556372 research-article2014
JHI0010.1177/1460458214556372Health Informatics JournalCraswell et al.
Original Article
Does use of computer technology for perinatal data collection influence data quality?
Health Informatics Journal 1–11 © The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1460458214556372 jhi.sagepub.com
Alison Craswell and Lorna Moxham University of Wollongong, Australia
Marc Broadbent
University of the Sunshine Coast, Australia
Abstract Population health data, collected worldwide in an effort to monitor mortality and morbidity of mothers and babies, namely, perinatal data, are mandated at a federal level within Australia. The data are used to monitor patterns in midwifery, obstetric and neonatal practice, health outcomes, used for research purposes, funding allocation and education. Accuracy in perinatal data is most often reported via quantitative validation studies of perinatal data collections both internationally and in Australia. These studies report varying levels of accuracy and suggest researchers need to be more aware of the quality of data they use. This article presents findings regarding issues of concern identified by midwives relating to their perceptions of how technology affects the accuracy of perinatal data records. Perinatal data records are perceived to be more complete when completed electronically. However, issues regarding system functionality, the inconsistent use of terminology, lack of data standards and the absence of clear, written records contribute to midwives’ perceptions of the negative influence of technology on the quality of perinatal data.
Keywords attitude to computers, Australia, (MeSH) data collection, data quality, midwifery, perinatal nursing
Introduction The introduction of Information, Communication and Technology (ICT) into midwifery, like other areas in healthcare, is proposed to combat rising costs, decreasing availability of skilled staff as well as provide improved efficiencies of healthcare delivery and increase the portability and comparability of health-related data.1–7 However, evidence that the use of technologies in healthcare has achieved these outcomes is yet to be seen within the literature.8–11 This has not stemmed the Corresponding author: Alison Craswell, School of Nursing and Midwifery, Faculty of Sciences, Medicine & Health, University of Wollongong, P.O. Box 1128, Noosaville BC, QLD 4566, Australia. Email: [email protected]
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
2
Health Informatics Journal
flow of technology into healthcare. Despite the increased use of technology in healthcare practice, confidence and competence of ICT use by nurses and midwives varies according to age, computer use at home, education and training.12–15 Research concerning the use of technology within nursing and midwifery practice finds midwives have favourable attitudes towards the use of technology.16,17 However, results from pre- and post-implementation studies16,18–20 show that these positive attitudes lessen after implementation of computer technology.16 One such technology recently introduced into midwifery practice in Queensland, Australia, relates to perinatal data collection. Perinatal data are collected for mothers and babies Australia wide as mandated at a commonwealth government level. The data are used to monitor patterns in midwifery, obstetric and neonatal practice as well as for the planning of health services, research and the education of midwives.21 The research presented in this article took place in Queensland, a north-eastern state of Australia, with a population of approximately 4.75 million people, about 20.1 per cent of the current population of Australia.22 However, the state covers a large space geographically as can be seen pictorially in Figure 1, with the population scattered up the eastern coast, requiring many and varied health services to support residing childbearing women. The collection of perinatal data in Queensland have recently moved to an eHealth platform, superseding the traditional paper form, which is reported to have improved data quality, enhanced accessibility and communication, produced cost savings and improved the timeliness of the availability of collected data items.23 The introduction of technology for perinatal data collection adds to the increasing use of technology within healthcare and the daily work of midwives. Data presented in this paper were gathered as part of a larger study the authors were undertaking which examines the influences on midwives during the process of entering perinatal data into the computer. Only data from the analysis that lies within the theme of perceived data accuracy are presented and discussed here.
Method Grounded theory (GT), a qualitative methodology that has an inductive orientation, was used in this study to add a depth of discovery that may otherwise not have occurred with a quantitative approach. The voice of the midwife was captured and is considered important in determining what happened as it provides an experiential point of view.25 Purposive sampling was used to interview participants in line with GT methodology. This approach ensured that data were captured from participants with knowledge and experience of entering perinatal data into a computer for collection. Later, theoretical sampling, an approach that adds and refines properties and dimensions to acquire an in-depth understanding of analytical categories,26 was used to gather data from participants who would fill the gaps in the developing theory. By the end of data collection, 14 midwives and 1 health information manager from 12 different hospitals across Queensland, Australia, using three different systems for entering perinatal data participated in this research. The participants held a variety of positions within their organisations and ranged in experience from Level 1 midwives to clinical nurse consultants and educators. Some participants also held the position of perinatal data coordinator for their unit, which had the added role of validating and extracting perinatal data regularly at the end of an allocated period to be sent to Queensland Health, the governing body for healthcare at state level. Adhering to GT methods, the sample size was not pre-determined but influenced by saturation of the data rather than a specific required number of participants to meet generalisable sampling requirements.27 Saturation of the data occurred when no new information emerged from interviews and the theoretical framework was sufficiently populated to explain the phenomena under study.28 Participants were asked an initial open-ended question which was designed to encourage a full meaningful answer using the midwife’s own knowledge and experiences around the research topic.
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
3
Craswell et al.
Figure 1. Comparative size of the State of Queensland, Australia, to Texas, USA.24
Further discussion during each interview was unstructured, led by the participant but focused around the topic, adhering to GT methodology with the interviewer focusing the discussion back to the topic of perinatal data entry into the computer when necessary. Interviews lasted between 30 min to 1.5 h in duration, were recorded and later transcribed by the primary researcher. Ethical approval for the research was obtained from the University of Wollongong, Human Research Ethics Committee in 2012 with the research design adhering to the principals of justice, respect and beneficence.29 Data were analysed using the constant comparative method, again maintaining consistency with GT methods. Data analysis occurred concurrently with data collection over a period of approximately 10 months. Initial open coding using ‘in vivo’ codes and the writing of early memos occurred with early analysis giving some direction for theoretical sampling requirements for further participant recruitment. Development of codes, then themes and finally allocation of substantive or theoretical codes occurred in line with the systematised process of GT methodology and conceptualising what is happening in the data. Throughout this process, NVivo data management software was used to assist with organisation of the data.
Findings The findings presented here deliberately use the participants’ own words with ensuing discussion on the emerging themes, categories and associated factors. The conventions used throughout this article identify direct quotes using ‘I’ for interviewer and ‘P’ for participant. A major theme that emerged from this research was ‘perceived data accuracy’ which highlighted participant concerns over the accuracy of perinatal data entries since perinatal data entry has moved from paper to computer. This theme is made up of the sub-themes: (a) accuracy in records, both completed, computer records and written records and (b) data standards, specifically obstetric data standards, system functionality and inconsistent terminology. The relationships between these can be seen in Figure 2. Participants themselves questioned: Data are entered into a computer for perinatal data collection but is the data entered accurate?
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
4
Health Informatics Journal
Figure 2. Theme – perceived data accuracy and contributing sub-themes and associated factors.
Perceived data accuracy Perceptions of data accuracy by participants varied across interviews. Participants believed the data they entered was accurate but communicated that they knew it could be more ‘complete’. Perception regarding the accuracy of the data was communicated as in this example of records being returned with errors for correction: I: ‘So in your mind, do you think the data that’s in there is fairly accurate?’ P: ‘I wouldn’t make that assumption’. I: ‘So why do you think that?’ P: ‘Because sometimes what comes back from the perinatal people that they want confirming, it just makes me wonder if that (data) hasn’t been input correctly, I wonder if everything else has?’ This theme, ‘Perceived data accuracy’, is contributed to by four sub-themes: accuracy in perinatal data records, data standards, system functionality and inconsistent terminology. The findings in relation to these contributing sub-themes and associated factors will now be presented.
Accuracy of perinatal data records Computer records. It was generally felt by participants that moving to an eHealth platform for submission of perinatal data had improved the accuracy of the data entered. They found that the process of validation of an individual record before it could be finalised and submitted, forced them to enter data into fields that were easily left blank on the old paper record. Field validation is a construct of particular software that means data entered must fit set parameters for the page to move on or be saved. Logical checks on the data entered into the database are made against predefined rules for value ranges of the data to ensure they fit these parameters.30 Participants with the added role of perinatal data coordinator for their particular unit perceived that the number of returning errors from the perinatal data collection unit, the department in Queensland Health governing perinatal data, to be less than when the collection was done by paper
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
5
Craswell et al.
form. This was confirmed during discussions with staff at the data collections unit in Brisbane (C. Morris, 9 February 2011, personal communication) and also when examining hospital error reports and graphs published in Perinatal Data newsletters.23,31 One participant reported a 2- to 3-month lag time while using paper perinatal data forms but communicated this had subsequently been reduced since using a computerised extract of data. This was thought to be due to the validation function described above. However, some participants also reported uncertainty regarding inaccuracies in the completed and validated records. One participant reported: I can look in it (the record), but unless I sit there with the chart and check the (computer) entries I don’t know whether the data’s correct or not. No one does.
The time required to do such in-depth checking was reported as being unavailable to midwives and perinatal data coordinators. Another participant reported knowledge of inaccuracies in the data from their unit: We had no faith in the data from here because we knew it was very inaccurate even with the validation process.
These inaccuracies in computer data entry were generally not considered acceptable. Participants reported that a means of improving the quality of the information being sent to Queensland Health was designation of a perinatal data coordinator to check, correct and complete the electronic forms that midwives have entered data into at some point during care. With this model in place, some participants report being able to ensure the data are accurate: I: P:
‘So you think once you’ve done your clean, you’re fairly confident the data’s accurate? Yep, complete and accurate’.
This participant uses the term ‘clean’ in the context of correcting obvious errors in the data. Other participants still believed accuracy was inadequate and completeness could be improved: What I’m sending off is accurate but it’s not as robust as it could be.
The value of accurate perinatal data can also be seen by some maternity units giving perinatal data coordinator midwives between 4 and 8 h a week off-line time to check, complete and validate records prior to extracting the data to the data collections unit. The perceived level of accuracy reported from participants of this research varies from good to unknown. Accuracy does depend on a number of influencing factors including ‘busy-ness’ in the unit. Participants reported that when they were busy or pressed for time, they enter less data into the perinatal data record: I try not to skip over things, but I’m sure if I went down through all the, you know, adding things in, I’m sure we could pick up a lot more. Because I’m in a rush and someone else is just about to deliver and I need to go in there as well.
These busier times result in a requirement for additional staff, usually more casual or agency midwives, who may not be familiar with the system for entering perinatal data but are in some units still required to enter their own perinatal data. Midwives reported this as resulting in less-complete and less-accurate records.
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
6
Health Informatics Journal
Finally, concerns were communicated about the quality of the data that are printed off on discharge and go with the mother to the General Practitioner (GP) as this occurs before the coordinator has done a clean and check: You know at the end of the day you are supposed to print off a perinatal data once it’s all complete so you can send it to the GP and some of the things you find there, you say well you know the information these ladies are getting, that they’re reading is wrong and what’s going back to their GP is wrong.
Written records. When entering perinatal data into the computer, participants report utilising a combination of memory and the written chart. It was reported that the easiest and quickest entry of perinatal data occurred immediately after a woman had birthed when all the data were fresh in the midwife’s mind. However, this was not always possible due to the time constraints of the midwives’ role or a birth occurring on the change of shift, leaving no time for data entry. In these cases, perinatal data entry was handed over to another midwife or entered some time later. This could be on another shift or another day, entered by the birthing midwife or someone else. The resulting accuracy of such records was thought to be inadequate: Certainly if the accoucher’s (birthing midwife) not doing it, and you’re handing it on, then it can just be completely NOT accurate.
Participants suggested that the best practice for this occurrence involved all data to be entered electronically being written somewhere in the paper chart but also that this was not always the case or the process utilised for the systems in use by participants. The worst case scenario reported was when the discharge midwife went to enter the small amount of required discharge information and then check and validate the record to find there was no record created and therefore no data entered for the mother or baby at all. In these cases, transcribing data from the written chart was undertaken and participants voiced concern over relying on the accuracy of the written record: I: ‘And do you think the paper records are accurate? P: Probably not. Often they’re not. I’ve done documentation audits and there’s either things missing or … I find the same thing going through the paper record to complete the perinatal data. You know I’ve found records where I can’t find documentation of the apgars (a score of 5 items each being 0, 1 or 2 awarded at 1, 5 and 10 minutes post birth by the birth attendant) anywhere in the mother’s or baby’s notes, or a birth weight or something. So one would assume other things were missing that you’re not necessarily looking for’. Another participant reported looking up information in sources other than the relevant paper records to ensure accuracy: I always go in, I do always check the lady’s blood group. I don’t just take it as a given, what’s written in the handheld record in case something has been transcribed incorrectly.
Participants were concerned about the completeness and accuracy of data in written patient records that they are sourcing for entry into the computer for perinatal data collection.
Data standards Obstetric data standards. Participants reported that there is no standard for obstetric data collected via private medical consultations. Therefore, the data sourced from the private medical record
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
7
Craswell et al.
could be missing altogether, inconsistent with the field definitions of perinatal data or misrepresented: P: So sometimes the electronic data isn’t recorded anywhere on paper, only electronically and vice versa. I: Does that include the obstetric record as well? P: Yeah, because they’re all different from each obstetrician too, so no standard, no. An example of the misrepresentation of data is the number of ultrasound scans (USSs) which may be recorded to include only the mandatory clinical scans at 12 and 18 weeks: Realistically a lot of that stuff should be in the chart and some of it is inaccurate. Like the scans for instance.
Participants reported that when they questioned the mother further about the number of USSs performed, it was reported that the obstetrician has used the scanner to determine the foetal position and foetal heart rate at every visit. The data thus are a misrepresentation of the actual number of times the mother and the developing foetus is exposed to ultrasound. This may also relate to obstetricians interpretation of the term USSs to include only those for morphology purposes, clinical USSs looking specifically for congenital abnormalities.
System functionality Participants reported that some midwives have limited understanding of the nuances of the various software systems in use in Queensland for entry of perinatal data. An example was that midwives did not know that particular fields with drop-down boxes had scroll bars providing multiple items, some of which were hidden, for selection. This would indicate midwives had a lack of computer skill or training with the software for perinatal data entry and that the data selected are not always reflective of the appropriate category. Either another category is chosen or the information is left out: I actually had to teach someone about the drop downs the other day. About BGL’s (Blood Glucose Levels) and BSL’s (Blood sugar levels). They didn’t know there was a drop down box (for pregnancy diabetes) and that was only self discovery.
When the user does not know to use the scroll bar to scroll down a box to select an appropriate option, data go unrecorded which directly affect the collection of statistics relating to women and/ or the neonate.
Inconsistent terminology Field definitions in a perinatal data collection system are defined as the question that relates to a field within perinatal data for which information is entered. Findings from this research suggest that multiple understandings of field definitions exist across jurisdictions. An example can be seen with entering an apgar score (a score of 5 items each being 0, 1 or 2 awarded at 1, 5 and 10 min post birth by the birth attendant): … for instance the apgar scoring in our version, the electronic version we have, you need to actually put in the score for each of the 5 factors and then it automatically adds it up. Rather than put in a total, like 9 or whatever. And I quite regularly find somebody’s put 1 for the respirations at both 1 minute and 5 minutes
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
8
Health Informatics Journal and yet they’ve written respirations established at birth (in the medical record). And again I’ve spoken to people about that and of course they always deny it’s them that does that – they understand how to do apgars!
Similarly, where one unit may define ‘midwife led care’ as birthing with a midwife who has met the woman at an antenatal visit prior, another defined it as requiring a minimum of four or five visits with that midwife in a ‘know your own midwife’ scheme. Inconsistencies were apparent across the various systems and across fields within each perinatal data system used. In some systems, the field definitions written into the software for data extraction to Queensland Health did not exactly match the field definition required by Queensland Health. Therefore, the data extracted for that field are consistently incorrect and returned for correction or clarification to the perinatal data coordinator. These system inconsistencies increase the workload of error correction and clarification as well as potentially collect mismatched data between healthcare institutions: I: In terms of the ability of staff to enter complete and accurate information, if we did away with the paper chart tomorrow, do you think it would be a success? P: No, As I say we’re not at that level of literacy yet and I think our databases would have to be much more user friendly to do that. And you would have to have some little help boxes as well to explain to people. Something that you could call up to verify what it is, the information that you’re putting down.
Discussion Obstetric and midwifery practice today is primarily evidence based utilising research from data derived from sources such as the perinatal data collection in an effort to improve health outcomes and planning for future health service delivery needs.32 The findings of this research show that generally participants are concerned about the accuracy of data in the perinatal data entries they complete using a computer. This was demonstrated by participants communicating a clear understanding of the validation process and knowledge that data entered could be successfully validated by the system, yet not match the written record. Participant concerns relating to data quality persist despite perceived improvements in completion of fields and error return rates that moving to an eHealth platform for perinatal data collection are reported to have brought.33 Other Australian research comparing electronic discharge summaries to written versions found that moving to computers does not always improve the data quality, supporting this assertion.34 Perceptions of the accuracy of entered data are influenced by processes such as the transcription from paper to computer when written records are inconsistent. A reliance on memory results in entering of different information than that which exists on the paper record. Unavailability or inconsistency of information from sources such as obstetric records also contributes to inaccuracy due to differing data standards. In addition to this, system functionality when perceived as poor by participants may result in less information being entered or incorrect selections of data made. Finally, while inconsistent terminology persists, the standardisation of the data entered cannot be certain. Inconsistent terminology between staff and across workplaces in relation to data definitions is not a problem specific to midwifery but seen across the healthcare sector.35 Evidence suggests a solution is the use of standardised data but this is reported as difficult in other system implementations in the healthcare sector.35 The research undertaken and presented in this article found a lack of standard data in written sources used for transcription and the use of different computer systems for perinatal data collection potentially reduced the accuracy of data even when the midwife was
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
9
Craswell et al.
committed to completing the record accurately and in a timely fashion. The necessity of a perinatal data coordinator, which is both a solution and a recognition of a problem, to correct errors and complete entries prior to validation and extraction of data to Queensland Health, arises as a result of these inconsistencies within data standards of field definition consistency, written records and the computer. However, the varied nature of this position, the irregular hours provided for error correction and differing levels of concern about the quality of data between coordinators result in inconsistencies between units in the accuracy of data submitted to the governing data collection unit. In addition, the concern that summaries are printed off for the mother and the GP on discharge prior to this check suggests this is not always successful in preventing dissemination of inaccurate information. Midwives entering the data need to input the correct data at point of care rather than the coordinators being left responsible for error corrections. However, while barriers such as issues with system functionality and inconsistent understanding of field definitions persist, this expectation seems unreasonable. The consequences of inconsistent and inaccurate perinatal data entries are potentially enormous with ironically, inaccurate data directly affecting the areas serviced by the information resulting from very data being collected. Assuming that the perinatal data collection is of high quality when potentially flawed data are known to be entered leads to the risk of these same data being used to make major decisions in the evaluation and future planning for maternity services. Such misinformation puts the health of mothers and babies at risk. Involving the users in all areas of the development and implementation of computer systems for perinatal data collection will not only increase their knowledge of system function and therefore accuracy of perinatal data entries but also enhance buy-in of managers and staff, motivating them to use the system in a more appropriate way. With accuracy in perinatal data being most often reported via quantitative validation studies both internationally and within Australia,36 research that investigates user issues underlying these studies provides valuable information on how practices can be changed to ensure perinatal data are of a high quality within Australia.
Limitations This research is not without limitation as this study utilised a small purposive sample and use of a methodology that prevents results being generalisable to the midwife population at large or to other computer systems for population data collection. Further research to test the findings with a large population using quantitative methods would strengthen these results.
Conclusion and recommendations Participants of this research communicated concern about the accuracy of the perinatal data they enter for women in their care and believe the data they enter into each field in response to each question are accurate. However, they reported uncertainty regarding whether the data entered online actually matched that written in the chart. Issues of inaccuracy within the perinatal data collection place at risk the planning of health services in Queensland across all jurisdictions that rely on accurate information and statistics and as a result, the potential health of women and their babies utilising these services. The move from paper to eHealth collection of perinatal data is believed to have resulted in a more complete data collection than was previously experienced using paper forms, with turnaround times from submission of data to availability of statistics thought to have improved. However, while a validation process can improve the completeness of data, it cannot ensure the data collected by midwives match those in the written record. A lack of data standards for written records, inconsistent field definitions across maternity units and computer systems, as
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
10
Health Informatics Journal
well as the persistence of inaccuracies despite records being validated and complete, offer areas for improvement to ensure the data quality of the perinatal data collection is paramount. Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
References 1. Deloitte. National E-Health and information principal committee: National E-Health strategy (ed Deloitte). Canberra, ACT, Australia: Deloitte Touche Tohmatsu, 2008, pp. 1–121. 2. Eley R, Fallon T, Soar J, et al. The status of training and education in information and computer technology of Australian nurses: a national survey. J Clin Nurs 2008; 17: 2758–2767. 3. Healy J, Sharman E and Lokuge B. Australia: health system review. Health Syst Transit 2006; 8: 1–158. 4. National eHealth Transition Authority (NEHTA). NEHTA Strategic Plan Refresh 2011/2012. 2013, http://www.nehta.gov.au/about-us/our-strategy 5. Smedley A. The importance of informatics competencies in nursing: an Australian perspective. Comput Inform Nurs 2005; 23: 106–110. 6. Vimarlund V and Koch S. Identifying where the values come from IT-Innovations in health and social care. Intell Inform Manag 2012; 4: 296–308. 7. Yu P and Comensoli N. An exploration of the barriers to the adoption of information technology in Australian aged care industry. In: Walduck K, Cesnik B and Chu S (eds) HIC 2004: proceedings. Brunswick East, VIC, Australia: Health Informatics Society of Australia, 2004, pp.19–223. 8. BlackAD,CarJ,PagliariC,etal.TheimpactofeHealthonthequalityandsafetyofhealthcare:asystematicoverview. PLoS Med 2011; 8: e1000387 http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal. pmed.1000387 (accessed 8 February 2011). 9. Car J, Black A, Anandan C, et al. The impact of eHealth on the quality and safety of healthcare: a systematic overview and synthesis of the literature. Report for the NHS Connecting for Health Evaluation Programme. 2008, https://www1.imperial.ac.uk/resources/32956FFC-BD76-47B7-94D2FFAC56979B74/ 10. Chaudhry B, Wang J, Wu S, et al. Systematic review: impact of health information technology on quality, efficiency, and costs of medical care. Ann Intern Med 2006; 144: 742–752. 11. Urquhart C, Currell R and Hardiker N. Nursing record systems: effects on nursing practice and healthcare outcomes. Cochrane Database 2009; (1): CD002099. 12. Campbell CJ and McDowell DE. Computer literacy of nurses in a community hospital: where are we today? J Contin Educ Nurs 2011; 42: 365–370. 13. Chan T, Brew S and De Lusignan S. Community nursing needs more silver surfers: a questionnaire survey of primary care nurses’ use of information technology. BMC Nurs 2004; 3: 1–8. 14. Darbyshire P. ‘Rage against the machine?’: nurses’ and midwives’ experiences of using Computerized Patient Information Systems for clinical information. J Clin Nurs 2004; 13: 17–25. 15. Dillon T, Lending D, Crews TR Jr, et al. Nursing self-efficacy of an integrated clinical and administrative information system. Comput Inform Nurs 2003; 21: 198–205. 16. Mills J, Chamberlain-Salaun J, Henry R, et al. Nurses in Australian acute care settings: experiences with and outcomes of e-health. An integrative review. Int J Inform Tech Manag 2013; 3: 1–8. 17. Hwang J and Park H. Factors associated with nurses’ informatics competency. Comput Inform Nurs 2011; 29: 256–262. 18. Chao C and Goldbert J. Lessons learned from implementation of a perinatal documentation system. J Obstet Gynecol Neonatal Nurs 2012; 41: 599–608. 19. Eley R, Soar J, Buikstra E, et al. Attitudes of Australian nurses to information technology in the workplace: a national survey. Comput Inform Nurs 2009; 27: 114–121. 20. Laramee AS, Bosek M, Shaner-McRae H, et al. A comparison of nurse attitudes before implementation and 6 and 18 months after implementation of an electronic health record. Comput Inform Nurs 2012; 30: 521–530.
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
11
Craswell et al.
21. Data Collections Unit. Perinatal data collection manual (ed Data Collections Unit Queensland Health). Brisbane, QLD, Australia: Queensland Health, 2010. 22. Queensland Government. Queensland Population Counter. 2014, http://www.qgso.qld.gov.au/products/ reports/pop-growth-qld/qld-pop-counter.php 23. Data Collections Unit. E-Bullitin, No. 28 (ed Health Statistics Centre Performance & Accountability Division Queensland Health). Brisbane, QLD, Australia: Queensland Health, 2011, pp. 13-22, http:// www.health.qld.gov.au/hic/ebulletin/facts_of_life28.pdf 24. Jones-Kelley A. Born Leader. Site Selection: The Magazine of Corporate Real Estate Strategy and Area Economic Development, May 2011, http://www.siteselection.com/issues/2011/may/images/Q-landTexasComparisonMapCMS.jpg 25. Creswell JW. Qualitative inquiry and research design. 2nd ed. Thousand Oaks, CA: SAGE, 2007. 26. Stern PN and Porr CJ. Essentials of accessible grounded theory. Walnut Creek, CA: Left Coast Press, 2011. 27. Pope C, Ziebland S and Mays N. Qualitative research in health care: analysing qualitative data. BMJ 2000; 320: 114–116. 28. Glaser BG. Theoretical sensitivity: advances in the methodology of grounded theory. Mill Valley, CA: Sociology Press, 1978, p. 164. 29. Australian Government. National statement on ethical conduct in human research (ed National Health and Medical Research Council ARC), 2007, pp. 1–98, http://www.nhmrc.gov.au/guidelines/publications/e72 30. Gliklich RE and Dreyer NA. Chapter 11. Data collection and quality assurance. In: Gliklich RE, Dreyer NA and Leavy MB (eds) Registries for evaluating patient outcomes: a user’s guide. 3rd ed. Rockville, MD: US Department of Health and Human Services, 2014, pp. 251–276, http://www.effectivehealthcare.ahrq.gov/ registries-guide-3.cfm. 31. Data Collections Unit. E-Bullitin, No. 22 (ed Health Statistics Centre Performance & Accountability Division Queensland Health). Brisbane, QLD, Australia: Queensland Health, 2009, pp. 7-20, http:// www.health.qld.gov.au/hic/ebulletin/facts_of_life28.pdf 32. Cook Carter M, Corry M, Delbanco S, et al. 2020 Vision for a high-quality, high-value maternity care system. Womens Health Issues 2010; 20: S7–S17. 33. Data Collections Unit. E-Bullitin, No. 29 (ed Health Statistics Centre Performance & Accountability Division Queensland Health). Brisbane, QLD, Australia: Queensland Health, 2011, pp. 15-23, http:// www.health.qld.gov.au/hic/ebulletin/facts_of_life29.pdf 34. Callen JL, Alderton M and McIntosh J. Evaluation of electronic discharge summaries: a comparison of documentation in electronic and handwritten discharge summaries. Int J Med Inform 2008; 77: 613–620. 35. Anderson S, Hardstone G, Procter R, et al. Down in the (data) base (ment): supporting configuration in organizational information systems. In: Ackerman MS, Halverson CA, Erikson T, et al. (eds) Resources, co-evolution and artifacts. London: Springer. 2007, pp. 221–253. 36. Craswell A, Moxham L and Broadbent M. Perinatal data collection: current practice in the Australian nursing and midwifery healthcare context. HIM J 2013; 42: 11–17.
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
JHI0010.1177/1460458214556372Health Informatics JournalCraswell et al.
Original Article
Does use of computer technology for perinatal data collection influence data quality?
Health Informatics Journal 1–11 © The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1460458214556372 jhi.sagepub.com
Alison Craswell and Lorna Moxham University of Wollongong, Australia
Marc Broadbent
University of the Sunshine Coast, Australia
Abstract Population health data, collected worldwide in an effort to monitor mortality and morbidity of mothers and babies, namely, perinatal data, are mandated at a federal level within Australia. The data are used to monitor patterns in midwifery, obstetric and neonatal practice, health outcomes, used for research purposes, funding allocation and education. Accuracy in perinatal data is most often reported via quantitative validation studies of perinatal data collections both internationally and in Australia. These studies report varying levels of accuracy and suggest researchers need to be more aware of the quality of data they use. This article presents findings regarding issues of concern identified by midwives relating to their perceptions of how technology affects the accuracy of perinatal data records. Perinatal data records are perceived to be more complete when completed electronically. However, issues regarding system functionality, the inconsistent use of terminology, lack of data standards and the absence of clear, written records contribute to midwives’ perceptions of the negative influence of technology on the quality of perinatal data.
Keywords attitude to computers, Australia, (MeSH) data collection, data quality, midwifery, perinatal nursing
Introduction The introduction of Information, Communication and Technology (ICT) into midwifery, like other areas in healthcare, is proposed to combat rising costs, decreasing availability of skilled staff as well as provide improved efficiencies of healthcare delivery and increase the portability and comparability of health-related data.1–7 However, evidence that the use of technologies in healthcare has achieved these outcomes is yet to be seen within the literature.8–11 This has not stemmed the Corresponding author: Alison Craswell, School of Nursing and Midwifery, Faculty of Sciences, Medicine & Health, University of Wollongong, P.O. Box 1128, Noosaville BC, QLD 4566, Australia. Email: [email protected]
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
2
Health Informatics Journal
flow of technology into healthcare. Despite the increased use of technology in healthcare practice, confidence and competence of ICT use by nurses and midwives varies according to age, computer use at home, education and training.12–15 Research concerning the use of technology within nursing and midwifery practice finds midwives have favourable attitudes towards the use of technology.16,17 However, results from pre- and post-implementation studies16,18–20 show that these positive attitudes lessen after implementation of computer technology.16 One such technology recently introduced into midwifery practice in Queensland, Australia, relates to perinatal data collection. Perinatal data are collected for mothers and babies Australia wide as mandated at a commonwealth government level. The data are used to monitor patterns in midwifery, obstetric and neonatal practice as well as for the planning of health services, research and the education of midwives.21 The research presented in this article took place in Queensland, a north-eastern state of Australia, with a population of approximately 4.75 million people, about 20.1 per cent of the current population of Australia.22 However, the state covers a large space geographically as can be seen pictorially in Figure 1, with the population scattered up the eastern coast, requiring many and varied health services to support residing childbearing women. The collection of perinatal data in Queensland have recently moved to an eHealth platform, superseding the traditional paper form, which is reported to have improved data quality, enhanced accessibility and communication, produced cost savings and improved the timeliness of the availability of collected data items.23 The introduction of technology for perinatal data collection adds to the increasing use of technology within healthcare and the daily work of midwives. Data presented in this paper were gathered as part of a larger study the authors were undertaking which examines the influences on midwives during the process of entering perinatal data into the computer. Only data from the analysis that lies within the theme of perceived data accuracy are presented and discussed here.
Method Grounded theory (GT), a qualitative methodology that has an inductive orientation, was used in this study to add a depth of discovery that may otherwise not have occurred with a quantitative approach. The voice of the midwife was captured and is considered important in determining what happened as it provides an experiential point of view.25 Purposive sampling was used to interview participants in line with GT methodology. This approach ensured that data were captured from participants with knowledge and experience of entering perinatal data into a computer for collection. Later, theoretical sampling, an approach that adds and refines properties and dimensions to acquire an in-depth understanding of analytical categories,26 was used to gather data from participants who would fill the gaps in the developing theory. By the end of data collection, 14 midwives and 1 health information manager from 12 different hospitals across Queensland, Australia, using three different systems for entering perinatal data participated in this research. The participants held a variety of positions within their organisations and ranged in experience from Level 1 midwives to clinical nurse consultants and educators. Some participants also held the position of perinatal data coordinator for their unit, which had the added role of validating and extracting perinatal data regularly at the end of an allocated period to be sent to Queensland Health, the governing body for healthcare at state level. Adhering to GT methods, the sample size was not pre-determined but influenced by saturation of the data rather than a specific required number of participants to meet generalisable sampling requirements.27 Saturation of the data occurred when no new information emerged from interviews and the theoretical framework was sufficiently populated to explain the phenomena under study.28 Participants were asked an initial open-ended question which was designed to encourage a full meaningful answer using the midwife’s own knowledge and experiences around the research topic.
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
3
Craswell et al.
Figure 1. Comparative size of the State of Queensland, Australia, to Texas, USA.24
Further discussion during each interview was unstructured, led by the participant but focused around the topic, adhering to GT methodology with the interviewer focusing the discussion back to the topic of perinatal data entry into the computer when necessary. Interviews lasted between 30 min to 1.5 h in duration, were recorded and later transcribed by the primary researcher. Ethical approval for the research was obtained from the University of Wollongong, Human Research Ethics Committee in 2012 with the research design adhering to the principals of justice, respect and beneficence.29 Data were analysed using the constant comparative method, again maintaining consistency with GT methods. Data analysis occurred concurrently with data collection over a period of approximately 10 months. Initial open coding using ‘in vivo’ codes and the writing of early memos occurred with early analysis giving some direction for theoretical sampling requirements for further participant recruitment. Development of codes, then themes and finally allocation of substantive or theoretical codes occurred in line with the systematised process of GT methodology and conceptualising what is happening in the data. Throughout this process, NVivo data management software was used to assist with organisation of the data.
Findings The findings presented here deliberately use the participants’ own words with ensuing discussion on the emerging themes, categories and associated factors. The conventions used throughout this article identify direct quotes using ‘I’ for interviewer and ‘P’ for participant. A major theme that emerged from this research was ‘perceived data accuracy’ which highlighted participant concerns over the accuracy of perinatal data entries since perinatal data entry has moved from paper to computer. This theme is made up of the sub-themes: (a) accuracy in records, both completed, computer records and written records and (b) data standards, specifically obstetric data standards, system functionality and inconsistent terminology. The relationships between these can be seen in Figure 2. Participants themselves questioned: Data are entered into a computer for perinatal data collection but is the data entered accurate?
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
4
Health Informatics Journal
Figure 2. Theme – perceived data accuracy and contributing sub-themes and associated factors.
Perceived data accuracy Perceptions of data accuracy by participants varied across interviews. Participants believed the data they entered was accurate but communicated that they knew it could be more ‘complete’. Perception regarding the accuracy of the data was communicated as in this example of records being returned with errors for correction: I: ‘So in your mind, do you think the data that’s in there is fairly accurate?’ P: ‘I wouldn’t make that assumption’. I: ‘So why do you think that?’ P: ‘Because sometimes what comes back from the perinatal people that they want confirming, it just makes me wonder if that (data) hasn’t been input correctly, I wonder if everything else has?’ This theme, ‘Perceived data accuracy’, is contributed to by four sub-themes: accuracy in perinatal data records, data standards, system functionality and inconsistent terminology. The findings in relation to these contributing sub-themes and associated factors will now be presented.
Accuracy of perinatal data records Computer records. It was generally felt by participants that moving to an eHealth platform for submission of perinatal data had improved the accuracy of the data entered. They found that the process of validation of an individual record before it could be finalised and submitted, forced them to enter data into fields that were easily left blank on the old paper record. Field validation is a construct of particular software that means data entered must fit set parameters for the page to move on or be saved. Logical checks on the data entered into the database are made against predefined rules for value ranges of the data to ensure they fit these parameters.30 Participants with the added role of perinatal data coordinator for their particular unit perceived that the number of returning errors from the perinatal data collection unit, the department in Queensland Health governing perinatal data, to be less than when the collection was done by paper
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
5
Craswell et al.
form. This was confirmed during discussions with staff at the data collections unit in Brisbane (C. Morris, 9 February 2011, personal communication) and also when examining hospital error reports and graphs published in Perinatal Data newsletters.23,31 One participant reported a 2- to 3-month lag time while using paper perinatal data forms but communicated this had subsequently been reduced since using a computerised extract of data. This was thought to be due to the validation function described above. However, some participants also reported uncertainty regarding inaccuracies in the completed and validated records. One participant reported: I can look in it (the record), but unless I sit there with the chart and check the (computer) entries I don’t know whether the data’s correct or not. No one does.
The time required to do such in-depth checking was reported as being unavailable to midwives and perinatal data coordinators. Another participant reported knowledge of inaccuracies in the data from their unit: We had no faith in the data from here because we knew it was very inaccurate even with the validation process.
These inaccuracies in computer data entry were generally not considered acceptable. Participants reported that a means of improving the quality of the information being sent to Queensland Health was designation of a perinatal data coordinator to check, correct and complete the electronic forms that midwives have entered data into at some point during care. With this model in place, some participants report being able to ensure the data are accurate: I: P:
‘So you think once you’ve done your clean, you’re fairly confident the data’s accurate? Yep, complete and accurate’.
This participant uses the term ‘clean’ in the context of correcting obvious errors in the data. Other participants still believed accuracy was inadequate and completeness could be improved: What I’m sending off is accurate but it’s not as robust as it could be.
The value of accurate perinatal data can also be seen by some maternity units giving perinatal data coordinator midwives between 4 and 8 h a week off-line time to check, complete and validate records prior to extracting the data to the data collections unit. The perceived level of accuracy reported from participants of this research varies from good to unknown. Accuracy does depend on a number of influencing factors including ‘busy-ness’ in the unit. Participants reported that when they were busy or pressed for time, they enter less data into the perinatal data record: I try not to skip over things, but I’m sure if I went down through all the, you know, adding things in, I’m sure we could pick up a lot more. Because I’m in a rush and someone else is just about to deliver and I need to go in there as well.
These busier times result in a requirement for additional staff, usually more casual or agency midwives, who may not be familiar with the system for entering perinatal data but are in some units still required to enter their own perinatal data. Midwives reported this as resulting in less-complete and less-accurate records.
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
6
Health Informatics Journal
Finally, concerns were communicated about the quality of the data that are printed off on discharge and go with the mother to the General Practitioner (GP) as this occurs before the coordinator has done a clean and check: You know at the end of the day you are supposed to print off a perinatal data once it’s all complete so you can send it to the GP and some of the things you find there, you say well you know the information these ladies are getting, that they’re reading is wrong and what’s going back to their GP is wrong.
Written records. When entering perinatal data into the computer, participants report utilising a combination of memory and the written chart. It was reported that the easiest and quickest entry of perinatal data occurred immediately after a woman had birthed when all the data were fresh in the midwife’s mind. However, this was not always possible due to the time constraints of the midwives’ role or a birth occurring on the change of shift, leaving no time for data entry. In these cases, perinatal data entry was handed over to another midwife or entered some time later. This could be on another shift or another day, entered by the birthing midwife or someone else. The resulting accuracy of such records was thought to be inadequate: Certainly if the accoucher’s (birthing midwife) not doing it, and you’re handing it on, then it can just be completely NOT accurate.
Participants suggested that the best practice for this occurrence involved all data to be entered electronically being written somewhere in the paper chart but also that this was not always the case or the process utilised for the systems in use by participants. The worst case scenario reported was when the discharge midwife went to enter the small amount of required discharge information and then check and validate the record to find there was no record created and therefore no data entered for the mother or baby at all. In these cases, transcribing data from the written chart was undertaken and participants voiced concern over relying on the accuracy of the written record: I: ‘And do you think the paper records are accurate? P: Probably not. Often they’re not. I’ve done documentation audits and there’s either things missing or … I find the same thing going through the paper record to complete the perinatal data. You know I’ve found records where I can’t find documentation of the apgars (a score of 5 items each being 0, 1 or 2 awarded at 1, 5 and 10 minutes post birth by the birth attendant) anywhere in the mother’s or baby’s notes, or a birth weight or something. So one would assume other things were missing that you’re not necessarily looking for’. Another participant reported looking up information in sources other than the relevant paper records to ensure accuracy: I always go in, I do always check the lady’s blood group. I don’t just take it as a given, what’s written in the handheld record in case something has been transcribed incorrectly.
Participants were concerned about the completeness and accuracy of data in written patient records that they are sourcing for entry into the computer for perinatal data collection.
Data standards Obstetric data standards. Participants reported that there is no standard for obstetric data collected via private medical consultations. Therefore, the data sourced from the private medical record
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
7
Craswell et al.
could be missing altogether, inconsistent with the field definitions of perinatal data or misrepresented: P: So sometimes the electronic data isn’t recorded anywhere on paper, only electronically and vice versa. I: Does that include the obstetric record as well? P: Yeah, because they’re all different from each obstetrician too, so no standard, no. An example of the misrepresentation of data is the number of ultrasound scans (USSs) which may be recorded to include only the mandatory clinical scans at 12 and 18 weeks: Realistically a lot of that stuff should be in the chart and some of it is inaccurate. Like the scans for instance.
Participants reported that when they questioned the mother further about the number of USSs performed, it was reported that the obstetrician has used the scanner to determine the foetal position and foetal heart rate at every visit. The data thus are a misrepresentation of the actual number of times the mother and the developing foetus is exposed to ultrasound. This may also relate to obstetricians interpretation of the term USSs to include only those for morphology purposes, clinical USSs looking specifically for congenital abnormalities.
System functionality Participants reported that some midwives have limited understanding of the nuances of the various software systems in use in Queensland for entry of perinatal data. An example was that midwives did not know that particular fields with drop-down boxes had scroll bars providing multiple items, some of which were hidden, for selection. This would indicate midwives had a lack of computer skill or training with the software for perinatal data entry and that the data selected are not always reflective of the appropriate category. Either another category is chosen or the information is left out: I actually had to teach someone about the drop downs the other day. About BGL’s (Blood Glucose Levels) and BSL’s (Blood sugar levels). They didn’t know there was a drop down box (for pregnancy diabetes) and that was only self discovery.
When the user does not know to use the scroll bar to scroll down a box to select an appropriate option, data go unrecorded which directly affect the collection of statistics relating to women and/ or the neonate.
Inconsistent terminology Field definitions in a perinatal data collection system are defined as the question that relates to a field within perinatal data for which information is entered. Findings from this research suggest that multiple understandings of field definitions exist across jurisdictions. An example can be seen with entering an apgar score (a score of 5 items each being 0, 1 or 2 awarded at 1, 5 and 10 min post birth by the birth attendant): … for instance the apgar scoring in our version, the electronic version we have, you need to actually put in the score for each of the 5 factors and then it automatically adds it up. Rather than put in a total, like 9 or whatever. And I quite regularly find somebody’s put 1 for the respirations at both 1 minute and 5 minutes
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
8
Health Informatics Journal and yet they’ve written respirations established at birth (in the medical record). And again I’ve spoken to people about that and of course they always deny it’s them that does that – they understand how to do apgars!
Similarly, where one unit may define ‘midwife led care’ as birthing with a midwife who has met the woman at an antenatal visit prior, another defined it as requiring a minimum of four or five visits with that midwife in a ‘know your own midwife’ scheme. Inconsistencies were apparent across the various systems and across fields within each perinatal data system used. In some systems, the field definitions written into the software for data extraction to Queensland Health did not exactly match the field definition required by Queensland Health. Therefore, the data extracted for that field are consistently incorrect and returned for correction or clarification to the perinatal data coordinator. These system inconsistencies increase the workload of error correction and clarification as well as potentially collect mismatched data between healthcare institutions: I: In terms of the ability of staff to enter complete and accurate information, if we did away with the paper chart tomorrow, do you think it would be a success? P: No, As I say we’re not at that level of literacy yet and I think our databases would have to be much more user friendly to do that. And you would have to have some little help boxes as well to explain to people. Something that you could call up to verify what it is, the information that you’re putting down.
Discussion Obstetric and midwifery practice today is primarily evidence based utilising research from data derived from sources such as the perinatal data collection in an effort to improve health outcomes and planning for future health service delivery needs.32 The findings of this research show that generally participants are concerned about the accuracy of data in the perinatal data entries they complete using a computer. This was demonstrated by participants communicating a clear understanding of the validation process and knowledge that data entered could be successfully validated by the system, yet not match the written record. Participant concerns relating to data quality persist despite perceived improvements in completion of fields and error return rates that moving to an eHealth platform for perinatal data collection are reported to have brought.33 Other Australian research comparing electronic discharge summaries to written versions found that moving to computers does not always improve the data quality, supporting this assertion.34 Perceptions of the accuracy of entered data are influenced by processes such as the transcription from paper to computer when written records are inconsistent. A reliance on memory results in entering of different information than that which exists on the paper record. Unavailability or inconsistency of information from sources such as obstetric records also contributes to inaccuracy due to differing data standards. In addition to this, system functionality when perceived as poor by participants may result in less information being entered or incorrect selections of data made. Finally, while inconsistent terminology persists, the standardisation of the data entered cannot be certain. Inconsistent terminology between staff and across workplaces in relation to data definitions is not a problem specific to midwifery but seen across the healthcare sector.35 Evidence suggests a solution is the use of standardised data but this is reported as difficult in other system implementations in the healthcare sector.35 The research undertaken and presented in this article found a lack of standard data in written sources used for transcription and the use of different computer systems for perinatal data collection potentially reduced the accuracy of data even when the midwife was
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
9
Craswell et al.
committed to completing the record accurately and in a timely fashion. The necessity of a perinatal data coordinator, which is both a solution and a recognition of a problem, to correct errors and complete entries prior to validation and extraction of data to Queensland Health, arises as a result of these inconsistencies within data standards of field definition consistency, written records and the computer. However, the varied nature of this position, the irregular hours provided for error correction and differing levels of concern about the quality of data between coordinators result in inconsistencies between units in the accuracy of data submitted to the governing data collection unit. In addition, the concern that summaries are printed off for the mother and the GP on discharge prior to this check suggests this is not always successful in preventing dissemination of inaccurate information. Midwives entering the data need to input the correct data at point of care rather than the coordinators being left responsible for error corrections. However, while barriers such as issues with system functionality and inconsistent understanding of field definitions persist, this expectation seems unreasonable. The consequences of inconsistent and inaccurate perinatal data entries are potentially enormous with ironically, inaccurate data directly affecting the areas serviced by the information resulting from very data being collected. Assuming that the perinatal data collection is of high quality when potentially flawed data are known to be entered leads to the risk of these same data being used to make major decisions in the evaluation and future planning for maternity services. Such misinformation puts the health of mothers and babies at risk. Involving the users in all areas of the development and implementation of computer systems for perinatal data collection will not only increase their knowledge of system function and therefore accuracy of perinatal data entries but also enhance buy-in of managers and staff, motivating them to use the system in a more appropriate way. With accuracy in perinatal data being most often reported via quantitative validation studies both internationally and within Australia,36 research that investigates user issues underlying these studies provides valuable information on how practices can be changed to ensure perinatal data are of a high quality within Australia.
Limitations This research is not without limitation as this study utilised a small purposive sample and use of a methodology that prevents results being generalisable to the midwife population at large or to other computer systems for population data collection. Further research to test the findings with a large population using quantitative methods would strengthen these results.
Conclusion and recommendations Participants of this research communicated concern about the accuracy of the perinatal data they enter for women in their care and believe the data they enter into each field in response to each question are accurate. However, they reported uncertainty regarding whether the data entered online actually matched that written in the chart. Issues of inaccuracy within the perinatal data collection place at risk the planning of health services in Queensland across all jurisdictions that rely on accurate information and statistics and as a result, the potential health of women and their babies utilising these services. The move from paper to eHealth collection of perinatal data is believed to have resulted in a more complete data collection than was previously experienced using paper forms, with turnaround times from submission of data to availability of statistics thought to have improved. However, while a validation process can improve the completeness of data, it cannot ensure the data collected by midwives match those in the written record. A lack of data standards for written records, inconsistent field definitions across maternity units and computer systems, as
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
10
Health Informatics Journal
well as the persistence of inaccuracies despite records being validated and complete, offer areas for improvement to ensure the data quality of the perinatal data collection is paramount. Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
References 1. Deloitte. National E-Health and information principal committee: National E-Health strategy (ed Deloitte). Canberra, ACT, Australia: Deloitte Touche Tohmatsu, 2008, pp. 1–121. 2. Eley R, Fallon T, Soar J, et al. The status of training and education in information and computer technology of Australian nurses: a national survey. J Clin Nurs 2008; 17: 2758–2767. 3. Healy J, Sharman E and Lokuge B. Australia: health system review. Health Syst Transit 2006; 8: 1–158. 4. National eHealth Transition Authority (NEHTA). NEHTA Strategic Plan Refresh 2011/2012. 2013, http://www.nehta.gov.au/about-us/our-strategy 5. Smedley A. The importance of informatics competencies in nursing: an Australian perspective. Comput Inform Nurs 2005; 23: 106–110. 6. Vimarlund V and Koch S. Identifying where the values come from IT-Innovations in health and social care. Intell Inform Manag 2012; 4: 296–308. 7. Yu P and Comensoli N. An exploration of the barriers to the adoption of information technology in Australian aged care industry. In: Walduck K, Cesnik B and Chu S (eds) HIC 2004: proceedings. Brunswick East, VIC, Australia: Health Informatics Society of Australia, 2004, pp.19–223. 8. BlackAD,CarJ,PagliariC,etal.TheimpactofeHealthonthequalityandsafetyofhealthcare:asystematicoverview. PLoS Med 2011; 8: e1000387 http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal. pmed.1000387 (accessed 8 February 2011). 9. Car J, Black A, Anandan C, et al. The impact of eHealth on the quality and safety of healthcare: a systematic overview and synthesis of the literature. Report for the NHS Connecting for Health Evaluation Programme. 2008, https://www1.imperial.ac.uk/resources/32956FFC-BD76-47B7-94D2FFAC56979B74/ 10. Chaudhry B, Wang J, Wu S, et al. Systematic review: impact of health information technology on quality, efficiency, and costs of medical care. Ann Intern Med 2006; 144: 742–752. 11. Urquhart C, Currell R and Hardiker N. Nursing record systems: effects on nursing practice and healthcare outcomes. Cochrane Database 2009; (1): CD002099. 12. Campbell CJ and McDowell DE. Computer literacy of nurses in a community hospital: where are we today? J Contin Educ Nurs 2011; 42: 365–370. 13. Chan T, Brew S and De Lusignan S. Community nursing needs more silver surfers: a questionnaire survey of primary care nurses’ use of information technology. BMC Nurs 2004; 3: 1–8. 14. Darbyshire P. ‘Rage against the machine?’: nurses’ and midwives’ experiences of using Computerized Patient Information Systems for clinical information. J Clin Nurs 2004; 13: 17–25. 15. Dillon T, Lending D, Crews TR Jr, et al. Nursing self-efficacy of an integrated clinical and administrative information system. Comput Inform Nurs 2003; 21: 198–205. 16. Mills J, Chamberlain-Salaun J, Henry R, et al. Nurses in Australian acute care settings: experiences with and outcomes of e-health. An integrative review. Int J Inform Tech Manag 2013; 3: 1–8. 17. Hwang J and Park H. Factors associated with nurses’ informatics competency. Comput Inform Nurs 2011; 29: 256–262. 18. Chao C and Goldbert J. Lessons learned from implementation of a perinatal documentation system. J Obstet Gynecol Neonatal Nurs 2012; 41: 599–608. 19. Eley R, Soar J, Buikstra E, et al. Attitudes of Australian nurses to information technology in the workplace: a national survey. Comput Inform Nurs 2009; 27: 114–121. 20. Laramee AS, Bosek M, Shaner-McRae H, et al. A comparison of nurse attitudes before implementation and 6 and 18 months after implementation of an electronic health record. Comput Inform Nurs 2012; 30: 521–530.
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
11
Craswell et al.
21. Data Collections Unit. Perinatal data collection manual (ed Data Collections Unit Queensland Health). Brisbane, QLD, Australia: Queensland Health, 2010. 22. Queensland Government. Queensland Population Counter. 2014, http://www.qgso.qld.gov.au/products/ reports/pop-growth-qld/qld-pop-counter.php 23. Data Collections Unit. E-Bullitin, No. 28 (ed Health Statistics Centre Performance & Accountability Division Queensland Health). Brisbane, QLD, Australia: Queensland Health, 2011, pp. 13-22, http:// www.health.qld.gov.au/hic/ebulletin/facts_of_life28.pdf 24. Jones-Kelley A. Born Leader. Site Selection: The Magazine of Corporate Real Estate Strategy and Area Economic Development, May 2011, http://www.siteselection.com/issues/2011/may/images/Q-landTexasComparisonMapCMS.jpg 25. Creswell JW. Qualitative inquiry and research design. 2nd ed. Thousand Oaks, CA: SAGE, 2007. 26. Stern PN and Porr CJ. Essentials of accessible grounded theory. Walnut Creek, CA: Left Coast Press, 2011. 27. Pope C, Ziebland S and Mays N. Qualitative research in health care: analysing qualitative data. BMJ 2000; 320: 114–116. 28. Glaser BG. Theoretical sensitivity: advances in the methodology of grounded theory. Mill Valley, CA: Sociology Press, 1978, p. 164. 29. Australian Government. National statement on ethical conduct in human research (ed National Health and Medical Research Council ARC), 2007, pp. 1–98, http://www.nhmrc.gov.au/guidelines/publications/e72 30. Gliklich RE and Dreyer NA. Chapter 11. Data collection and quality assurance. In: Gliklich RE, Dreyer NA and Leavy MB (eds) Registries for evaluating patient outcomes: a user’s guide. 3rd ed. Rockville, MD: US Department of Health and Human Services, 2014, pp. 251–276, http://www.effectivehealthcare.ahrq.gov/ registries-guide-3.cfm. 31. Data Collections Unit. E-Bullitin, No. 22 (ed Health Statistics Centre Performance & Accountability Division Queensland Health). Brisbane, QLD, Australia: Queensland Health, 2009, pp. 7-20, http:// www.health.qld.gov.au/hic/ebulletin/facts_of_life28.pdf 32. Cook Carter M, Corry M, Delbanco S, et al. 2020 Vision for a high-quality, high-value maternity care system. Womens Health Issues 2010; 20: S7–S17. 33. Data Collections Unit. E-Bullitin, No. 29 (ed Health Statistics Centre Performance & Accountability Division Queensland Health). Brisbane, QLD, Australia: Queensland Health, 2011, pp. 15-23, http:// www.health.qld.gov.au/hic/ebulletin/facts_of_life29.pdf 34. Callen JL, Alderton M and McIntosh J. Evaluation of electronic discharge summaries: a comparison of documentation in electronic and handwritten discharge summaries. Int J Med Inform 2008; 77: 613–620. 35. Anderson S, Hardstone G, Procter R, et al. Down in the (data) base (ment): supporting configuration in organizational information systems. In: Ackerman MS, Halverson CA, Erikson T, et al. (eds) Resources, co-evolution and artifacts. London: Springer. 2007, pp. 221–253. 36. Craswell A, Moxham L and Broadbent M. Perinatal data collection: current practice in the Australian nursing and midwifery healthcare context. HIM J 2013; 42: 11–17.
Downloaded from jhi.sagepub.com at Monash University on November 16, 2015
