Title. Judging nursing information on the WWW: a theoretical understanding.
Aim. This paper is a report of a study of the judgement processes nurses use when evaluating World Wide Web information related to nursing practice.
Background. The World Wide Web has increased the global accessibility of online health information. However, the variable nature of the quality of World Wide Web information and its perceived level of reliability may lead to misinformation. This makes demands on healthcare professionals, and on nurses in particular, to ensure that health information of reliable quality is selected for use in practice.
Method. A grounded theory approach was adopted. Semi-structured interviews and focus groups were used to collect data, between 2004 and 2005, from 20 nurses undertaking a postqualification graduate course at a university and 13 nurses from a local hospital in the United Kingdom.
Findings. A theoretical framework emerged that gave insight into the judgement process nurses use when evaluating World Wide Web information. Participants broke the judgement process down into specific tasks. In addition, they used tacit, process and propositional knowledge and intuition, quasi-rational cognition and analysis to undertake these tasks. World Wide Web information cues, time available and nurses’ critical skills were influencing factors in their judgement process.
Conclusion. Addressing the issue of quality and reliability associated with World Wide Web information is a global challenge. This theoretical framework could contribute towards meeting this challenge.
• Nurses have identified the Internet as a popular source for knowledge acquisition in practice.
• Accessing and filtering reliable World Wide Web information related to practice has been challenging for nurses.
• Many theories exist on nurses’ decision-making in clinical practice.
What this paper adds
• Nurses use a range of tasks relating to cross-checking and appraising of evidence and assessing user-friendliness, outlook, authority and practice relationship to overcome the complexity of judging the reliability of World Wide Web information.
• Intuition, quasi-rational cognition and analysis are the cognitive modes employed by nurses in their judgement processes.
• World Wide Web information cues, time available and critical skills are influencing factors when nurses apply tacit, process and propositional knowledge to judge World Wide Web information.
Implications for practice and/or policy
• Nursing publishers could be guided by the theoretical framework that emerged from this study to include clear metadata that would assist nurses in judging the reliability of World Wide Web information.
• The theoretical framework could be included in nurse education to help ensure that nurses are equipped with relevant skills to judge the best available web evidence to inform their practice.
• Nurses should be allowed sufficient time in practice to access the World Wide Web.
Decision-making in nursing is a complex issue. Many theories have been put forward to offer an understanding of this process in clinical practice. Historically, decision-making in nursing has been studied using a broadly dichotomous approach: systematic-analytical and holistic-interpretive. However, as experienced nurses are adopting a combination of these approaches (Lauri & Salantera 1995), the need for a unifying approach has been identified (Thompson 1999, Kennedy 2002, Cader et al. 2005). It has been suggested that the Cognitive Continuum Theory (Hammond 1981) has the potential to offer this unifying framework (Thompson 1999). The theory has merits in aiming for accuracy in clinical decision-making and ensuring that nurses’ decision-making processes are more transparent. However, in an analysis of this theory we identified limited adequacy in its application (Cader et al. 2005) to nursing. In relation to the World Wide Web (WWW), we concluded that the Cognitive Continuum Theory can only partially explain the judgement processes of nurses. Recently, the theory has been developed to embrace the culture of nursing (Standing 2007). As a global medium, the content of the WWW has major quality implications for the international community of nurses in their search for an evidence base for practice, and as such requires further research into its application.
The rapid growth of the Internet (Internet and WWW are used interchangeably in this paper), with the facility to update and retrieve information quickly, is tempting many professionals to shift from traditional information sources. Many healthcare professionals are using the Internet as an information source for clinical purposes and for updating their professional knowledge (Tsai & Chai 2005). Nurses are accessing online information to support their practice (Gilmour et al. 2007). According to Cobb (2003), the WWW is becoming an important source of information for healthcare professionals, with nurses forming the largest group. Jadad et al. (2001) found that 72% of nurses in the United States of America were using the Internet. In a study in United Kingdom (UK), Tod et al. (2003) concluded that levels of WWW skills and confidence remained low among nurses. Nonetheless, due to its accessibility and convenience, Canadian nurses have identified the Internet as a popular source of information for knowledge acquisition in practice (Estabrooks et al. 2003). Before becoming part of nurses’ knowledge, WWW information must first be accepted in terms of its relevance to nursing. WWW information is not aimed at a specific audience. Hence, nursing information is not clearly identified as such, giving rise to many unknowns and uncertainties (Marshall & Williams 2006). In the first instance, the area of uncertainty that needs to be resolved is that of judging the nature of the information in order to ascertain its relevance to nursing and its validity. Once this is established, nurses can then proceed to the next level of judgement, that of judging the reliability of the information.
Finding quality websites in the search for evidence supporting practice is challenging for many nurses (Beyea 2000). In clinical settings, quality evidence is essential because of the risk that unreliable information may pose to patient care. The implications of applying such information to clinical practice impose considerable professional responsibility on practitioners. The selection of what can be regarded as relevant and reliable information can make the evaluation of WWW information difficult. The complexity of judgement and decision-making is related to the level of risk-taking involved, which in turn can give rise to uncertainty. Such uncertainties relate to the many attributes associated with WWW information, the potential for inaccurate and incomplete published information and the safety implications of using such information in practice. However, according to Kershaw (2003, p. 30), ‘the advent of widespread online health information appears to be having a growing cultural impact both on patients and on health professionals’. Studies from various countries have highlighted the issue of quality of online health information (Eysenbach et al. 2002, Butler & Foster 2003,Pérez-López & Pérez Roncero 2006). In the printed media, ensuring quality has been the prerogative of authors, publishers and editors (Dougherty 1999). In the WWW environment, there appears to be a blending of these roles, making quality control difficult. The publication of potentially misleading and harmful information, for example, has been identified in a study by Gagliardi and Jadad (2002). In another study, Kunst and Khan (2002) have revealed gaps in the credibility and accuracy of WWW health information. When analysing Internet health information for quality, Purcell et al. (2002, p. 558) have stated that quality ‘is in the eye and ear of the beholder’, indicating that WWW users need to be heard. As evidence now indicates that nurses are increasingly using the WWW to meet their professional demands for information, it is important that their views as WWW users are considered.
The aim of the study was to gain an understanding of how nurses judge the reliability of WWW information related to nursing practice.
The grounded theory (GT) approach (Glaser & Strauss 1967, Strauss & Corbin 1998) was used because there was little knowledge and understanding of how nurses judge information on the WWW. GT is suitable to investigate those areas where limited knowledge exists (Benton 2000).
Two samples of nurses were selected for each of the two phases of the study. During Phase 1, data saturation was reached with 20 nurses undertaking a postqualifying graduate course at a UK university. In Phase 2, data saturation was reached with 13 nurses from a local UK hospital. The inclusion criterion was prior experience in using clinical information on the WWW in practice.
Data were collected between January 2004 and June 2005. For Phase 1 of the study, semi-structured face-to-face interviews were conducted using a list of core questions as an interview guide. The interviews lasted for 45–60 minutes. Three focus groups were also carried out during Phase 1 of the study to seek the views of participants on the initial findings. A summary of these initial findings was used as a focus for discussion during this process, allowing participants to comment on ‘the fit of the findings with their experience’ (Paterson 2001, p. 576). The group discussions lasted for 60–75 minutes. As it was difficult to distinguish whether participants’ comments related to information being retrieved for academic assignments or for clinical practice, it became necessary to contextualize the data and raise the level of risk and consensus in the second phase of the study.
During Phase 2, participants were asked to judge and select information from three web pages related to a scenario on an extended role of the nurse relevant to their area of practice, as if they were preparing for a 15-minute talk to their colleagues. These scenarios were designed with the help of Consultant Nurses from the local hospital who were working in four critical care areas of nursing. Interviews took place immediately following the viewing of websites and lasted for about 45 minutes. Participants were asked about the suitability of the information on the web sites for their 15-minute talk, the reasons for selecting the information, and the types of knowledge used to validate the selected information. The tape-recorded semi-structured interviews and focus group discussions were transcribed verbatim to avoid loss of meaning to the data.
The study was approved by the appropriate ethics committee. Participation was voluntary and anonymity was assured.
Data analysis and data collection were undertaken concurrently. A personal research diary was used to keep a record of impressions of the interactions with participants in order to enhance reflexivity. These notes were valuable in informing future data collection and in data analysis.
Following transcription of interviews, a summary was forwarded to participants to comment on its accuracy, offering the opportunity to feedback to the researcher on the correct interpretation of their views on evaluating WWW information. The stages of analysis put forward by Strauss and Corbin (1998) were used to analyse the transcripts. Following the first level of coding (open coding), Phase 1 of the study yielded 106 codes; these were reassembled during the second level of coding (axial coding) and 20 categories were developed. Following the Phase 2 analysis, 29 categories were obtained for both phases. All the categories were grouped to develop key categories. In addition, during data analysis, notes were kept on analytical thoughts and how emerging concepts related to each other. These notes, termed ‘theoretical memos’ (Strauss & Corbin 1998), assisted with the process of theoretical coding. ‘Cognitive evaluation’ emerged as the theoretical code that integrated all the other key categories.
In Phase 1 of the study, the nursing experience of the 20 nurses (18 women and two men) ranged from 5 to 15 years. The study was contextualized in the clinical setting in Phase 2 with 13 critical care nurses (10 women and three men) whose nursing experience ranged from 2 to 15 years.
The findings reveal an explanatory framework (Figure 1) illustrating the cognitive evaluation of WWW information related to nursing practice. The data suggest that during the process of evaluation the participants use intuitive, quasi-rational and analytical modes of cognition. These cognitive processes involve the application of tacit, process and propositional knowledge when judging various aspects (user friendliness, outlook, authority, relationship to practice and nature of evidence) of the WWW information. The nurses’ ability to implement these strategies depends on their critical skills, the level of cues presented with the WWW information and the time available. All the concepts of the theoretical framework are discussed below and illustrated in Figure 1.
Cognitive evaluation: the core process
Intuition, quasi-rational cognition and analysis (InQuAn) were identified as the key modes of cognition used by participants to judge the reliability of WWW information related to practice. Therefore, this cognitive framework will be referred to as InQuAn.
The key categories – user-friendliness and outlook – appear to have been judged purely intuitively. For example, some of their associated categories –‘navigation’, ‘user control’, ‘physical impact’ and ‘structure’ of the web pages – appear to have been judged subjectively:
…but for me as a consumer using it…I found the visual effects and the way it’s laid out…extremely helpful…it’s hard to explain…
The fact that this participant found it ‘hard to explain’ why these aspects of WWW information are helpful indicates that the information was being judged intuitively. However, an interesting attempt to explain the internalized nature of intuition was made in one of the focus group discussions:
Intuition is based on training and experience. The internal mix of that tends towards an intuitive judgement. I think that mix can be quite valuable if it can go along. To ignore it is to try to externalize everything and some decisions are made internally rather externally. You can’t look at anywhere else. You have to…sometimes make decision on grey areas without the support of anything that’s going to give us a clear direction. At that point you’re talking about intuition. You’re talking about that internal mix.
These findings support existing evidence on the intuitive approach to nurses’ judgement and decision-making in various nursing settings at many levels of the nursing hierarchy (King & Appleton 1997). This evidence came from a range of expert nurses (Benner & Tanner 1987) and health visitors (Wheeler 1992) to nursing students (McCormack 1993).
There is evidence that both intuitive and analytical thinking are employed for the tasks of assessing relationship to nursing practice, assessing authority of websites and cross-checking information. Hammond (1996) has explained that quasi-rational cognition is used when a task is both intuition-inducing and analysis-inducing. There is also evidence that, depending on the nursing task and context, some nurses apply both intuitive and analytical models in their decision-making process (White et al. 1992, Lauri & Salantera 1995). In this study, we found that when cross-checking information using on-line facilities, in addition to comparing WWW information with on-line evidence, participants also sought verification from relevant authority:
One of the things I like about the Internet is that you may e-mail the author as well. This should be an interesting exercise. If something that’s very interesting and has implications for practice but I’m not really sure of whatever I’ve read…there’s an opportunity to contact the author. I don’t think I’ll be hesitant to do that. May be add some of my peers. It carries a little more weight if someone receives an e-mail from a group of practitioners.
When checking with evidence from the print medium, WWW information was compared with potentially well-validated information, making this process analytical. This suggests that nurses appear to place more trust in print publications:
I’ve gathered shelves and shelves of paper…all relevant to particular subjects to do with family-centred care and if I do come across something it’s simply a case of turning from the web to the printed medium behind me and checking it out.
However, when consulting peers and colleagues, judgement was based on the level of trust the nurses had in their colleagues, making judgement less objective:
…I’d probably share it with peers I think and get some opinions back from a peer group, particularly if it was related closely to my practice and I think the opinion of peers can often highlight things that in our enthusiasm we have overlooked.
This extract supports the finding of Thompson et al. (2001) that colleagues have a strong influence on nurses’ decision-making. Interestingly, Eraut et al. (1995) have acknowledged that professional deliberations include both intuition and analysis.
When participants undertook the task of evaluating the nature of evidence analytical judgment was mainly used:
…would probably first of all scan a reference list for it and see if there’s a major piece of work that has been derived from a reliable source. I’ll even seek out extra pieces of work from the reference list and read them all in depth to again qualify that the article has gathered its information from a source…that’s reliable.
The scanning of a reference list in search of reliable sources is an objective and systematic approach to decision-making, requiring time and critical skills. The following extract from one of the discussion groups clarifies that the nurses were looking for scientific evidence in WWW information:
Participant C: The more references they give, the wider it looks…the better. However, the quality of references is more valuable than the number.
Participant A: Yea! Personally I would rather see less of referencing as long as they are heavily research-based.
In the clinical setting, Hamm (1988) defined analysis as slow, conscious and consistent. In our study, participants identified processes that they would engage in constantly to validate nursing information. For example, they consciously scanned reference lists, looked for research-based evidence and judged the author’s motivation. By going through these processes, they were applying an analytical approach, as explained by Hamm (1988).
Tacit knowledge, in this study, is equivalent to ‘aesthetic knowledge’ (Carper 1978). It is knowledge accumulated over time in the form of ‘expert opinion’. Assessing the user-friendliness of web information is one example where participants employed tacit knowledge. According to Eraut (2000), this form of knowledge is used in situations requiring quick action and which are too complex to analyse immediately. The process of assessing user-friendliness has to be done in real time, calling for a speedy judgement. The following quotation clearly shows that this participant was unable to articulate the nature of the knowledge used:
Generally, I think, I probably use a little bit of my own, my own background and my own…if you know what I mean…What I know and what I pick up in my work. I’m judging it a little bit on that.
According to Welsh and Lyons (2001, p. 304), this type of knowledge is classed as tacit, i.e. ‘that which nurses have which is often implicit, unstated’.
This form of knowing is acquired from integrating and applying propositional knowledge in practice settings through frequent interpretation of practice situations, responding appropriately and reflecting on feedback (Edmond 2001). ‘Experiential theoretical knowledge’ is how Rolfe (1998) labelled knowledge acquired through experience in practice. Process knowledge was used for tasks requiring a combination of propositional knowledge and personal experience, e.g. assessing the outlook, authority and relationship to nursing practice of websites. Our data suggest that participants’ previous encounters with similar types of information provide them with the knowledge to judge WWW resources:
I read the headline and depending on the headline I’ll know if it’s something worth reading, whether it would be relevant…Probably through experience, reading stuff previously and I can judge where the author is coming from.
Eraut (2000), p. 15 indicates that propositional knowledge refers to discipline-based theories and concepts. He further explains that propositional knowledge ‘includes propositions about skills and procedure, for example sets of instructions about how to dress a wound’. This form of knowledge was termed ‘empirics’ by Carper (1978), referring to knowledge drawn from empirical sciences related to nursing. Nurses seemed to have used this form of knowing for the subtasks associated with applying cross-checking strategies, assessing authority and relationship to practice and evaluating the nature of evidence:
I think it’s just knowledge of the subject. I would say you’ve got to have knowledge of the subject. Otherwise you would accept everything that you read. You could just accept it. So, you’ve got to have some background knowledge.
If you are looking at information on the effects of smoking cessation on children whose parents smoke you’ll also need the research on the parents’ views…really…to get a good picture of the situation.
Specific tasks to evaluate WWW information
The cognitive modes and forms of knowledge used by nurses relate to a range of tasks in the form of the dynamic processes of assessing user friendliness, assessing outlook, assessing authority of websites, applying crosschecking strategies, assessing relationship to practice and appraising the nature of evidence.
Assessing user-friendliness of websites
Our data showed that user-friendliness of websites is an important factor that can appeal to nurses when seeking information on the WWW. The facets of user-friendliness identified are: user control, navigation and functional features. Although reflected in the work of Bantjes and Cronjé (2000), the key category ‘user-friendliness’ in our study is much more sophisticated, with the inclusion of the hyperlink Internet facility. The identification of this facility may reflect the perceptions of professionals in the WWW context:
…so ease of use is quite an important thing. The breadth of links to other places…I think…I find very useful…so a site that isn’t in itself authoritative…I can always get to…other sites…and make some checks…
In addition to ease of use, this quotation supports the view that the process of validation of WWW information requires additional evidence; obtaining this is facilitated by the inclusion of hyperlinks on web pages. It is likely that nurses might not proceed to judge the information on websites if the elements of user-friendliness are absent.
Assessing outlook of websites
This key category is comprised of physical impact, structure and readability. For this task, participants discussed the way information is presented on the Internet. There is evidence that the physical presentation of information can be influential when nurses access information on the WWW to evaluate it:
First I would look at the clarity of the website, I’d look at whether it was well laid out, well set out, whether it was logical.
Readability also seems important when appraising WWW information:
I like the information to be clear and concise…you know…the language used has to be appropriate for nurses…I mean…not too technical…
The language it uses is easy to understand. I like the way it was presented. It was…easy read.
Assessing authority of websites
The subtasks associated with this relate to assessing authors’ background, evaluating authority of sources and judging trustworthiness. In this evaluative process, participants expressed concerns on the commercial motives of published information and its trustworthiness:
I don’t feel it’s completely reliable. You can’t always believe what they are telling you.
Some of the information is from companies trying to sell products. It could be that they’re telling you anything. Trying to sell their product, really. There is no proof that whatever it is they’re trying to sell is going to work, really.
Besides qualifications of authors, the participants have also identified authors’ clinical experience as important, as illustrated below:
Who is the author? Are they a recognised expert in their field? So you still have to critically analyse this sort of data.
Participants also categorized websites into authoritative and non-authoritative. Authoritative web sites include those from government organizations, academic institutions, professional organizations and global organizations:
I’ve only visited authoritative sites like the Department of Health, the UKCC, the RCN, the King’s Fund. I’ve never gone on to the Internet to look for other information because most of the work I’m doing is, for example, looking up data for the National Service Frameworks. And if I want more information on that, for example, I know I can get it from the Department of Health. I don’t need to look anywhere else.
Participants identified checking strategies when validating the reliability of WWW information. These include checking with information via the Internet, with peers and with the print medium. Using hypertexts was perceived by the nurses as a form of instant checking:
It’s better if there are lots of appropriate links so that you can go and find out more for yourself. You can also use the links to similar web pages to check the information they’re giving you.
A study by Eysenbach and Kohler (2002) also showed that crosschecking information on various websites facilitates the evaluation of information quality on the Internet.
Checking information with peers and colleagues adds another dimension to this crosschecking process:
…if you see something up there and you think…that’s marvellous…then you go and have a word with colleagues…nursing colleagues…you know…sometimes If you see something that sounds good from a website…you know…I would maybe go over the paper with them.
Assessing relationship to nursing practice
Assessment of information relating to nursing practice includes the concepts of currency, practical application and cultural origin. The following excerpts indicate the value that the nurses have placed on the currency of information and its relevance to safe practice:
I’ve got a habit of particularly looking at the referencing…as well…first…to see if they’re more up to date.
I think, when it actually relates to practice. I mean. I would value quite high because at the end of the day, you’re looking at developing your own practice, really…and to improve on your practice.
Information…on the Internet…can be unsafe to use.
Some participants were concerned that information from United States sources might not transfer well to nursing practice in UK. However, others had positive comments to make on this issue. It seems therefore that the cultural origin of the WWW resources was very important when judging the relevance of WWW information to nursing practice:
…information that comes across…especially on the web…is quite…even though you’ve put the UK plus on…it comes up American.
You know…a lot of American data…it’s really interesting but doesn’t relate to nursing practices in England.
I don’t know…I may be wrong but I do think we have a lot to learn from America and I think there is some relevant sort of stuff out there, personally I do yes…So I think we are getting closer to America and I think America is still very relevant. I don’t think it matters where you are as long as the evidence is there to be focused on and gain that information from…you know.
Appraising the nature of evidence
Substantial sections of the data indicate that information on the WWW is expected to be supported by research evidence and references. To judge this aspect of information, participants were looking for specific clues on the level of research evidence, references and diversity of information. A research approach with a broader perspective was advocated:
You’ve got to have the right methodology in place to work out what the answer to the question that you’re asking. I think we tend to look at research in a qualitative area. Whereas quantitative…medical data is very quantitative. Very specific enquiry.
This participant evaluated references to judge the author’s field of expertise:
…if you look at what the author is referencing from, then you’d have a good idea of what his field of expertise or certainly his field of reading has been for that particular piece of work.
Publishing diverse information is within the democratic spirit of the Internet. However, it does make the validation process more difficult, as one participant seemed to caution against information generated on an emotional basis:
I’d look at who has compiled the website, whether it’s a relative of someone who has died, whether it’s a parent of a child with a particular illness…whether there is an emotional aspect to the website.
A theoretically grounded sampling technique (Silverman 2000) was employed to ensure richness of data from nurses with experience in judging the quality of WWW. Nurses inexperienced in judging WWW information or those from a different clinical setting might have provided a different perspective on the issues identified. It is recognized that the sample size used in this investigation may not be representative of the wider population (Polit & Hungler 1997).
The InQuAn framework offers greater understanding of nurses’ judgment processes in a WWW context, as proposed by Cader et al. (2005). Previous studies have been undertaken with lay consumers of WWW health information (Eysenbach & Kohler 2002, Childs 2004). With the focus on nurses, our findings make a contribution to the existing body of knowledge relating to web users’ voices in validating WWW information. Focusing on web users ensured that the democratic nature of the WWW was respected. However, our intention is not to dismiss the findings derived from other approaches. Rather, this framework offers a broader understanding of the process nurse web users adopt when judging the reliability of WWW information.
Rippen and Risk (2000) have recommended that health information publishers should provide users with information that would allow them to judge this information. The InQuAn framework contains information to help nursing publishers on the web in deciding on the type of information to be included in their publication, thus empowering nurse web users to judge the reliability of that information to their practice. The inclusion of metadata on author expertise and the cultural origin of WWW information will facilitate the judgement of users. In addition, making appropriate propositional knowledge relating to a topic area available to nurse users will greatly assist them in being analytical in their decision-making process, as indicated in the InQuAn framework.
The contribution of the InQuAn framework to nurses’ judgement and decision-making is also significant. In the clinical setting, the approach to decision-making appeared to be twofold: intuition and analysis. Our findings suggest that not all aspects of WWW information can be judged either analytically or intuitively. A combination of intuition and analysis in the form of the quasi-rational mode of cognition has been shown to be applicable for some judgement tasks, which supports the findings of Lauri and Salantera (1995). This suggests that a unifying approach to decision-making is appropriate. The different modes of cognition used by our participants gives a more comprehensive picture of the process of nurses’ judgement and decision-making.
Lack of skills in appraising WWW information by nurses is one of the deficits in Internet education identified by Foundation of Nursing Studies (2000) and Ciliska et al. (2001). The inclusion of the InQuAn framework in their teaching could assist nurse educators to meet this challenge, giving them the opportunity to reconceptualize Internet education in nursing curricula.
Estabrooks et al. (2003) have asserted that nurses need rapid access to contemporary evidence-based care and many international websites are available for that purpose. However, the volume of information they contain in some areas of nursing practice can be limited as these online databases are in early stages of development. Seeking information from less authoritative websites is becoming an increasing option for nurses. However, finding relevant and high quality information from these websites can be daunting for many nurses (Hamer & Collinson 1999). The InQuAn framework offers nurses a clear understanding of the process of judging the quality of WWW information by showing the relationship between WWW judgement tasks, the forms of knowledge involved and the relevant modes of cognition required. However, to enhance this practice it is important that nurses are given sufficient time to access the wider section of the Internet at work, as evidence suggests that they are doing so only in short time periods (Morris-Docker et al. 2004).
Untangling the web is a challenge requiring a range of strategies. Strategies to address the issue of quality on the WWW have been developed by institutions and individuals. The InQuAn framework provides one contribution towards meeting this challenge. With the assistance of this framework, practising nurses are offered the opportunity and confidence to exploit the full potential of the WWW.
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Conflict of interest
No conflict of interest has been declared by the authors.
RC, SC and DW were responsible for the study conception and design, and drafting of the manuscript. RC performed the data collection, data analysis and provided administrative, technical or material support. SC and DW made critical revisions to the paper for important intellectual content and supervised the study.