Abstract
Identifies Internet health user types according to three factors: site attributes most favoured, health topics most sought, and the health sites they visited. Knowing what type of consumer uses a site is important, as this should inform menu structure and provide an understanding of why certain kinds of people visit certain sites. Web sites even in the same field will not appeal to all users. Much of this differentiation will come down to design and feel of the site, although site attributes and information needs also impact here. Data were obtained from an online questionnaire placed on the SurgeryDoor Web site, a site which provides consumer health information. In all a total of 1,068 responses were received. Shows that useful groupings can indeed be constructed. Users were also classified according to additional health sites visited, also a preference metric, and this additionally is related to site attribute preferences, information needs and user characteristics.
Keywords
Citation
Huntington, P., Nicholas, D. and Williams, P. (2003), "Characterising and profiling health Web user and site types: going beyond “hits”", Aslib Proceedings, Vol. 55 No. 5/6, pp. 277-289. https://doi.org/10.1108/00012530310498851
Publisher
:MCB UP Ltd
Copyright © 2003, MCB UP Limited
Introduction
There has been much investment in providing consumers with health information on the Web. Indeed, the UK Government has proved to be a major investor with its NHS Direct Online Web site. The rapidly expanding NHS site and its numerous commercial and non‐commercial rivals have proved exceedingly popular, if the number of “hits” they attract are anything to go by. Thus, in the case of the SurgeryDoor site over the period October 2001 to September 2002 3,680,453 pages were viewed (excluding declared robots). As part of a three‐year investigation (Nicholas et al., n.d.) on behalf of the Department of Health we have been evaluating the use and impact of a number of UK sites, including NHS Direct Online and the commercial SurgeryDoor Web site, although in fact many were “promiscuous” users visiting a number of other health Web sites – about which we also sought data.
The prime purpose of this particular study is to go beyond the headline hit figures and to identify and profile meaningful groups of health users, using a variety of criteria. This should help Web site managers to target their services more effectively, something that is beginning to concern them as they move towards providing a new range of personalised services. Without such an analysis there is a real danger that site owners (and consumers) will find themselves lost in the digital information fog. Site owners need to move on from hit counting and start obtaining meaningful user data. Most importantly, they need these data to help in menu and content construction, in that menus and content can be structured around identified information need groups. Menu and content structure could then reflect actual user interest rather than the producers’ or the site designers’ interest. Further, knowing more about user profiles and their site use leads to questions about a site’s information provision, presentation and also to a classification of sites. This can be thought of as working towards a classification that we are happy to apply in traditional media, for example, when referring to broadsheet and tabloid papers. The study described here constitutes a step towards such classifications.
Aims and objectives
The purpose of the paper is to identify and distinguish meaningful groups of Internet health consumers. This is necessary because there are so many consumers (millions, probably) searching on such a range of health matters that generic health sites can never hope to meet all their needs as they are currently configured. Using feedback obtained via an online questionnaire mounted on a leading consumer health Web site, SurgeryDoor, and employing a range of statistical measures, we sought to examine the characteristics of the respondents and then categorise them according to these characteristics.
Literature review
Little work appears to have been undertaken in exploring the characteristics of Internet health seekers – the literature is, rather, preoccupied with issues of quality (e.g. Impicciatore et al., 1997) or accessibility (e.g. Eysenbach and Köhler, 2002). A wealth of literature has accumulated on the classification of Web use health information needs, although they have largely been concerned with the categorisation of Web documents (Norbert et al., 1999; Oates and Gibson, 2003). Health on the Net Foundation (HoNF) (1999) and the Pew Foundation (2000) have carried out a number of Internet surveys of health information consumers. However, these are reports on monitoring the basic statistics of the Internet health community and users over time by age and gender.
For example, the Pew Internet & American Life Project did identify “health seekers” (55 per cent of the Internet‐user population) (Pew Foundation, 2000). Among the health seekers, 29 per cent go online weekly for medical information and 30 per cent monthly, while HoNF has found that a “steady advance of female users of Web‐based health‐care information” no longer applies (HoNF, 1999).
In an early non‐digital study Pinder (1990) developed a typology of health information seekers on the basis of how sufferers and carers coped with the onset and development of Parkinson’s Disease. Three categories of user were identified – “seekers”, “weavers”, and “avoiders”. The former sought as much information as possible, and used it as a central weapon in coping with the disease. Weavers sometimes sought information, and were selective about what they took on board. Avoiders believed that “the anxiety of not knowing was preferable to having their … fears about the condition confirmed”. Leydon et al. (2000) in their study of cancer patients also found some reluctance to seek information. Huntington et al. (2002) in a study of text‐based health information broadcast over a digital interactive television project identified four user types: active traditional media users, passive traditional media users, electronic passive/isolated users, and electronic sociable users. In particular, the paper identified new user types willing to explore health information via a DiTV platform. These users had not previously made use of this type of information. The authors raised the question as to the extent to which health information providers could use identified user groups to target specific users.
Cyber Dialogue (2000) found that approximately half of all Internet‐using health information seekers advised a family member or friend to see a doctor, changed their exercise or eating habits or made a “positive” decision related to their health treatment. A rather mixed bag of information need is shown in these results, with seekers looking for information on behalf of others, to improve their general health or to decide on the next steps they should take with regard to a current condition. Many others joined an illness support group after visiting a disease‐specific Web site. Some work has been carried out on the impact of the Internet on such specialist online groups. Gann (1998) reports, for example, that participation in these forums focused around peer support, sharing of information on treatment advances and clinical trials.
As with much else on the Web, it is often the marketers and consultancies which are leading the way in usage research, albeit in a fairly superficial way and with commercial rather than altruistic or academic ends. The Internet generation has been identified as being critical, demanding and sophisticated (Modalis, 2001). Further, marketers have categorised Internet users into economic groupings, such as the e‐fluential “online movers and shakers” (Burson‐Marsteller, 2000), while Jepsen and Dipnall (2003) looked at the classification of Internet use on ethical lines. One interesting study looking at online health information seeking has been that carried out by the Boston Consulting Group (BCG, 2001; Poensgen and Larsson, 2001) with European health consumers. Their researchers found that users tend to have a focused and deep interest in information only about their specific condition or disease. They do not regularly surf the Web for general health‐related material, but want sites offering specific information, and show little interest in using the Web to obtain general health‐related information or products. Those more actively involved in their diagnoses and treatment decisions are more likely to use the Internet as a resource for information.
Methods
Data for the project were obtained from an online questionnaire hosted on the SurgeryDoor Web site[1] for the month of November 2000. In total 1,068 users answered the questionnaire, which represented 5 per cent of the 21,118 visitors (as denoted by unique IP addresses) to the site in November 2000. All questions were closed questions and covered a variety of topics including age, gender, income and family circumstance, views on site attributes and the health information needs of the user. A total of 32 per cent of respondents were aged under 34, 50 per cent were aged between 35 and 54 and 18 per cent were aged 55 and over. SPSS was used to analyse the data.
Platform background
The consumer health information company, InTouch with Health, owns the publicly accessible Web site www.surgerydoor.co.uk (Figure 1). The site is frames‐based, and has menus on both the left and right frames (with “content” displayed on a middle frame). When it was launched on 27 January 2000, SurgeryDoor became the UK’s first Internet health portal offering electronic versions of official NHS information, and the country’s biggest online health multi‐store UK‐specific health Web site (M2 Presswire, 2000).
The site features targeted sections providing health advice for groups such as ’teens, carers, women, travellers and disabled groups. Site content is divided into the following sections, each of which appears as main content entries on a side‐bar and contain a series of subheadings:
- •
Health Daily. Consisting of such things as a news and health alert, weather, tips (such as “what to look for when considering joining a gym”), and feedback and suggestions.
- •
Medical. The first entry of which is “Emergencies”, and gives basic information on such things as bleeding, scalding etc. A “Symptoms Index” leads to the National Health Service online booklet Home Healthcare Guide, embedded in the SurgeryDoor window. Other content under the main “Medical” entry includes accounts of “Diseases in Depth”, a “Medical Dictionary” and a Prescription Drug Guide.
- •
Healthy Living. Including such topics as Preventing Accidents, Dental Health Drugs and advice on alcohol consumption.
- •
NHS & Benefits. Much of the content for this section is taken from the Government Benefits Agency and, indeed, appears to be pages directly from its own Web site.
- •
Complementary Medicine.
- •
Travel Health. Including vaccination advice and Traveller’s Health Kit (a list of items that may be useful during a holiday).
- •
Community & Fun. Consisting of the interactive elements of the site and current health information. Content includes: “Patient experiences”, “SurgeryDoor Health Magazine”, feedback and suggestions, health surveys, quizzes and competitions.
- •
Shopping.
- •
Today’s Selection of Features and Topics. Featured topics and articles such as “Living with Asthma” and “Hypnosis – its place in medicine today”.
Results
Satisfaction with the key attributes of the SurgeryDoor site
Figure 2 shows how the SurgeryDoor users who responded rated nine attributes of the site, including breadth of content, trust in the site, ease of navigation and advertisements on the site. They rated each characteristic out of 4, where 4 was very satisfied and 1 very unsatisfied. Trust (3.4) was rated most highly. We might have expected this, as clearly these were users of the site and they would hardly be users if they did not trust the site, especially in the case of a health site. Breadth of content and site navigation were similarly highly thought of (3.3). The least satisfactory characteristics proved to be e‐commerce opportunities (2.8), advertisements (2.7) and topicality (2.4). We might expect advertisements and shopping to score poorly – clearly some users were annoyed by these functions on a health site, and the dissatisfaction with currency shows the demanding nature of the digital health consumer.
Categorising users according to site characteristics most rated
Not all users will respond or relate to site attributes in the same way, and it would be valuable to type them according to the characteristics to which they do relate. Of course, some attributes might be seen to conflict with one another – for example, ease of navigational structure and breadth of content. There are conflicts between other characteristics as well; for example, our previous research (Nicholas et al., 2003) has found that where advertising is obtrusive there is a reduction in perceived trustworthiness in the site.
A factor analysis, which identifies un‐correlated or independent combinations of variables on the responses on how users rated each attribute, was conducted to see if meaningful groupings could be found. The results are reported in Table I. Three groups could be identified: those that rated content attributes, those that rated what we can term site facilities (e‐mail, etc.) and those that thought most highly of system attributes, like speed of response/download. These three groupings explained about 60 per cent of the variance related to these variables.
Site visitors who favoured “content” were interested in breadth, depth and trust of content. Users most interested in “facilities” were happy with advertisements, shopping and e‐mail facilities, while “system people” related favourably to the characteristics of speed of delivery and navigational ease.
An important relationship was found between the user’s score on the content attribute (breadth, depth and trust) and the number of different health sites visited. The user’s rating fell, as the number of sites visited increased (Figure 3). This suggests that there is a group of people that believe they have to visit many a site to find what they want. We call these people site checkers or evaluators. They largely do this checking on the basis of long experience with searching the Web, practice in making constant comparisons and a process of trial and error. The Web provides huge opportunity to “suck it and see” and these people take advantage of this.
A relationship was also found between facility preference (i.e. for advertisements, shopping and e‐mail) scores and age and gender. Figure 4 demonstrates the relationship with age: the user’s facility score decreased as age increases. This might result from a number of factors including the fact that older people might be less familiar with Internet technology like chat rooms and e‐mails, have more fears about the security of online shopping and disapprove more of advertisements. In addition there was some gender bias with women being more likely than men to be happy with “facilities” (Figure 5). This relation was also repeated for system attributes. Women were also found to be happier than men with the system attributes – that is they tended to be more satisfied with speed of delivery and navigation ease.
A relationship was also found between system attributes and the number of visits made to SurgeryDoor (Figure 6). Those people who visited the site less frequently (just once a month) scored lower on system attributes (speed and navigation). Logic suggests that these people were irregular users because of these factors (which they did not like) or perhaps they were just unhappy with a site with which they were not familiar. Furthermore, some users may have a slow Internet connection and consequently blamed the site rather than their own computer set‐up for this. Attitudes towards system site attributes were also found to be related to the user’s health interest in the site (Figure 7). Carers and those users currently suffering from an illness were found to be much more positive with regard to the site’s system attributes.
Health topics most sought
Clearly content must be one of the most important reasons for using a health Web site. But precisely in what type of content are people interested? Figure 8 provides the details and shows user preference scores for 12 health topics, such as New Treatments, Natural Health, Pregnancy etc. Users rated each as very important, important, not so important or not important. Scores again were out of 4, where a 4 indicated that the topic was very important. The two most importantly rated topics were General Health information (3.3) and Diet (3.2). Poorly performing topics included information on Medical Conditions (1.8), Pregnancy (2.2) and Support Groups (2.3). This may well reflect user interest in coming to this particular site and maybe they get the latter information elsewhere.
Categorising users according to health topics sought
A factor analysis was conducted to identify specific user groups by topic of interest. The results are reported in Table II. Four types of “topic” user groups could be identified: “alternative remedy users”, “staying fit and healthy users”, “keeping up to date users”, and “ill but want to know more users”. The combined factors accounted for about 60 per cent of the variance.
Alternative remedy users rated the two topics Natural Health and Complementary Medicine most highly, while the fit and healthy group rated Healthy Living, General Health and Diet topics most highly. The third user type identified rated Medical News and Medical Research highly, which suggests that these users “want to keep up to date”. The fourth user type rated Prescription Drugs and New Treatments highly, which suggests a type of user who may be “ill but wants to know more” about what they have been prescribed and about new treatments.
Age, gender and for whom the person was searching were significant characteristics of the alternative remedy group. Thus, alternative remedy users tended to be women under the age of 34 (Figure 9) and tended to search on behalf of friends and children (Figure 10). Previous research by the team has shown that information passed on by a friend is a poorly trusted information source and hence was unlikely to be used (Nicholas et al., 2002). However, these users were also using this information for themselves and for their dependants (Figure 10).
The fit and healthy information user group was found to relate to the user’s current health status (Figure 11), and those who were currently healthy scored highly. These users, it seems, were accessing content so as to “stay” fit and healthy and to check health requirements etc. A low positive correlation (0.27) was also found between “content” attributes and this type of user, suggesting that these users were also interested in depth, quality and trust of content.
The “ill but wants to know more” type of user was found, perhaps unsurprisingly, to relate to the person’s current health status (Figure 12), with carers, those currently suffering and those who had a long‐term illness being more likely to feature in this group. This was also true to a certain extent of users identified as wanting to keep up to date. Those people with a long‐term illness also sought to keep up to date with health information. Furthermore, those users searching for other people and for children also featured strongly in the “Ill but wants to know more” user group (Figure 13).
Profiling users by the health site searched
It is evident that different user types might be attracted to different types of health sites. A one‐size, all‐purpose health site that fits all user types and interests of users is unlikely to exist. People may be attracted to a site for style differences as well as site attribute factors and content differences. Also, old and young users may distinguish themselves by adopting different site‐visiting behaviour. At present we do not have access to comparable data for other sites. However, we can, in the context of the current study, look at user differences revealed between SurgeryDoor users who have visited different combinations of health sites. Here we will look at site attribute, topic interest and user characteristic differences between SurgeryDoor users who also use NHS Direct Online and with SurgeryDoor users who also visited Net Doctor. The aim of the analysis is to see if these differences reveal any information about the profiles of users for the various consumer health sites.
A total of 29 per cent of respondents said that they visited just one site for their health information, 71 per cent of respondents visited two or more sites and 39 per cent visited three or more different sites. Figures 14 and 15show which site people visited first and which ones they additionally used. Not surprisingly, perhaps, about nine out of ten users said that SurgeryDoor was their first choice. NHS Direct Online, Net Doctor (www.netdoctor.co.uk), Medic Direct and Health in Focus were also mentioned as a first choice but the percentages were insignificant, between 1 to 4 per cent. These four sites, however, were significant as second sites visited. Net Doctor and NHS Direct Online were clearly important, each attracting around a third of responses as a second site preference.
It was decided to restrict the between‐site analysis to only those respondents who said that they had visited two or more sites. This was done so as to exclude from the analysis comparisons between those users just visiting one site and those visiting more than one site. In addition it was decided to exclude those respondents who said that they visited both NHS Direct Online and Net Doctor.
A logistic regression model[2] was used to determine which site attributes, topics and user characteristics would explain differences between people using different combinations of sites. All the variables reveal something about users and variables indicated by the analysis will give an idea of the likely profile/characteristics of the different user groups. Three models were considered:
- 1.
(1) a comparison between those SurgeryDoor users that used the NHS Direct Online site and those that did not;
- 2.
(2) a comparison between those SurgeryDoor users that used and did not use the Net Doctor site; and
- 3.
(3) a comparison between NHS Direct Online and Net Doctor users.
Not all users here were SurgeryDoor users; we are simply grouping and comparing users of the two most important second sites visited with those that did not use these sites. Table III lists the significant coefficients.
There is an increased tendency for Net Doctor users on the SurgeryDoor site to be younger compared with other SurgeryDoor users. Net Doctor users were just under half as likely to be aged 55 and over and were about one‐third less likely to be aged between 34 and 54 compared with other SurgeryDoor users. The over‐55s made up 18 per cent of total questionnaire responses, while the age group 35 to 54 made up about 50 per cent of responses. A total of 40 per cent of Net Doctor users were aged 34 and under compared with 29 per cent of other SurgeryDoor users. Age of user was also a significant difference between NHS Direct and Net Doctor users. The Net Doctor site seemed to attract a younger user type. Why should this be? This age profile might say something about how the site is presented. For example, there is a greater use of menu icons on this site compared with either SurgeryDoor or NHS Direct Online. Content is also considered an important attribute in attracting a younger user profile.
In comparing NHS Direct Online users with other users, salary was found to be an important distinguishing factor. There was an increased incidence of a person being an NHS Direct Online user as income levels increased. This relationship is shown in Figure 16. It shows that 41 per cent of those SurgeryDoor users earning between £45,000‐59,000 were also NHS Direct Online users but this was only true of 24 per cent of those earning less than £15,000. NHS Direct Online seems to attract users with a wealth profile. The wealth variable here may be an indicator for other variables that relate to it, for example, the user’s education and class. The NHS Direct Online site might be appealing to this type of user. Hence the NHS Direct Online may well be perceived as a more upmarket or a traditional, safe and conservative site – a broadsheet rather than a tabloid type of health site. NHS Direct Online users were also significantly more likely to rate general health and to a lesser extent medical news information as important or very important.
Family characteristics was also important, with NHS Direct Online users being less likely to be “live alone” users. This was truer in a comparison with other SurgeryDoor users than with Net Doctor users. Family characteristics may be important in that families may be more likely to have regular contact with the NHS generally and might well search specifically for NHS Direct Online because of this.
Information interest types
Users identified as wanting to stay fit and healthy do differentiate or discriminate between the sites. Thus the staying fit and healthy user type tended to be more likely to be also an NHS Direct Online user and less likely to be a Net Doctor user. Figure 17 describes this relationship and includes those users who had visited both sites.
Those who used SurgeryDoor, who also used NHS Direct Online, were more likely to be a “staying fit and healthy” type user, while Net Doctor users scored poorly and this type of user was not attracted to the Net Doctor site. Perhaps Net Doctor users are not so interested in staying fit and healthy but are more interested in the health limitations and implications for their life style. Again this may say something about the content difference between the two sites (Net Doctor and NHS Direct Online) and how each site markets itself.
System attributes
The user rating of navigational structure was found to be a distinguishing factor. Thus, Net Doctor users were just over twice as likely to rate the navigational attribute as being important compared with either other SurgeryDoor users or users who additionally used NHS Direct Online. NHS Direct Online users were half as likely to rate navigation as very important. This may reflect real navigational differences between the three sites. This argues that the market reaction is that the NHS Direct Online navigational structure is poor, while for Net Doctor it has been rated as good. Alternatively, this may say something about the attitude differences between NHS Direct Online users and Net Doctor users. NHS Direct Online users do not seem to rate navigational structure as important and this might imply a stronger brand attachment to NHS Direct Online. These users are not so interested in site attributes as in NHS Direct Online itself. However, for Net Doctor users navigational structure is important – these users seem to be site attribute‐driven and may well represent more of a floating, less brand‐loyal audience, and may well jump site to another offering a better attribute combination.
Conclusion
This paper has identified discrete sub‐groups of health consumers and developed an understanding of the relationship between user types, user characteristics and site preferences. This type of research should be useful in informing designers of Web sites, especially with regard to menu construction. Patently, sites should now move on from broadcasting information generally to the “average” user to assisting specific groups of users to find what they are seeking. Thus, for instance, alternative medicine information seekers should be able to use the menu structure to quickly retrieve the information they need. Menus should reflect actual user interest rather than – as is too often the case – the producers’ or the site designers’ interest. To assist in this our study has provided user profiles for a range of consumer health sites and developed arguments around the sites’ information provision and presentation. Site producers and information providers can determine interests both by surveying users (and potential users, offline, if possible) and by analysing pages accessed – log analysis.
Users were grouped according to the site characteristics favoured – content provided, facilities offered and system functions. Users were also identified by their topic of interest: “alternative remedy” user; “staying fit and healthy” user; “keep up to date” user; and “I’m ill but want to know more” user. Previous Internet health researchers have not used these groupings. The key findings about these groups were:
- •
With regard to user groups by site characteristics favoured, a relationship was found between the user’s score regarding the importance of content (breadth, depth and trust) and the number of health sites visited. It was found that the user’s content rating decreased as the number of sites visited increased, characteristic, one would have thought, of the “promiscuous” digital information consumer. Furthermore, the user’s facilities (advertisements, shopping and e‐mail) rating decreased as age increased. There was some gender bias with women being more likely than men to be happy with advertisements, shopping and e‐mail services.
- •
Four types of Internet users were identified by their topic of interest: “alternative remedy” user; “staying fit and healthy” user; “keeping up to date” user; and “I’m ill but want to know more” user. Alternative remedy users tended to be women under the age of 34 and further tended to search for a friend or for a child. The “ill but wants to know more” type of user was found to relate to the user’s current health status and whether they were carers or not. Not surprisingly perhaps, those currently suffering from an illness and those who had a long‐term illness were more likely to feature in this group, as well as those who were looking for information on someone else’s behalf and were interested in both new drugs and treatments.
- •
It was possible to characterise users according to the pattern of health sites visited. Net Doctor users on the SurgeryDoor site were more likely to be younger compared with either other SurgeryDoor users or NHS Direct Online line users. The staying fit and healthy type of SurgeryDoor user tended to be more likely to be also an NHS Direct Online user and less likely to be a Net Doctor user. This may say something about the content difference between the two sites (Net Doctor and NHS Direct Online) and how the site attracts users. In was found that there was an increased incidence of a user being an NHS Direct Online user as income level increased.
Notes
- 1.
1 www.surgerydoor.co.uk 3W Marketing Ltd were responsible for gathering the information but not its analysis.
- 2.
2 Logistic regression was used in the second part of the analysis. Logistic regression is similar in “concept” to least squares linear regression, though its procedures, assumptions and underlying statistical model are different. Logistic regression is used whenever an outcome event can be classified into two populations. It is used here with regard to the outcome if a user made use of a secondary site (NHS Direct or Net Doctor): the user either did or did not. In particular the model is used to decide which users’ characteristics are predictive or significant for the outcome occurring. Logistic regression in addition estimates odds, tt is it says something about the impact of each information source on the outcome. The odds are the likelihood of the outcome occurring, given the user‐graded importance of each information source.
References
Boston Consulting Group (BCG) (2001), “Most European e‐health ventures are failing because they rely on typical, retail e‐commerce business models”, available at: www.bcg.com/media_center/media_press_releases.asp (accessed 13 June 2003).
Burson‐Marsteller (2000), “The e‐fluentials”, available at: www.efluentials.com. (accessed 13 June 2003).
Cyber Dialogue (2000), “Cybercitizen Health 2000”, available at: www.cyberdialogue.com (accessed 5 January 2001).
Eysenbach, G. and Köhler, C. (2002), “How do consumers search for and appraise health information on the World Wide Web? Qualitative study using focus groups, usability tests, and in‐depth interviews”, British Medical Journal, Vol. 324, pp. 573‐7.
Gann, R. (1998), “Empowering the patient and public through information technology”, in Lenaghan, J. (Ed.), Rethinking IT and Health, Institute for Public Policy Research, London.
Health on the Net Foundation (HoNF) (1999), 5th HON Survey on the Evolution of Internet Use for Health Purposes: October‐November 1999, available at: www.hon.ch/Survey/ResultsSummary_oct_nov99.html (accessed 13 June 2003).
Huntington, P., Nicholas, D., Williams, P. and Gunter, B. (2002), “Characterising the health information consumer: an examination of the health information sources used by digital television users”, Libri, Vol. 52 No. 1, pp. 16‐27.
Impicciatore, P., Pandolfini, C., Casella, N. and Bonati, M. (1997), “Reliability of health information for the public on the World Wide Web: systematic survey of advice on managing fever in children at home”, British Medical Journal, Vol. 314, pp. 1875‐81.
Jepsen, D. and Dipnall, J. (2003), “Ethics in leadership: the Internet perspective”, available at: www.mrsnz.org.nz/Files/Conf‐JoD.PDF (accessed 13 June).
Leydon, G.M., Boulton, M., Moynihan, C., Jones, A., Mossman, J., Boudioni, M. and McPherson, K. (2000), “Cancer patients’ information needs and information‐seeking behaviour: in‐depth interview study”, British Medical Journal, Vol. 320, April, pp. 909‐13.
M2 Presswire (2000), “UK‐specific health Web site aims to capture a fifth of online health market”, M2 Presswire news release, (Lexis‐Nexis Universe UK News database).
Modalis (2001), “Modalis market tracks Internet user – post‐hype report”, May, available at: www.modalis.com (accessed 13 June 2003).
Nicholas, D. Huntington, P. and Williams, P. (n.d.), “The Web, the kiosk, digital TV and the changing and evolving face of consumer health information provision: a national impact study, April 2000‐January 2002”, DoH.
Nicholas, D., Huntington, P. and Williams, P. (2002), “The impact of location on the use of digital information systems: case study: health information kiosks”, Journal of Documentation, Vol. 58 No. 3, pp. 284‐301.
Nicholas, D., Huntington, P. and Williams, P. (2003), “Perceptions of the authority of health information: case study: digital interactive television and the Internet”, Health information and Libraries Journal, (in press: September 2003).
Norbert, G., Mounia, L. and Norbert, F. (1999), “A probabilistic description‐oriented approach for categorizing Web documents”, Proceedings of the 8th International Conference on Information and Knowledge Management 1999, Kansas City, MO.
Oats, S. and Gibson, R. (2003), ECPR 2003 Workshop: The Changing Media and Civil Society, available at: www.essex.ac.uk/ecpr/jointsessions/edinburgh/details/ws%2020.pdf (accessed 7 February).
Pew Foundation (2000), Pew Internet and American Life Report: The Online Health Care Revolution: How the Web Helps Americans Take Better Care of Themselves, available at: www.pewinternet.org/ reports/reports.asp?Report=26& Section=Report Level2&Field=Level2ID&ID=134
Pinder, R. (1990), The Management of Chronic Illness, Macmillan, Basingstoke.
Poensgen, A. and Larsson, S. (2001), Patients, Physicians, and the Internet: Myth, Reality, and Implications, Boston Consulting Group, Boston, MA.