IJE Advance Access published online on January 17, 2008
International Journal of Epidemiology, doi:10.1093/ije/dym294
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Cohort Profile: The Nurses and Midwives e-Cohort Study—A Novel Electronic Longitudinal Study
1School of Nursing and Midwifery, The University of Queensland, Brisbane, Australia.
2School of Population Health, The University of Queensland, Brisbane, Australia.
3Faculty of Health and Environmental Sciences, AUT University, Auckland, New Zealand.
4Accident Research Centre, Monash University, Victoria, Australia.
5School of Health Sciences, Massey University, Wellington, New Zealand.
* Corresponding author. School of Nursing & Midwifery, Faculty of Health Sciences, The University of Queensland, Room 228 Edith Cavell Building, Royal Brisbane and Women's Hospital, Herston, QLD4029, Australia. E-mail: catherine.turner{at}uq.edu.au
Accepted 19 December 2007
 |
How did the study come about? |
|---|
Top
How did the study...
What does the study...Nurses and midwives comprise the largest professional group in most national health systems, so shortfalls in numbers can have a substantial impact on health care delivery. A scarcity of human resources in health has been internationally recognized and has led the International Council of Nurses to launch the Global Workforce Project in 2004, and the World Health Organization to announce the Health Workforce Decade 2006–15 in 2006.1,2 Efforts to address workforce needs through coherent workforce planning and policy setting are hampered by the complexity of predicting the supply of and demand for nurses and midwives, and the challenges associated with understanding drivers of workforce retention.3 Available workforce descriptors among regulatory authorities vary considerably; and collections are mostly cross-sectional, frequently incomplete and typically limited to administrative databases. In Australia and New Zealand, workforce issues include the migration of staff between states and countries, and critical personnel shortages in rural and remote communities; although data limitations preclude clear delineation.3,4
One response to the problem of insufficient available information on which to base workforce planning, the Nurses and Midwives e-cohort study, was conceived by its director and developed by a team of researchers at The University of Queensland, Brisbane, Australia. In the first instance the study focused on an investigation of the Queensland workforce. Pilot data and other findings,5–7 suggested a desirable initial aim was to conduct a longitudinal examination of recruitment and retention in the Queensland nursing and midwifery workforce to better understand time-related patterns affecting retention and loss. Encouraged by the planned move of professional registrations to an on-line format, we attempted to maximize efficiency by adapting traditional longitudinal research methods to a novel electronic cohort (e-cohort) design. We envisioned that participants would primarily engage with and respond to the study using email and the Internet; a strategy which has not yet entered mainstream epidemiology.8 From the project's inception, professional associations and industry partners committed funds and other support and helped establish credibility and relevance to stakeholders and potential participants.
As the Queensland component was being developed, the value of expanding the cohort to increase sample size and to cover a variety of settings was apparent so academic colleagues elsewhere in Australia and in New Zealand were invited to join the research team. All nursing and midwifery councils in Australia and the Nursing Council of New Zealand were then approached to seek their collaboration to enable this expansion.
|
What does the study cover? |
|---|
The broad aims of this study are to establish a longitudinal e-cohort to examine factors associated with recruitment and retention of the nursing and midwifery workforce in order to develop retention strategies; and to assess the current and future utility of information technology to transform traditional longitudinal epidemiological study methods of recruitment, data capture and follow-up. We label the themes of the first aim as (i) Work/Life Balance that describes and quantifies the factors associated with retention of the existing nursing and midwifery workforce, new graduates’ transition to practice, patterns of employment and early retention in the workforce, and recruitment and retention of students; and (ii) Staying Healthy which measures the prevalence, incidence and associated risk factors of musculoskeletal disorders and work-based injury outcomes within the cohort. The potential for studying other outcomes, especially psychological health and changes in cohort patterns of risk factors (such as cigarette smoking, body weight, limited exercise and poor diet), will be realized as the study extends further into the future.
|
Who is in the sample? |
|---|
There are approximately 290 000 registered and enrolled nurses and midwives in Australia9 and 44 400 nurses in New Zealand10 eligible for inclusion in the study. Anticipating the planned move of professional registrations to an online format, we envisaged recruiting participants using mixed (but primarily electronic) modes, similar to those described and successfully adopted elsewhere11,12 for cross-sectional surveys. We planned to send personalized emails to all eligible nurses and midwives introducing the study and inviting participation, which could be initiated by clicking a URL link within the email that directly transferred them to the study website. Strategies to reduce non-response were to include sending personalized reminder emails to those who did not register and, where possible, mailed and telephoned reminders. However, due to issues of completeness and privacy, no authority provided an electronic list of its registered members. Instead restrictions imposed by the involved councils varied, and pragmatic non-electronic recruitment strategy compromises were required. Three nursing councils offered ‘in-kind’ support by including a hard copy invitation to participate with annual practicing certificate renewal documents; three others used study funds to mail postcards to all practicing nurses and midwives on their databases; and three nursing councils and the Midwifery Council of New Zealand refused our request to contact their members. Regular mass newsletters and publications circulated to registrants were also targeted with advertisements and articles featuring the study and urging participation. Targeted or personalized reminder protocols for non-responders were not sanctioned by any council. The staggered negotiations with Australian nursing councils and the year-long rolling certificate renewal practice of the New Zealand Nursing Council spread the mailing of invitations over a calendar year, from April 1, 2006 to March 31, 2007, with a cut-off for recruitment planned for March 31, 2008.
Undergraduate nursing students were also recruited from all universities offering nursing programs in Queensland (seven universities) and South Australia (three universities), covering approximately 6500 students.13 Australian education providers issue all students with their own email address but none was willing to provide them to us within a database. Instead, Nursing Program Directors sent bulk email messages to their students after the first semester census date (March 31st) in 2006 with an attachment containing an invitation to the study, and also spoke to those students who attended large year group lectures about the study at this time. When permitted, project staff also attended large year group lectures of students within the first few weeks of the semester to promote their recruitment into the study and handed out bookmarks advertising the study and website address. Directors were then contacted again and requested to email a global reminder message and the invitation attachment in the following semester.
As of October 9, 2007, the sample consists of 8247 registered nurses and midwives including 540 students undertaking undergraduate nursing studies within Queensland and South Australia of whom 135 were also registered with their respective councils as enrolled nurses. The current response rates appear in Table 1; all are very low, uniformly below 5%, with no meaningful differences among invitation strategies, and are consistent with response rates for a recent Internet-based birth cohort study using similar non-traditional recruitment methods.14
|
|
How often have they been followed-up? |
|---|
Follow-up measurement waves will be conducted annually commencing April 2008, with funding secured to survey participants until 2010, and further funding applications to support data collection beyond this point pending.
|
What has been measured? |
|---|
All participant data are being collected electronically using Internet-based survey software and stored in secure structured query language (SQL) databases customized for the study. Potential participants must access the study website to register at: http://www.e-cohort.net. After providing informed consent on-line, participants establish their personal profile by entering a username and password and then proceed to the registration page which requests baseline demographic, contact and work-related details. At all future logins each participant's personalized profile page is displayed (Figure 1).
|
Participant details are stored in a registration database with survey responses stored separately. The databases are linked by a study ID, generated automatically upon registration and accessible only to a core group of investigators. Data are entered on a question-by-question basis so that if a participant suspends the survey or loses their Internet connection entered data are not lost; at next login participants are automatically re-directed to the last question they completed.
The registration database includes age, gender, marital status, income, contacts (including a second email address, an alternative contact person's email and their telephone numbers), employment details and reasons for leaving work; and all can be updated at any time (with previous entries archived automatically). Table 2 includes basic demographic data of registered participants, which show them to be quite similar to the Australian15 and New Zealand16 workforces overall, low uptake notwithstanding. For Australian nurses and midwives, the national average age is 43.3 years compared to 43.5 years for those registered in the cohort; 91% are female compared with 94% in the cohort; and 0.5% were Aboriginal and/or Torres Strait Islander compared to 0.6% registered in the cohort. Also included in Table 2 are basic professional descriptors of registered participants. The cohort encapsulates a diverse group of nurses and midwives working in a variety of settings. While the majority are primarily based in hospitals, the cohort includes several hundred participants in each of a number of other workplace settings allowing key workforce issues to be investigated over time between various hospital and non-hospital settings.
|
New Zealand ethics clearance was conditional on making all questions voluntary, apart from participant's electronic addresses, resulting in many missing observations for core demographic and professional variables (Table 2). Participants have been urged in quarterly electronic newsletters to complete these registration fields, with limited success. A core set of these variables will be re-elicited in each future follow-up wave.
Once registered, participants progress to their personal profile page, which includes a link to the baseline survey, which takes about 40 min to complete. The Work/Life Balance theme is examined via capturing a range of factors describing work practices and stresses, family responsibilities and study, and then relating them to changes in employment status. Workplace attributes elicited include work setting of main employment specialty, employment sector, usual work group or unit size and the nature and duration of shift work; and job stresses are captured by the Job Content Questionnaire (JCQ),17 the Effort-Reward Imbalance subscale18 and the Hostility Score.19 The Staying Healthy theme is addressed by recording a range of physical and mental health attributes at baseline, using a number of validated instruments, and linking these to prevalent and incident indicators of musculoskeletal disorders and injury outcomes. Measurement instruments include the International Physical Activity Questionnaire (IPAQ),20 Centre for Epidemiologic Studies Depression (CES-D) score (Version 10),21 SF-36 (Version 2)22 and the Diet Quality Score.23 Data assessing musculoskeletal disorders are collected using a modified Nordic Musculoskeletal score24 and the Pain Catastrophizing Scale (PCS).25 Follow-up measurement waves will utilize the same instruments in staggered cycles to diminish participant burden. Additional questions will be included as required to address emerging hypotheses, as well as capturing a wide range of morbidity.
|
What is attrition like? |
|---|
The first follow-up survey has not yet been administered, so attrition from the cohort remains unquantified. The overarching strategy for minimizing attrition is to create a personalized survey experience for participants. To this end, all have password-protected access to a personalized home page that they use to view and modify their personal details, and to view completion status of past and current surveys. Email contact with participants is also personalized; they receive a welcome email 2 weeks after registration, emailed birthday cards, thank you emails for completing a survey and quarterly emailed newsletters. Reminders to complete surveys are emailed as necessary. Two-thirds of participants who register, attempt and complete the survey immediately, increasing to 80% following reminder emails. The reminder system also helps maintain current contact details for participants, with incorrect email addresses corrected using the sequence of additional contacts noted earlier. Furthermore, the nurses and midwives are encouraged to update their personal contact details as needed by logging in to their personalized e-portal.
|
What are the main strengths and weakness? |
|---|
The web-based recruitment and data collection techniques used by the Nurses and Midwives e-cohort study are novel in large-scale epidemiological cohort research. We are not aware of other studies that have attempted to systematically sample, recruit and maintain a large cohort from a defined population using a traditional sampling frame entirely via electronic methods; although existing studies, such as the Nurses Health Study II, and new studies, such as the Millennium Cohort Study,14 now offer participants the opportunity to complete surveys electronically. It is recognized that recruitment of Internet-based cohorts is becoming more feasible,14,26 as is the feasibility of using web-based questionnaires in large population-based epidemiological studies.26,27
As traditional mechanisms of data elicitation are becoming less attractive for robust epidemiological studies, with declining landline telephone ownership and response to mailed questionnaires,28,29 our approach is to embrace and assess this evolving technology. Increasingly, many organizations and government agencies provide options for direct electronic interaction and data transfer; for example, within Australia, the national Census and annual personal tax return is now offered electronically. It seems natural that epidemiological researchers should harness these technologies once there is mainstream acceptability of these electronic interactions.
Electronic technologies have the potential to significantly improve survey methods with increased cost-effectiveness in labour, printing, distribution and storage of records, increased survey accessibility for participants and earlier access to data for researchers.12,29,30 Other anticipated advantages are improvements in data quality and, integrity, and reduced data management needs, although these will depend on the design of the electronic interface and construction of the survey.29 For example, unlimited free-form text fields may induce many variant responses engendering considerable data cleaning. Automated branching, skips, range and consistency check methods programmed within the survey software can control the range of responses to questions thereby minimizing such problems. Electronic methods will also enable the timing and targeting of follow-up surveys administrations to vary.
The most profound limitation of this study was the consistently poor response rate. Past web-driven population surveys have also generally yielded low to moderate response rates,26,31,32 although the effect of the resulting coverage biases is argued.33 A notable exception, with response rate of 82%,12 arose from a survey of tertiary students who received personalized email information and invitations to participate with embedded hyperlinks to the survey; all received incentives to participate, and non-responders were sent email reminders. Our intended recruitment strategy aimed to parallel this successful protocol. Unfortunately, electronic database unavailability and data protection concerns necessitated adoption of non-electronic recruitment procedures, with the consequent extra barriers to recruitment likely to be largely responsible for the response rate. We are currently seeking permission to survey nurses electronically (as per our original protocol) within another jurisdiction and are planning a graduate follow-up study to assess this directly.
The low response rate is also likely to be at least partially due to access and acceptability of the electronic medium within this population. While electronic technology is now a fundamental requirement for professional practice, with most Australasian nurses and midwives allocated email addresses by their employers, obstacles within professional environments remain. Access to email is not always available, work-based firewalls can inhibit full Internet functions, and time pressures mean that research participation receives low priority—if indeed permitted at all within the workplace. Furthermore, it is likely that some nurses and midwives remain uncomfortable with this technology. Outside of work, there are issues of incomplete coverage with not all members of the target population having access to the Internet. In 2005–06, 70% of Australian households had a computer and 60% had home Internet access.34 Internet penetration is likely to be higher in the study sample, as rates increase rapidly with higher incomes.34 Notwithstanding, predominantly female nurses and midwives have many competing demands and if confronted with barriers, such as slow dial-up connections or household/family obligations, are unlikely to participate in potentially lengthy research surveys.
Another major limitation within the study was the ethics requirement that no questions within the registration page were mandatory (except a valid email address) and the option to submit this mandatory field was prior to voluntary fields. Consequently many participants partially completed the important socio-demographics and professional questions that will be vital in describing the sample, act as confounders and be used to provide work-force data. While a core set of these variables will be moved to the survey and re-elicited in each future follow-up wave, it is difficult to make accurate assertions about the sample representativeness or composition within the interim.
With such a low uptake of participants within the study, our a priori position is that prevalence data and cross-sectional workforce descriptors may be unrepresentative. However, if the socio-demographic comparators prove reasonably similar to the national data then resultant prevalence estimates and descriptors still may have utility, especially in the current vacuum of systematically collected empirical information on nurses and midwives. Notwithstanding this, in absolute terms, the cohort is large and, providing participant retention is good, the acquisition and analysis of these prospective data is likely to yield robust and generalizable findings. It is these findings, which will provide real insight into understanding the drivers of nurses and midwives workforce retention and work-related factors associated with their health.
With advancing electronic technologies and increasing broad population acceptance, we are poised to harness the Internet for epidemiological research. Target population coverage and barriers are rapidly diminishing—but gatekeeper and ethical issues remain challenging and require resolution. While this large-scale international study faltered in achieving a respectable response rate, we expect that learning from our collective experiences and resolving these challenges will pave the way for future success.
|
Can I get hold of the data? Where can I find more information? |
|---|
The Nurses and Midwives e-cohort study team welcomes collaborative research opportunities. Access to the data is subject to a number of study protocols and approval processes. Contact information can be obtained from the study website, www.e-cohort.net.
|
Acknowledgements |
|---|
This project is supported by grants from the Australian Research Council (LP0562102), (SR0566924), 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. 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 the Nursing Council of New Zealand. Corporate sponsors include Virgin Blue, Virgin Atlantic and MessageNet.
Conflict of interest: None declared.
|
Notes |
|---|
The Nurses & Midwives e-cohort Group includes: Anthony Tuckett, Desley Hegney, David Thompson and Simon Stewart. 
|
References |
|---|
1 International Council of Nurses. The Global Nursing Shortage: Priority Areas for Intervention (2006) Geneva: International Council of Nurses.
2 World Health Organization. The World Health Report 2006: Working Together for Health (2006) Geneva: World Health Organization.
3 Canadian Nurses Association. Planning for the Future: Nursing Human Resource Projections (2002) Ottawa: Canadian Nurses Association.
4 Gaynor L, Gallasch T, Yorkston E, Stewart S, Turner C. Where do all the undergraduate and new graduate nurses go and why? A search for empirical research evidence. Aust J Adv Nurs (2006) 24:26–32.[Web of Science][Medline]
5 Hegney D, McCarthy A, Rogers-Clark C, Gorman D. Why nurses are resigning from rural and remote Queensland health facilities. Collegian (2002) 9:33–39.[Medline]
6 Hegney D, McCarthy A, Rogers-Clark C, Gorman D. Why nurses are attracted to rural and remote practice. Aust J Rural Health (2002) 10:178–86.[CrossRef][Medline]
7 Hegney D, McCarthy A, Rogers-Clark C, Gorman D. Factors affecting the retention of the rural and remote area nursing workforce in Queensland, Australia. J Nurs Adm (2002) 32:128–35.[CrossRef][Web of Science][Medline]
8 Hartge P. Participation in population studies. Epidemiology (2006) 17:252–54.[CrossRef][Web of Science][Medline]
9 The College of Nursing. Statistical facts on nursing in Australia. Accessed August 16, 2007. Available at: http://www.nursing.aust.edu.au/StatFacts/.
10 Clark M, Ayling B. Workforce statistics 2006 update. In: Forum 2006. Accessed November 19, 2006. Available at: http://www.nursingcouncil. org.nz/forum.html.
11 McCabe SE, Diez A, Boyd CJ, Nelson TF, Weitzman ER. Comparing web and mail responses in a mixed mode survey in college alcohol use research. Addict Behav (2006) 31:1619–27.[CrossRef][Web of Science][Medline]
12 Kypri K, Gallagher SJ, Cashell-Smith ML. An internet-based survey method for college student drinking research. Drug Alcohol Depend (2004) 76:45–53.[CrossRef][Web of Science][Medline]
13 Department of Education Science and Training. Nursing Education and Graduates: Profiles for 1999, and 2000, with projections for 2001. (2007) Accessed August 16, 2001. Available at: http://www.dest.gov.au/archive/highered/eippubs/eip01_13/chapter3_4.htm.
14 Richiardi L, Baussano I, Vizzini L, Douwes J, Pearce N, Merletti F. Feasibility of recruiting a birth cohort through the Internet: the experience of the NINFEA cohort. Eur J Epidemiol (2007); doi: 10.1007/s10654-10007-19194-10652.
15 Australian Institute of Health and Welfare. Nursing and Midwifery Labour Force 2004. (2006) Canberra: AIHW. National Health Labour Force Series No. 37. Cat No. HWL 38.
16 New Zealand Health Information Service. New Zealand Health Workforce Statistics 2004: Nurses and Midwives. (2004) November 22. Accessed September 11, 2007. Available at: http://www.nzhis.govt.nz/stats/nursestats.html.
17 Karasek R, Brisson C, Kawakami N, Houtman I, Bongers P, Amick B. The Job Content Questionnaire (JCQ): an instrument for internationally comparative assessments of psychosocial job characteristics. J Occup Health Psychol (1998) 3:322–55.[CrossRef][Medline]
18 Siegrist J. Adverse health effects of high-effort/low-reward conditions. J Occup Health Psychol (1996) 1:27–41.[CrossRef][Medline]
19 Koskenvuo M, Kaprio J, Rose RJ, et al. Hostility as a risk factor for mortality and ischemic heart disease in men. Psychosom Med (1988) 50:330–40.
20 Craig CL, Marshall AL, Sjöström M, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports and Exerc (2003) 35:1381–95.[CrossRef]
21 Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas (1977) 1:385–401.[CrossRef]
22 Ware JE, Kosinski M, Dewey JE. How to Score Version 2 of the SF-36 Health Survey (2000) Lincoln: RI Quality Metric Inc.
23 Collins C. Diet Quality Score (2004) Newcastle: University of Newcastle.
24 Kuorinka I, Jonsson B, Kilbom A, et al. Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms. Appl Ergonom (1987) 18:233–37.[CrossRef][Web of Science][Medline]
25 Sullivan MJL, Bishop SR, Pivik J. The Pain Catastrophizing Scale: development and validation. Psychol Assess (1995) 7:524–32.[CrossRef][Web of Science]
26 West R, Gilsenan A, Coste F, et al. The ATTEMPT cohort: a multi-national longitudinal study of predictors, patterns and consequences of smoking cessation; introduction and evaluation of internet recruitment and data collection methods. Addiction (2006) 101:1352–61.[CrossRef][Web of Science][Medline]
27 Ekman A, Dickman PW, Klint A, Weiderpass E, Litton JE. Feasibility of using web-based questionnaires in large population-based epidemiological studies. Eur J Epidemiol (2006) 21:103–11.[CrossRef][Web of Science][Medline]
28 Galea S, Tracy M. Participation rates in epidemiologic studies. Ann Epidemiol (2007) 17:643–53.[CrossRef][Medline]
29 Ekman A, Litton JE. New times, new needs; e-epidemiology. Eur J Epidemiol (2007) 22:285–92.[CrossRef][Web of Science][Medline]
30 Dillman DA. Mail and Internet Surveys: The Tailored Design Method (2000) New York: Wiley.
31 Bälter KA, Bälter O, Fondell E, Lagerros YT. Web-based and mailed questionnaires: a comparison of response rates and compliance. Epidemiology (2005) 16:577–79.[CrossRef][Web of Science][Medline]
32 Link MW, Mokdad AH. Alternative modes for health surveillance surveys: an experiment with web, mail, and telephone. Epidemiology (2005) 16:701–4.[CrossRef][Web of Science][Medline]
33 Nohr EA, Frydenberg M, Henriksen TB, Olsen J. Does low participation in cohort studies induce bias? Epidemiology (2006) 17:413–18.[CrossRef][Web of Science][Medline]
34 Australian Bureau of Statistics. Household Use of Information Technology, Australia, 2005-06 (2006) Canberra: Australian Bureau of Statistics. 8146.0.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||