The Internet as a research site: establishment of a web-based longitudinal study of the nursing and midwifery workforce in three countries
Abstract
Title. The Internet as a research site: establishment of a web-based longitudinal study of the nursing and midwifery workforce in three countries.
Aim. The aim of this paper is to describe the development of a web-based longitudinal research project, The Nurses and Midwives e-cohort Study.
Background. The Internet has only recently been used for health research. However, web-based methodologies are increasingly discussed as significant and inevitable developments in research as Internet access and use rapidly increases worldwide.
Method. In 2006, a longitudinal web-based study of nurses and midwives workforce participation patterns, health and wellbeing, and lifestyle choices was established. Participating countries are Australia, New Zealand and the United Kingdom. Data collection is handled through a dedicated website using a range of standardized tools combined into one comprehensive questionnaire. Internet-specific data collection and a range of recruitment and retention strategies have been developed for this study.
Discussion. Internet-based technology can support the maintenance of cohorts across multiple countries and jurisdictions to explore factors influencing workforce participation. However, barriers to widespread adoption of web-based approaches include website development costs, the need for fast broadband connection for large data collection instruments, and varying degrees of Internet and computer literacy in the nursing and midwifery workforce.
Conclusion. Many of the issues reported in this paper are transitional in nature at a time of rapid technological development. The development of on-line methods and tools is a major and exciting development in the world of research. Research via the world-wide web can support international collaborations across borders and cultures.
What is already known about this topic
- •
There is a lack of empirical nursing and midwifery workforce data.
- •
The Internet is gaining importance as a research site.
- •
Nurses and midwives are required to be familiar with, and use, the Internet in clinical practice.
What this paper adds
- •
Description of Internet-mediated research processes supporting the maintenance of international cohorts in a longitudinal study.
- •
A variety of follow-up strategies are required to ensure ongoing email contact with the cohort.
- •
Access to fast broadband Internet is critical for nursing participation.
Implications for practice and/or policy
- •
Internet-mediated research can support international collaborations aimed at addressing nursing and midwifery workforce issues.
- •
Internet access and literacy issues will lessen, as e-technology is increasingly a core aspect of nursing and midwifery education and practice.
- •
Current workforce data are essential in a rapidly changing global context to inform workforce policy and health service delivery.
Introduction
In 2005 an innovative, web-based longitudinal cohort study of the nursing workforce in Australia and New Zealand (NZ) was established – The Nurses and Midwives e-cohort Study ‘http://www.e-cohort.net’– and in 2006 the study was extended to the United Kingdom (UK). The aim of this research is twofold: first, to examine factors associated with the recruitment and retention of nurses and midwives by exploring their workforce participation patterns, health and wellbeing, and lifestyle choices and, second, to pilot the use of an electronic survey method for a longitudinal study with a number of international nursing and midwifery cohorts.
The establishment of a comprehensive, longitudinal web-based study is novel and the literature on the implementation of web-based studies in health is sparse. In this paper, we discuss issues pertinent to Internet-mediated research, and then go on to discuss the Internet-specific data collection and recruitment strategies used in this study. Finally, the advantages and disadvantages identified through the establishment of the Nurses and Midwives e-cohort Study are discussed. A demographic profile of a partial cohort analysed during the recruitment phase has been published by Turner et al. (2008).
Background
There is an increasingly intense focus on nursing and midwifery workforce issues as awareness of the impact of current and future shortages has grown and the multi-factorial and complex nature of the problem has been identified [International Council of Nurses (ICN) 2006]. Although specific issues differ between countries, it is clear that retention and recruitment of the nursing and midwifery workforce requires urgent attention. Consistently, when the issue of nursing and midwifery shortages and the challenges faced in terms of recruitment and retention are discussed by bodies such as the ICN (2006), the ICM (Koblinsky et al. 2006) and World Health Organisation (WHO) 2006, the lack of contemporary workforce research is highlighted and the need for robust data identified.
The Nurses and Midwives e-cohort study was established in this context of increasing recruitment and retention concerns and in response to calls for research to underpin future policy and planning (International Council of Nurses 2006, WHO 2006, 2007). It was also envisaged that this study would make a significant contribution in two key areas other than nursing and midwifery. First, the project is designed to collect a considerable amount of data on women’s health due to the population of interest being predominantly female: 95% in NZ (Clark 2006), more than 90% in Australia (Australian Productivity Commission 2006) and 89% in UK (Nursing and Midwifery Council 2006). The second key area of anticipated benefit is to advance epidemiological methodology by establishing and maintaining an electronic cohort who complete annual Internet-based surveys. The primary mode of Internet-mediated engagement is not yet an aspect of typical epidemiological studies (Turner et al. 2008).
A web-based approach relies on participants having access to the Internet, and both Internet access and use is rapidly increasing worldwide. In Australia, Internet access has risen to 60% or 4·7 million households (Australian Bureau of Statistics 2007); in NZ, 2006 census data shows Internet access at 1 million households, and 69% of individuals had used the Internet in the preceding year (Statistics New Zealand 2007). The Oxford Internet Institute (2007) identifies Internet access in the UK as of 2005 as approximately 60%, with the most favoured place of access being at home. The quality of Internet access is also important for Internet-based research. By March 2007 broadband access in Australia had increased to 67% of total Internet subscribers (Australian Bureau of Statistics 2007), which was also the case in the UK, with 61% of access being via broadband (Oxford Internet Institute 2007). Broadband access in NZ is lower, with 33·2% of households having it and 30·9% still using dial-up connections (Statistics New Zealand 2007).
Although Internet access in the study countries appears to be in the range 60–69%, it is important to recognize that these figures are based on data from the total population. Internet access is significantly higher in professional groups (Australian Bureau of Statistics 2007). In a study of New Zealand postgraduate nursing students, 95% of respondents had home access and 87·5% access at work (Gilmour et al. 2008). Based on this evidence, it is probable that the Internet will become an increasingly commonplace site for research involving professional groups. Although Internet surveys are used in some research areas such as communications, and in disciplines such as sociology, the Internet has only recently been used for health research but is increasingly being acknowledged as an inevitable and significant development in the area (Eysenbach & Wyatt 2002, Scriven & Smith-Ferrier 2003, Bälter et al. 2005, Baernholdt & Clarke 2006, Ekman & Litton 2007). The Nurses and Midwives e-cohort Study is one of the few studies attempting to recruit and maintain a longitudinal cohort using an electronic web-based mode.
Although Internet use is increasing, there is little research to date on its use by nurses and midwives, although a study of Canadian nurses showed that use had increased, particularly at home (Estabrooks et al. 2003). Use of information communications technology (ICT) generally, and the Internet specifically, is now considered a core aspect of modern practice and is a basic element in all nursing and midwifery programmes (Wishart & Ward 2001, Nursing Council of New Zealand 2005, Gilmour 2007). This development reflects the requirement that nurses and midwives source evidence to support clinical decision-making, and that they are able to advise patients in relation to the validity of material readily available on the Internet. In the past, while web-based studies may not have been considered appropriate based on assumptions related to nurses’ and midwives’ possible general lack of confidence in Internet and computer use, it is clear that this situation is rapidly changing.
There were two main reasons for choosing a web-based design for this project. First, the cost of maintaining traditional paper-based longitudinal studies is considerable, whereas Internet-based survey administration and participant follow-up costs appear substantially lower. Second, it was envisaged that recruitment and retention of participants would be enhanced by using a website designed to engage and interact with individual participants. In a recent study comparing response rates and compliance in web-based and mailed questionnaires, Bälter et al. (2005) found that interactivity in Internet-based research led to increases in retention and completion of subsequent questionnaires. It was also thought that use of Internet-based technology would facilitate the maintenance of cohorts across multiple countries and jurisdictions so that, for the first time, comprehensive international data on workforce participation, health and wellbeing and lifestyle could be collected using a range of data collection instruments.
Methodology and methods
As multiple issues affect the practitioner shortage, a comprehensive survey was designed to provide workforce participation data in combination with data on health and wellbeing, and lifestyle challenges – factors which have an impact on practitioners’ ability or willingness to remain in the workforce. It was also considered essential to formulate a study which captured trends and patterns over time. The strengths of a longitudinal design are well-known and centre on the ability to follow an individual’s development over time and to correlate this with exposure to a range of variables in that period (Bryman 2004). The ability to make causal statements, together with significant biostatistical and computational advances, has increased the popularity of this study design in recent years. In addition, the design is considered flexible and allows new and contemporaneously important questions and issues to be introduced and obsolete questions and issues to be deleted as the study progresses.
The original design of the Nurses and Midwives e-cohort Study was undertaken by researchers at The University of Queensland, Australia in 2005, and from the beginning an international scope was envisaged for the research. New Zealand was invited to collaborate in the study at this stage. In common with many international research collaborations, multi-disciplinary teams of researchers were initially established in Australia and NZ, with the overall Project Director at The University of Queensland. The synergies of a multi-disciplinary team with in-depth knowledge of workforce issues, functional relationships within the disciplines – particularly with regulatory authorities – coupled with extensive experience in longitudinal study design and implementation, has been a factor in the successful establishment of the project. Team protocols have been developed for data access and registration of publications, along with the development of a data dictionary. Following the initial establishment of the study, in 2006 the UK became involved, largely through connections with individual researchers and the Nursing and Midwifery Council.
Data collection is handled through a dedicated website: http://nurses.e-cohort.net. The potential advantage of this design is that data can be gathered, analysed and reported in comparatively short timeframes with fewer resource pressures. The use of ICT virtually eliminates the expense of data entry associated with traditional paper-based epidemiological studies. However, considerable resources are still required for follow-up procedures to ensure retention and for tasks such as data-cleaning. Direct data entry, however, does allow for rapid follow-up of non-responding participants via email. Early access to the data is also possible, reducing the time-lag between entry and publication of findings. The time-lag in the analysis of data and publication in traditional research can be a significant issue for planners, and can result in workforce policy being based on data which, in this rapidly changing sector, can be out of date. A particularly salient advantage of a web-based longitudinal study is that participation can continue irrespective of geographical location and employment status. As long as participants can access the Internet, for example through an Internet café if travelling, they can remain in the study. The identification of particular groups of participants for inclusion in sub-studies is also possible, for example those working in aged care or on night duty, or those planning to retire. A question can be included requesting participants’ agreement to be approached for further studies focused on particular issues or for qualitative research purposes.
The research team in Australia worked with ICT specialists to develop a front-end study website which interfaces with a secure structured query language (SQL) server database. Participants register on the website and a unique study identification number (ID) is automatically generated for research purposes; this matches survey responses entered by each participant. Contact information is provided by participants through their personalized e-portal, which they can log into and update at any time. This information is held in a separate database from the survey responses to ensure confidentiality of survey data. The website design brief included making the site as attractive, user-friendly and relevant as possible for nurses and midwives, and included the design of a unique study logo for all promotional material. Photographs depicting aspects of clinical practice were used on the various pages, and these images change frequently to ensure that the site maintains a fresh ‘look’. The site also includes public access pages where current research projects and publications arising from the research are posted to keep participants informed about the outcomes of the study. Other pages profile the lead investigators and give further information on associate investigators and sponsoring organizations. Links are also provided from the site to those of sponsoring organizations such as funding bodies and regulatory boards and councils.
Instruments
A range of internationally known and extensively used standardized tools were included in one comprehensive questionnaire for each measurement wave. The annual questionnaire collects data on demographics and a range of variables relevant to the three main areas of the study –‘Workforce Characteristics’, ‘Work-Life Balance’ and ‘Staying Healthy’. Items in the section ‘Workforce Characteristics’ include employment status, hours worked per week and employment settings, such as public or private health providers. The ‘Work-Life Balance’ section investigates a range of workplace attributes and job stresses, along with family and study responsibilities. Three measures are used to measure job stress and workplace attributes, (i) the Job Content Questionnaire (Karasek et al. 1998), (ii) Effort Reward Imbalance subscale (Siegrist 1996) and (iii) hostility score (Koskenvuo et al. 1988). Musculo-skeletal health is measured using a modified Nordic Musculoskeletal score (Kuorinka et al. 1987). The ‘Staying Healthy’ section records physical and mental health variables using several validated instruments, such as the (i) International Physical Activity Questionnaire (Craig et al. 2003), (ii) Centre for Epidemiologic Studies Depression Score (Radloff 1977) (iii) SF-36 (Version 2) (Ware et al. 2002) and (iv) Diet Quality Score (Collins 2004). Other items include disease diagnoses and health screening behaviour.
These international data collection instruments have been subject to validity and reliability testing, and their use enables comparisons between other countries as well as between NZ, Australia and the UK. In the case of Australia, comparisons can also be made between states and territories. For example, the psychometric performance of the International Physical Activity Questionnaire (Craig et al. 2003) was tested for reliability and validity across in 12 counties, and showed comparable measurement properties to other self-report instruments. The Job Content Questionnaire (Karasek et al. 1998), which has been used in very large studies, has been tested in six populations from four developed nations for across-country reliability, which was found for most scales.
These tools were combined and loaded into survey software customized for the project. The sample database enables the questionnaire timing to be variable (annual/bi-annual) and various subgroups can be identified and targeted. The instruments have cycles, with the core questions administered on an annual basis and the others administered in variable cycles depending on the nature of the questions and the issues being explored. This is also important to reduce the burden on participants in terms of the time required to complete the questionnaire.
An iterative process took place between the Australian and NZ teams to determine the overall structure of the questionnaire and to ensure that wording was consistent with other data sources such as census instruments and labour force research. For example, in the demographic section, ethnicity choices were extended to mirror the parameters used in census documents, and physiological measures were standardized to ensure consistency across countries. The questionnaire was piloted with a random sample of nurses (n = 100) selected from the email database of the Queensland Nursing Council. The pilot participants gave feedback on layout, content, and technical issues. Completion data was requested of all pilot participants and the average time taken to complete the questionnaire of 108 questions was 40 minutes. Members of the research team also completed the pilot survey to evaluate ease of navigation around the site and the measurements. Modifications were made on the basis of all the feedback received. Slight changes were also made after ethics review by The University of Queensland and the Massey University (NZ) Ethics Committees.
Participants complete the questionnaire, which takes at least 40 minutes; the time required can vary between participants due to issues such as speed of Internet connections. In recognition of the time required to complete a questionnaire of this size and possible poor quality Internet connections such as dial-up, the survey tool stores data in a relational database in real time, on a question-by-question basis. This means that data are not lost if the participant’s connection times out or if the survey window is inadvertently closed. When logging on to the site in the future, the participant is taken back to the last completed question. This facility also means that practitioners can complete the survey over several sittings. The ITC system is not able to determine how many participants complete over several sessions. However, it does identify how many complete, partially complete and have not started the survey at any given time. Sixty-six per cent of participants who registered went on to complete the survey straight away.
Recruitment
The process of contacting nurses and midwives for recruitment was via the regulatory bodies in NZ and the UK and in each State and Territory in Australia, as these are the only organizations that engage with every Registered Nurse and Midwife for regular license or practising certificate renewal. In Australia and the UK many midwives also hold nursing registrations and/or are regulated by the same authority as nurses. In NZ, however, midwifery is a completely separate discipline from nursing and has a separate regulatory council which, through choice, is not currently involved in the research, although some individual midwives who are also Registered Nurses have opted to participate. The process for contacting potential participants varied between countries and jurisdictions within Australia, depending on the form of engagement determined by individual regulatory authorities. Study proposals were submitted to the authorities in NZ and Australia requesting financial or ‘in-kind’ support for recruitment purposes.
The response from councils and boards differed. In NZ the Nursing Council provided both direct financial and in-kind support by including a hardcopy invitation to participate with annual practising certificate renewal documents. In Queensland and Tasmania this form of in-kind support was also provided. Council and boards in the Australian states of New South Wales, Western Australia and the Australian Capital Territories facilitated a paid mailing inviting all practising nurses and midwives on their database to participate. In the UK, the Nursing and Midwifery Council agreed to publish an advertizement promoting the study and the website address in their quarterly newsletter.
In four states and one territory in Australia and NZ all Registered Nurses, Enrolled Nurses and nursing assistants and, in Australia, midwives received an initial hardcopy invitation to join the study. Those who wished to participate registered for the study by logging on to the website and recording their basic personal details. Participant’s name, date of birth, practitioner registration number and a home or work email address are mandatory for registration. The participant’s email address and date of birth are entered as the username and password for all future logins to the website. A back-up email address, phone number and alternative contact are all requested to enable follow-up contact if required. During the registration process, participants access all the standard research documentation such as the information sheet, consent, ethics approval details and information about the research team. Once registered, participants are able to view and complete annual surveys. Two weeks after registering, participants receive a personal thank you email - if the survey has not been completed, the email contains details on how to complete it as a reminder.
As the initial contact with potential participants was only through either a hardcopy invitation or advertizement, the study has required intensive promotion at all levels of the workforce to raise its profile. An important aspect of this promotion was gaining the support of those in senior leadership positions in policy, regulation and planning areas of nursing and midwifery workforce management. This group had to be convinced that this research would be of use, as support could be translated into practical measures such as allowing nurses to complete the survey on work-based computers and during work time as workload permitted. Different approaches were taken to promotion in NZ and Australia, reflecting the very different contexts. As NZ is a comparatively small national jurisdiction, a country-wide approach could be taken. In Australia, the nursing and midwifery workforce is regulated by individual States and Territories and management and organizations can differ between jurisdictions. Thus, the active promotion of the study has taken many forms, for example presentations to Directors of Nursing and employers, attendance and presentations at national nursing conferences, meetings with professional nursing organizations and promotion through their media such as newsletters, websites and circulars, general media interviews and articles in nursing ‘magazines’, and presentations at a wide variety of professional development forums such as grand rounds and clinical specialty and workforce meetings.
As with all such studies, the success of the project depends on potential participants deciding that being in the study has some value in terms of possible impact on their working environment and/or the profession at large. In addition, if sufficient participants join the study, specific issues for various sub-populations could be explored. It has, therefore, been critical to promote the value of the research and explain that data will be used to provide a comprehensive profile of the profession as a whole which can inform policy and planning for recruitment and retention purposes. Taking part in this type of research was clearly going to be a new experience for practitioners, many of whom would be much more familiar with paper-based surveys.
Cohort retention
To retain participants, a range of strategies have been used. Some of these are individualized, such as sending personalized birthday greetings by email, and others are general, such as regular newsletters outlining study progress and research updates. Once analysis and publication begins, results will be posted on the website. Initially, a major recruitment incentive was provided: on completion of their study registration participants were automatically entered in a draw for international travel prizes provided by a travel company.
A major challenge has been maintaining contact via email for two reasons: the need for participants to advise the study each time they change email address, and the number of incorrect email addresses provided. Follow-up of ‘bounce-back’ emails has proved more time-consuming than originally planned, as it is usually undertaken on an individual basis and often requires several contacts via alternative email addresses and phone calls when email contact has been unsuccessful. Frequent changes of email addresses have been identified as a significant factor in the amount of work required for follow-up. Also, some of the unplanned consequences of changing email addresses require an understanding of the email environment. A recent email to 10,079 participants resulted in 1794 emails requiring follow-up due to (i) delivery problems because of full mailboxes and defunct or incorrect email addresses, (ii) personal messages notifying new information, (iii) automated replies such as out of office and (iv) spam messages which needed to be checked in case of mis-labelling as spam. Each email takes on average 10 minutes to resolve. For example, one institution changed its email server; this was noted when several emails bounced back from participants sharing the same provider, which was recognized as an institutional mail server. Rapid follow-up identified all relevant participants and the address could be changed for all those participants. Retention strategies such as e-birthday wishes and newsletters were also used as a point for checking email validity. Intensive follow-up has occurred during recruitment of the cohort and, if needed, will be undertaken after each annual survey to ensure retention in the study.
Discussion
In theory, there are a range of benefits to using Internet-mediated research. In particular, the administrative aspects of the data collection via the Internet are an attractive proposition. The cost of mailings, and especially data entry, are potentially minimized, particularly as the participants enter data directly. While substantial cost savings have been reported with Internet-mediated studies (Adams & White 2007, Smith et al. 2007), some authors state that the cost benefits are currently unknown (Ekman & Litton 2007). The instruments can also be customized to the participants in terms of appearance, sequence and delivery modes. The ability to administer instruments internationally from one site is also a key strength of this approach. However, as with postal questionnaires, there is little opportunity for guidance in answering questions, which increases the potential for invalid responses (Best & Krueger 2004). Web design can also be an issue; for instance, in the e-cohort questionnaire, progress indicators signalling the completion of a section were misinterpreted by some participants as the end of the survey and follow up of incomplete items has been a time-consuming process. The lower costs of administration are therefore offset by other costs, including website design, data organization and retrieval, and development of appropriate protection strategies to ensure secure access to the database.
There have also been barriers to using an Internet-mediated approach for a longitudinal study. The baseline survey, which will be the largest data collection instrument, has 108 questions. In addition to the direct time required to fill out the questions, participants have reported that downloading and uploading the survey using a dial-up connection is very time-consuming. As noted earlier, access to fast broadband connections is still limited in the domestic sphere and this will have an impact on the response rate. In the second survey, currently being developed, there are questions on Internet accessibility from home and work and type of connection used to provide data on the prevalent mode used by the study participants.
Another issue is that nurses and midwives are still in a transitional process of incorporating information technology into everyday practice, and the age of the workforce is of relevance. In 2006, the average age of nurses in NZ was 45 years (Clark 2006), in Australia 43 years (Australian Institute of Health and Welfare 2006) and in the UK 63% of Registered Nurses and Midwives are over the age of 40 years (Nursing and Midwifery Council 2006). Many of these practitioners will have completed their education before information technology became imbedded in everyday activities. A study of nurses’ work-related computer skills found 28% of the total group (454 nurses and nursing support staff) self-assessed as poor or fair proficiency; nurses 50 years or over or graduating before 1984 rated themselves as lower in proficiency (Wilbright et al. (2006). Other studies have shown that the nursing profession has been slower to adopt and use online resources than other health professional groups, and indeed the public (Estabrooks et al. 2003, Gosling et al. 2004,Janes et al. 2004). Internet-mediated research is likely to affect participation rates for nurses and midwives with limited computer skill proficiency.
Barriers to web-based participation raise the issue of whether it is possible to obtain a representative sample using a web-based design. Researchers have examined the characteristics of postal and face-to face interview responders compared to on-line responders. In the Millennium Cohort study of American defence force personnel (77,047 participants), 54·7% chose to answer online (Smith et al. 2007). Most demographic and health status metrics were comparable but there were also some differences, such as the on-line responders’ being more likely to share health-related characteristics, for example being overweight and having higher rates of smoking. Another smaller study comparing health behaviours of web-based responders compared to face-to-face interviews and found some difference in diet, mean body mass index (BMI), patterns of alcohol consumption, smoking and diet (Adams & White 2007). Until the Internet is fully incorporated as a fundamental communication medium throughout every group in society, the ability to gain representative samples will be limited.
A summary of the advantages and disadvantages of implementing web-based modes of data collection identified during the establishment of the Nurses and Midwives e-cohort study is presented in Table 1.
| Advantages | Disadvantages |
|---|---|
| Data accessibility – potential for earlier analysis and publication of resultsLess costly to maintainPotential to maintain participation as the survey can be undertaken anywhere in the world via InternetInnovative engagement with participants possible e.g. via regular newsletters and birthday greetingsAbility to analyse data on sub-groupsPotential for international research collaborations that are cost efficient | Expensive upfront costs for web design and buildingExtensive promotion required to profile the studyVariation in computer confidence and skillVariation in computer facilities such as broadband versa dial upEmail ‘bounce-backs’ requiring constant follow-upWebsite maintenanceSample bias in favour of competent Internet usersLow response rates |
Conclusion
While in this paper we have discussed some of the issues associated with Internet mediated research, it is also important to acknowledge the opportunities created by new and rapidly evolving communication technologies. Disadvantages such as limited access to fast Internet connections and personal aversion to computer modes of data collection are transitional issues at a time of rapid technological development. Establishing the Nurses and Midwives e-cohort Study has been a challenging experience at this point in time. As with all designs, whether web-based or using any other mode, there are inherent advantages and disadvantages. However, as familiarity with ICT increases, particularly amongst younger healthcare professionals, and increasingly sophisticated technology becomes commonplace, research via the Internet will clearly increase. This is certainly an area where nursing research can use the developing technologies to increase international research links and explore global issues related to the workforce. The development of on-line methods and tools is a major and exciting development in the world of research.
Acknowledgements
Industry partners providing in kind support for the project include: Queensland Nursing Council, Nurses and Midwives Board of New South Wales, Nurses Board of Tasmania, Nurses Board of Western Australia, Nurses Board of the Australian Capital Territory and Nursing Council of New Zealand. Corporate sponsors include Virgin Blue, Virgin Atlantic and MessageNet.
Funding
This project is supported by grants from the Australian Research Council (LP0562102), Australian National Health and Medical Research Council (2005002108) and New Zealand Health Research Council (456163). Industry partners providing additional funding include: Queensland Health, the South Australian Department of Health, Injury Prevention and Control Australia (Pty Ltd), Nursing Council of New Zealand and the Macquarie Bank Foundation.
Conflict of interest
No conflict of interest has been declared by the authors.
Author contributions
CT was responsible for the study conception and design. AH was responsible for the drafting of the manuscript. JG, PS, EY, AT & CT made critical revisions to the paper for important intellectual content. AH & CT obtained funding.
