differences in pre-planned subgroups. Emergency medicine is a pragmatic specialty and the evidence it is based on needs to be undertaken in a wide range of settings, with results applicable at all hours of the day and relevant for clinicians in a range of ED settings. Multicentre trials can increase the range of settings and variability in the population studied, improving the applicability or relevance of the findings to the broader ED population.

There are also substantial indirect benefits from multicentre research. To be applicable and acceptable in the broader medical community, research needs the support and participation of non-academic emergency physicians.[6] The networking, communication, joint research training required, and the sharing of resources, such as research templates and biostatistical expertise to set up and perform a multicentre trial, will build capacity for future research. The involvement of more centres and the resultant increase in research staff needed to perform the trial can help build an important base of trained and experienced research staff.[4] The improved ability to answer important clinical questions and the advantages of capacity building are attractive to funding organisations looking for the greatest outcome for their dollar investment.[4, 7] Participation in multicentre research with experienced grant holders will also allow site researchers to develop a track record for subsequent grant applications. A major benefit of multicentre research is that it is a vehicle to standardise and improve clinical practice across multiple sites for a given condition. Well-designed clinical trials aim to test an intervention in the setting of current best practice, which is embedded in the trial protocol. These best practices then tend to be sustained at participating sites, and by clinical staff, long after the research programme has stopped.

Multicentre research is, however, not without challenges. Although clinical research is inherently complex, this increases in multicentre clinical trials. The principal challenges are to maintain the internal validity of the study and the sustained commitment and collaboration of all sites over the study period. This means that sustained leadership, interpersonal skills, communication and determination are much more important to the success of a multicentre trial than the academic brilliance of the research idea – a good academic is not necessarily a good leader for a large trial.

Careful forethought is required for the development, implementation and execution of the study from its inception. This invariably increases the time and organisational requirements before and during its conduct.[8] The coordination of multiple sites requires personnel to maintain protocol compliance and communication, troubleshoot, and ensure recruitment rates are maintained. Data entry, data cleaning and data queries are more complex and time-consuming. There are often difficulties in arranging communication with sites, potentially over different time zones and using different languages as well as with different cultural expectations. Preventing possible disagreements over authorship, intellectual property and budget allocation require careful planning and trial agreements.

Ethics committee approval to run the project will be required at each study site. Traditionally, this has required an application to each site's committee. However, agreements between regions for multisite ethics approval are beginning to assist with this task and lessen the workload. International trials need to follow the local research governance procedures, which follow the same basic pattern, but vary widely in the details and speed of processing. Indemnity arrangements also vary widely, and might be prohibitively costly in some countries (such as the USA). Coordinating governance and administrative requirements for a large trial is a specialist skill that requires an experienced clinical trial unit and trial manager. Finally, the principal investigator has to be able to control the content of the protocol, and ensure there is agreement between all investigators on the study process. This can be difficult if there is not a mutual desire to succeed and a willingness to cooperate.[7, 9] In a large trial, it is obviously impossible for all participants to be involved in directing the research, so it is important for the principal investigator and the Trial Management Group to stay ‘grounded’ and to see the trial through the eyes of the local investigators at each trial site. Leadership skills (and time) from the core trial group are vital to foster a mutual desire to succeed and willingness to cooperate from all participating centres.

Types of multicentre research

Trials can be broadly grouped into sponsored and investigator-led studies. Sponsored studies, such as by a pharmaceutical company, are important for generating information to enable a drug or device to be brought to market. However, the focus of this article is investigator-initiated multicentre trials undertaken by non-commercial researchers, either as collaborative group research, or as large-scale investigator-led research. As set out in Table 1, collaborative group research is based on a group of researchers who tend to know each other and are collaborating because they are part of a pre-existing research network or because of shared interest in a specific topic. While often led by one or a small group of researchers, academic input is solicited from all participating researchers/sites.[10] In large-scale investigator-led research, a core group of academic researchers invites a large number of sites to participate. Academic input and control in this larger setting is only by the core group, whereas the other sites only contribute patients to the study. An example of such a large-scale investigator-led multicentre study is the CRASH-2 study, conducted at 254 EDs in 40 countries.[11]

Table 1. Research settings: single-site research, collaborative group research and large-scale investigator-led research
 Single-site researchCollaborative group researchLarge-scale investigator-led research
FormatSingle or multiple investigators at a single site.Group of investigators from a few research sites (<10–12 sites).Group of investigators, with large pool of collaborating research sites.
Type of questionExperimental. Hypothesis formation.Explanatory. Ranging from initial testing to definitive study of specific topic in defined setting.Pragmatic. Generalisable.
Type of effect that can be studiedLarge effect in a common condition.Moderate effect in a common condition or large effect in a rare condition.Small effect in a common or rare condition.
Infrastructure requiredResearch time and staff.Research time and staff plus facilitation of communications. Study coordinator.Research time and staff. Clinical trial unit support. Trial manager.
Academic inputInvestigator has all academic input.One leader, but all investigators have significant academic input.One leader with small steering group. Others recruit patients but have little academic input.
OrganisationInformal contact with peers or mentor. Data Monitoring and Safety Committee (DMSC), Trial Steering Committee (TSC) depending on study.Trial group in regular contact. DMSC. TSC.Central trial management group. DMSC. TSC.
Formality of relationshipsInformal.Implicit or explicit (written) understanding between all parties.Legal contracts define relationship with research sites.

Unless the effect size is very large and in a very common condition, single-centre research can often only address an experimental question or help form a hypothesis. Collaborative group research can provide definitive answers to less common conditions but is often limited in the applicability of the results to a specific setting. Large-scale investigator-led studies can investigate small effect sizes and provide answers that are generalisable and pragmatic for a broad range of settings. Increasing number of sites increases the need for funded infrastructure (experienced clinical trial unit and trial manager), organisational support (Trial Steering Committee [TSC] and Data Monitoring and Safety Committee [DMSC]) and legal contracts to define relationships between participating sites.

Creating a team

Collaborators for multicentre research can be found through colleagues and friends, eminent people in the field, or existing research networks. For example, much multicentre paediatric emergency medicine research in Australia and New Zealand will engage interested sites through the local research network – Paediatric Research in Emergency Departments International Collaborative.[10] Collaboration with organisations possessing research skills and experience, such as universities or research institutes, might be beneficial.[4, 8]

The selection of collaborating centres is critical. It needs consideration of possible recruitment figures, previous recruitment records, research infrastructure (clinical research time, research staff and administrative support), whether the local ethics and administrative systems understand emergency care, and the representative spread of types of centres (large urban EDs vs others). Although past performance is the best guide to future success, enthusiasm and motivation are key ingredients and there should be room in any collaborative research for the enthusiastic new researcher.

As important as finding a team of collaborators, is the creation of an ethos for a multicentre trial that makes the staff in each research site feel part of the team, rather than just functionaries who are doing the hard ‘leg work’ of the research for a bunch of academics sitting in an ivory tower. This is particularly true for large-scale investigator-led studies. The academic research leaders have to work hard at generating this positive feeling, which is difficult to define and measure, but is vital to success. Motivation for participation will not be funding (investigator-led research is not generously funded) or names on papers (especially in large-scale trials there are too many names for all to be on the paper with recent examples being the 3CPO[12] and CRASH-2[11] studies). Each local site lead needs to be motivated by wanting to improve patient care, a powerful motivation for doctors. They must feel that the research question is relevant, and that it addresses a question that they need answers for in their daily practice. Creating a team feeling is the main role of the academic leaders during the trial, both for the trial as a whole and for the site representatives at each trial site. This is a large investment in time and energy, with communication as the main task (see Communication). The trial leaders for the study overall, and for the site leaders at each recruitment site, need to be visible, positive and enthusiastic. The trial leaders need to travel, meet, talk and invest time and energy to create a positive ‘team feeling’, preferably supported by funded research staff to reinforce, motivate and assist with troubleshooting.

Research organisation

Developing the research protocol

Once collaborators have been found, the research protocol needs to be developed in detail. The initial study proposers usually take the lead with this and regular discussion with the study group will refine the detail.[4] The participant population should be described in detail (inclusion and exclusion criteria) and the study outcomes very well defined. Collaboration with a statistician from the beginning is important to ensure adequate sample size calculations and that the data collected will answer the questions posed. A decision on the need for early pilot data to assist with sample size calculation will need to be made.[8, 13] Particular thought should be given to likely adverse events related to the study and documenting the processes for informing the Ethics and Data Monitoring and Safety Committees.

Piloting of the research tools

Once the protocol and data collection tools have been developed, piloting their implementation at one or two sites can identify improvements needed and practical problems that are likely to arise in the conduct of the study. The entire study protocol from patient screening, enrolment, randomisation, intervention, data collection, data entry, data storage, and analysis can be tested and refined from a pilot study before conducting the definitive trial. In addition, a pilot study can evaluate the utility of the data collection instruments and case report forms. Modifications and improvements to the documents might follow.[8, 14]

Production of the study materials

This is led by the principal investigator(s) and the study coordinator/trial manager and with input and approval from the wider research group in the case of collaborative research. It is important that research material be generic and identical, in order to be used across all sites, with consideration to branding, readability and visibility. Branding with consistent colouring and logos of all material helps to establish a study in the minds of researchers and clinical staff involved. These materials are vital for the education of the broader study personnel and ease of use. In addition to study materials, the package for distribution to sites should include materials for education and advertising (see Box 1). All materials can be fitted with local hospital names and logos. Research materials will need to follow local ethics requirements. It is vital to keep study material as simple as possible. Clinicians in emergency care do not have time for complex data recording – one side of paper at each stage is ideal. For example, in CRASH-2[11] there was one side of A4 for trial entry and one side of A4 for outcome. Minimising data collection requires the investigators to be ruthless in eliminating the ‘nice to have’ data points leaving only the ‘must have’. It is much better to collect a small complete dataset than to have a large amount of data, but with missing values. The simpler the study design and data collection, the more likely a trial is to succeed.

Box 1. Study materials to be prepared for multicentre research

Case report forms

Created in a generic fashion in which individual sites can add their hospital name and logos, but including site-specific study numbers

Detailed explanations for case report forms

Each data point should be thoroughly explained in order to achieve identical data collection across all sites

Should be self-explanatory

Participant information statement and consent form (modified per local ethics requirements)

Log books

For enrolled and randomised participants

For potential participants, including missed, excluded and refused

Logs to include study number, patient name, date of birth, unit record number, key individual study data (admission date, intervention, discharge date, contact details); for non-enrolled patients (i.e. missed, excluded and refused patients) extent of log information depends on ethics committee guidance

Promotional material for the ED

Posters and signs for ED staff

PowerPoint slides for talks in the ED or wider hospital

Data query forms

Audit tool for site visits

Communication list with names and contact details of all study staff

Every page of all data collection documents (e.g. clinical report form) and those with identifying information (e.g. consent form) needs a space for unique study numbers different from the participants' medical record numbers. It is helpful to include site codes into the study number for easy identification of site-specific information. Often, patients will need multiple study-related numbers, such as overall study number (includes all patients, including missed, excluded and refused), enrolment number (all enrolled patients) and a separate randomisation number (randomised patients only). The numbering systems for all study patients need to be carefully thought out before assignment and distribution to sites.

Log books (which link identifiable patient data to the non-identifiable research data) are crucial to keep track of study patients at each site. Often, hard copy and electronic copy logs are created. Electronic logs must be password-protected. Once patient records are de-identified after study closure, the patient logs are the only way to re-identify study materials.

Ideally, all study materials and associated promotional materials are based on existing well-designed documents from other successful studies, which can be modified.

The study coordinator needs to keep copies of all versions of all site-specific documents, including protocols, clinical report forms, consent forms, and ethics approvals and communication.

Training of site research staff

All site research staff should have identical training that provides a thorough understanding of the study protocol, and good clinical research practice (GCRP) training that conforms to the local standards. This will include principles of good research practice, ethics requirements, study requirements and documentation standards.[8] This can occur at individual sites. However, depending on the size of the study and the geographical spread of sites, for collaborative group research this might be more practically and efficiently arranged centrally with all researchers attending a seminar that encompasses education on both the study content and GCRP. Elements of such a seminar are listed in Box 2. For studies that span multiple years, consideration should be given to running a similar seminar annually to ensure all researchers have equivalent and consistent knowledge of the study and GCRP. A coordinated education package should be produced that can be tailored for each site to educate all persons likely to be involved in managing participants in the study.

Box 2. Content of multisite research study seminars for site investigators and research staff

The protocol and background

Completing clinical report forms and case documentation

Adverse events and serious adverse events

Good clinical research practice

Consent process

Communication between sites

Study binder

A study binder (Site File) must be created by each site, in order to help research teams identify, maintain and organise all documents that are crucial for conducting and sustaining a clinical research study. This binder must be identical across all sites. The materials should be comprehensive and self-explanatory and allow the continuation of the study if key personnel leave the study or become incapacitated.

Making participation easy

To increase generalisability and to recruit the required number of patients, it is important for large-scale investigator-led studies to include EDs that are not traditional academic centres. Outside the traditional research centres there is likely to be little infrastructure support for research, so the design of the study must make it as easy as possible for these centres to take part. The administrative burden of the research has to be taken on by the clinical trial unit, as it might be too much to expect the local investigators to do the paperwork associated with their local research governance structures. The local researcher will have to sign the forms, but the time-consuming task of filling in the forms needs to be done by the central trial administrative team. The design of the trial must make it easy for a busy clinician to enter/recruit the patient. The randomisation system must be easily accessible at all times, with a back-up system in the event of technical failure. The intervention must be as simple as possible. For example, in the CRASH-2 trial it was decided that an infusion pump would not be required, as this would present an additional barrier to patient recruitment. This might seem trivial when designing the trial, but in the hurly-burly of a trauma resuscitation, finding and setting up a pump might well be a significant barrier. The data collection form must be as simple as possible. Finally, the central trial team must produce materials that make it easy for each local site lead to engage with colleagues, such as pre-prepared presentations of the trial background, rationale and procedures. There are obviously some things that have to be done by each local lead, but the central trial team or clinical research unit should take on as much of the work as possible, leaving the minimum that can only be done locally. The trial team should also be easily available (both administrative and academic staff) and responsive to questions and inquiries from the research sites, so that local leads feel supported.

Funding

Multicentre research is resource-intensive and usually requires external funding. It is important to include the additional costs of the coordination of the participating sites in the initial costing. Funders expect realistically costed applications – so cutting corners to make your project look ‘good value’ will be immediately spotted and makes it easy for the application to be rapidly rejected as ‘unrealistic’. Potential funding sources need to be identified early and, once the protocol is finalised, grant writing should be given priority to avoid any delay in the commencement of the project. Expertise in study design, set-up and analysis should be sought from experienced researchers with a proven research track record in the type of study concerned, or in the field under investigation; from collaborating research institutes and university departments; and from within the research group. Funding options include large-scale national funding (e.g. National Health and Medical Research Council) for the entire project, or smaller grants for each research site to provide the dollars and personnel necessary to adequately perform the study.[4, 13] It needs to be remembered that rejections from grant bodies are common (even for experienced academics) and persistence is needed to achieve success.

Investigator-led trials, even if large, are never going to be as well-funded as industry-sponsored trials and will, inevitably, partly depend on the goodwill of clinical staff and institutions. However, there are key differences between industry trials (mainly directed at obtaining regulatory approval for a product license) and investigator-led trials (mainly directed at answering a specific question). This difference in perspective often makes the investigator-led approach much less costly. In the UK, past experience of successful grant applications gives a rough estimate of the total cost of investigator-led clinical research of about A$4000 (£3000) per subject entered into the trial (pers. comm., Prof. T. Coats, Leicester, UK, 2012). A recently completed collaborative group research randomised controlled trial of 750 patients at seven EDs in Australia and New Zealand[15] cost A$1400 per randomised patient. Although multicentre trials need to be realistically costed, the funding application will be more competitive if each participating hospital contributes some staff time. In return for contributing this resource to the trial, the institution gains the benefits of being ‘research-active’, which include enhanced reputation and staff satisfaction, and an improvement in clinical standards and quality of care.[16] Cost (usually staff time) to each participating institution should be a large consideration in the design of the trial, in particular minimising the volume of data collection and making the follow-up schedule as easy as possible.

Governance

Each research site will have its own local research governance system – the requirements of which might vary considerably.[17] The core trial unit should assist local investigators as much as possible in fulfilling these local requirements.

Agreements

There are many issues in running a study and publishing the findings that can cause tension between the researchers. These issues are best addressed early, and should be thought of as part of the study setup. Such issues include the budget for each site, responsibilities of each researcher member, recruitment targets, managing approaches from media, and a process for managing disagreements between the researchers.[7] Membership of the writing group and authorship (and even the likely authorship order) of the major study manuscripts need to be decided at the outset. For large groups, it might be better for the author to be ‘The X Trial Collaborators’ with a list of contributors. Many institutions will require legal ‘Clinical Trials Agreements’ between collaborating institutions. When preparing these documents, try to keep to accepted templates (such as a modification of the Medicines Australia Clinical Trial Research Agreement) as any deviation from accepted wording might require new legal opinion with associated cost and delay.

Oversight committees for the study

At least two oversight committees should be formed for the study. A TSC should comprise the principal investigator and one or two other investigators, a multicentre research expert, and other research experts who can provide advice on the running and implementation of the study. This group should oversee the entire study and have power to resolve conflicts between researchers. Decisions from it should be binding. In some circumstances, it might be worth appointing an independent member, not an investigator, as chair of the TSC.

A DMSC should comprise a small number of researchers independent of the study who can review the study data regularly and impartially. The frequency with which this is done is determined by the risk profile and rate of patient recruitment.[8] An overview of the ethics and governance of the research will be kept by the DMSC committee, which, during the trial, will consider any ethics issues arising, any new information that is published, review adverse events, and will periodically review unblinded data against the trial stopping criteria. If stopping criteria exist, the DMSC will determine if these criteria have been met and take appropriate action if required.

Data handling

Ideally, all data should be entered into the database as soon as they become available and collated with those from all participating sites. Data errors or uncertainties need to be clarified with the respective site research team on specified forms. A timeline for response should be included. This allows for a close monitoring of recruitment, data quality and progress of the study.[18] A track record of all site communication and responses needs to be kept.

If paper-based, clinical report forms should be de-identified at individual sites before their dispatch to the central site. This can be done by blacking out or removing identifying information. The original forms will be stored at the central site, whereas the peripheral site will store a photocopied version. If electronic data entry systems at each site are in place, secure password-protected back-up needs to be put in place.

Processes for data handling, transfer, data validation and quality assurance should be outlined with clarity around reporting lines and responsibilities for correcting errors.[4, 8]

Tracking individual site performance

Recruitment is usually reviewed about every 3 months. Each site should have an individual recruitment target to allow the trial as a whole to deliver to ‘time and target’. There will be variation in the rate at which each site recruits patients (key staff off sick, pressure on the clinical system etc.), so the central trial team needs to have a monitoring system and an intervention plan if a centre falls behind. This plan might require considerable communication with the site lead to identify the problem, usually a site visit and an agreed plan to improve. This usually needs input from both the trial manager and the clinicians leading the trial followed by a ‘relaunch’.

An audit plan should be included in the trial protocol with specific audit triggers, which might be related to time (annual etc.), recruitment (audit after X participants), on data review (one centre appears to be an outlier on demographics etc.) or in response to a GCRP issue.

Site visits should be based on a pre-planned schedule, which usually occur after a specific time or after a specific number of patients have been recruited. There are also triggered site inspections, which might be initiated by complaints, adverse events or unusual data patterns (too many patients, too few patients, too many missed patients, data completeness too good to be true or an outlier on the characteristics of recruited patients). There might be additional site inspections by the national regulatory organisations. An audit tool should be developed and issues should be documented and formally fedback to the site research team and the TSC. A timeline for rectification of any problems should be included and followed up. A central data monitoring system will highlight sites that have unusual patterns in their data, which can be followed up by a site visit.

Study closure

A discussion on the procedures for study closure should occur well in advance. This allows sites to plan staff employment and orderly closure at each site. Procedures for the termination of recruitment, final data queries and storage of the data should be clearly set out in the trial protocol.

Communication

Once participant recruitment has commenced, communication between researchers is a key issue.[7-9] Telephone/video conferences should occur regularly and frequently (depending on the complexity of the study) between site investigators (led by the principal investigator and study coordinator) and between site research assistants (led by the study coordinator). This allows all to learn from the issues faced at each site. Initial considerations include patient recruitment, study flow, staff engagement and data management. Other issues for consideration include missed patients, adherence to the study protocol, questions about inclusion and exclusion criteria, and functioning of the data collection tools.

Review meetings should be scheduled at predetermined intervals to discuss site performance, finances, and to plan the next steps, including study closure.

Site visits should be undertaken by the study coordination team and the principal investigator. This should focus on principles of data quality, good research practice, data collection and storage. It is an opportunity to address participant recruitment and retention problems, and to identify any deficiencies in the knowledge or capability of the research personnel.

Rewarding performance can help improve recruitment and maintain awareness of the study. Little prizes on reaching recruitment targets or for the top-recruiting site might foster healthy competition between sites. Small items like trial-branded pens or mugs can be given out to staff as a reward for recruiting a patient – this also has the benefit of publicising the trial. Other items such as mouse mats for the ED can carry trial information such as the inclusion criteria.

A newsletter should be produced to maintain researcher and ED staff interest and disseminate crucial information. It should include information on study recruitment and highlight well-performing sites. A human interest section highlighting study staff can engender team spirit and enthusiasm.

Creating and maintaining a team spirit is particularly important and challenging in large-scale investigator-led studies. Multiple methods of communication need to be used in addition to newsletters, including journal editorials, updated websites, podcasts, YouTube videos, local meetings and satellite meetings at conferences. A ‘hotline’ for collaborators, which allows direct contact with one of the lead researchers, should be set up.

Conclusion

Multicentre research brings with it complexities and challenges not seen in single-centre studies. However, careful planning and implementation will be more likely to result in successful completion. A sense of teamwork, satisfaction and achievement will be gained, and research knowledge, infrastructure and linkages made that will aid in future research endeavours.

Acknowledgements

We acknowledge grant support from the Murdoch Children's Research Institute, Melbourne, Australia and the Victorian Government's Operational Infrastructure Support Program.

Competing interests

DMT is a Section Editor (original research) for Emergency Medicine Australasia.

Ancillary