What Is eHealth (4): A Scoping Exercise to Map the Field

Assessing the Taxonomic Structure of Research Databases and the Presence of eHealth

In the formative stage of the project, we explored the subject taxonomies, or thesauri, of multiple databases of abstracts in order to identify high-level subject headings which could be used to profile the volume and content of the medical informatics literature and to construct searches for pertinent evidence. In the case of Medline the thesaurus containing a hierarchical controlled vocabulary is referred to as Medical Subject Headings, or MeSH (see below). As part of this we sought to assess whether eHealth was explicitly represented within these thesauri. A further objective was to determine the ontological structure of the databases in relation to medical informatics and eHealth and the implied relationships between alternative subfields.

The databases examined were Medline (PubMed), the Cumulative Index of Nursing and Allied Health Literature (CINAHL), the Science Citation Index (SCI), the Social Science Citation Index (SSCI), the Cochrane Library Database (including Dare, Central, NHS Economic Evaluation Database [NHS EED], Health Technology Assessment [HTA] database, NHS EED bibliographic) and Index to Scientific and Technical Proceedings (ISTP, now known as ISI proceedings), all of which predate the targeted search period.

Exploring the Composition of the Medical Informatics Literature Using the Existing MeSH Thesaurus

MeSH has been developed (and is constantly updated by) the US National Library of Medicine. It consists of sets of terms naming descriptors in a hierarchical structure that permits searching at various levels of specificity. At the most general level of the hierarchical structure are very broad headings such as Anatomy or Information Science. More specific headings are found at more narrow levels of the eleven-level hierarchy, such as Ankle or Medical Informatics. There are 22568 descriptors in MeSH.

Historical trends in the literature indexed by the individual Medline MeSH terms subsumed within the broad Medical Informatics category were assessed for the period 1987 to 2003, and part way through 2004. Individual MeSH definitions were examined to assess the range and nature of the topics covered and to clarify which are most clearly related to common conceptions of eHealth (eg, specific applications of information technology (IT) to health care versus technical issues). The number of publications in Medline was profiled by year, as was the type of publication, subject to the limitations of the Medline categorization scheme (Randomized Controlled Trial/Controlled Trial/Meta-analysis/Review). In addition, the MesH tree was compared with an expert-derived taxonomy from the International Medical Informatics Association (IMIA) in order to assess its coverage of key areas and its merits as a means of identifying appropriate literature.

Using eHealth as a Search Term

Applying eHealth as free-text search term to multiple databases offered a “grounded” method of defining the field, as represented in the research literature. In order to identify publications specificially relating to eHealth and to place the concept within the wider medical informatics literature, all the databases described previously were searched for the presence of the word eHealth or its variants in the title or abstract for the period January 1, 1997 to December 31, 2003 (search string: Ehealth OR e-health OR e*health). Results were organized to show the number of articles arising each year, the journals in which they appeared, and the range of topics covered.

Profiling the Literature From Wider Web-Based Sources

Mixed methods were used to (a) identify current commentary and analysis relating to the emergence, nature, scope and potential of eHealth, and (b) locate evidence and opinions on general trends in technology and technology adoption with direct or indirect relevance to eHealth now or in the future. Relevant terms (including e health, e-health, ehealth, healthcare information technology and healthcare computing) were applied, singly and in combinations, to the Google search engine, which indexes over 8 billion URLs and ranks results by relevance and link popularity. In addition, websites previously identified as being likely to contain information relevant to eHealth were visited directly and scrutinized for pertinent information. In some cases, this was guided by the results of preliminary Google searches or by following up leads suggested in documents found earlier on, while in others it was guided by the existing knowledge of team members. As the searches were predominantly opportunistic and iterative in nature, it is inappropriate to try to document them exhaustively; however, the following types of information were targeted:

• previous exercises to map, scope or define eHealth;

• white papers, technical reports, predictions and early research reports on aspects of technology in health care, eHealth related policy, evaluation and trends, from the United Kingdom, Europe and beyond;

• funding programmes for eHealth- and/or health-and-technology - focused research and development;

• relevant articles from computing and information science-focused academic publications;

• eHealth and health technology-focused websites, web logs and online journals, online ehealth news feeds, email discussion groups and email newsletters;

• online sources with a focus on human-computer interaction, usability and accessibility, with specific attention on health care issues;

• technology-oriented news websites profiling general and health-related trends and developments;

• online studies, reports and statistical surveys relating to general technology take-up; consumer purchasing trends; attitudes and strategies of consumers and clinicians towards adoption of technology in general and for health care-focused tasks in particular; evaluation of the effectiveness of technological innovation, in the health care sector and beyond.

Given the increasing online availability of refereed academic literature there was inevitably some overlap between the information identified by the two searches.

Aggregating and Analyzing Definitions of eHealth

Scientific abstracts identified using the key word search were examined in order to assess the presence of definitions. While hand searching of full text articles was not a primary objective, this was done where easy Web-based access to this information was available. In the case of Web-based reports or commentary the definition was extracted from the page in which it appeared or was quoted. In both cases the initial extraction was performed by one research fellow and the results checked for inclusion eligibility by a second investigator. Our aim was not to perform an exhaustive and systematic review of definitions (because of time constraints) but to aggregate those appearing most easily and commonly in the research and wider arenas, as a means of supplementing our wider scoping study. The aggregated definitions were then analyzed thematically in order to assess the applications, stakeholders, contexts and theoretical perspectives targeted, so that the heterogeneity of conceptualizations could be determined. They were also considered with reference to the perspectives of the defining individual or organization and associated clarifications within the source document.

Assessing the Taxonomic Structure of Research Databases and the Presence of eHealth

Of the databases of scientific abstracts consulted, only Medline has a comprehensive hierarchical taxonomy of descriptors for the broad field of medical informatics. This part of the MeSH tree is shown in Figure 1. Medical informatics is also represented on CINAHL; however the subtree is relatively shallow and undifferentiated, forming only a small branch of the higher Information Science category, with many potentially relevant areas subsumed within other branches.

That eHealth has yet to be explicitly included among these thesauri, indicates the relative youth of the topic and the lack of an agreed conceptual definition. The literature relevant to eHealth is thus distributed among a range of existing MeSH fields.

The Medline MeSH structure for Medical Informatics contains 3 main subbranches: Public Health Informatics, Medical Informatics Computing, and Medical Informatics Applications. Examining the definitions of these and their lower order MeSH descriptors indicates that the Medical Informatics Applications tree encompasses the greatest number of component categories relevant to eHealth, taken broadly as the use of information and communication technologies to facilitate health care. For example, it subsumes the lower-order categories of Decision Making, Computer Assisted (which subsumes Computer Assisted Therapy and Diagnosis, among others); Information Systems (electronic information systems, networks, clinical decision support) and Information Storage and Retrieval (databases, laboratory information systems, etc). In contrast, Medical Informatics Computing is mainly characterized by an emphasis on systems and hardware, although it does contain MeSH descriptors relevant to eHealth — most importantly Internet, which may appear in eHealth publications as a specific technology or an application of technology. Public Health Informatics is concerned with the application of information and computer sciences to public health practice, research, and learning. Although this potentially encompasses eHealth-relevant research (for example, use of information and communications technologies for population health surveillance), the term was only recently introduced and has yet to contain any subcodes, limiting its usefulness at the present time. While the broader taxonomic categories each have their own character, there is clearly overlap between them. For example, decision support systems appear within both Medical Informatics Applications and Medical Informatics Computing, and electronic databases are a common feature in medical informatics applications, as well as representing a type of system.

Comparison of the MeSH tree with an expert-derived conceptual map endorsed by the International Medical Informatics Association (IMIA) revealed interesting differences in terms of the breadth of included concepts and their structural relationships (Table 2) [12]. For example, human and organizational factors appear to be underrepresented within Medline, while applications for consumers do not have a specific MeSH term (however, the IMIA taxonomy also appears to underrepresent consumer issues). This reflects the historical evolution of the MeSH hierarchy, which has been added to as the need arose by elaborating upon existing structures. Nonetheless, all the main areas apparently relevant to eHealth were encompassed by the MeSH tree and we are confident that using it as the basis of our search enabled the majority of pertinent literature to be identified.

Exploring the Composition of the Medical Informatics Literature Using Existing Taxonomic Systems

Figure 2 describes trends in the volume and nature of the literature indexed by the Medical Informatics MeSH descriptor (note that searching for MeSH terms in PubMed automatically includes the more specific MeSH terms in a search). There has been a steady growth in the volume of medical informatics research literature. The annual number of publications increased from 1987 to 2003 five-fold.

Publications indexed with MeSH keywords from each of the 3 main medical informatics MeSH subtrees (medical informatics computing, medical informatics applications, public health informatics) all follow this steady upwards trend, as do most narrower MeSH (eg, Information Systems; Therapy, Computer Assisted). However, the frequency of publications concerned with Clinical Laboratory Information Systems (Figure 3), appears to be decreasing, while research concerned with computer-assisted diagnosis increased rapidly in 2003 (Figure 4).

A breakdown of Medical Informatics MeSH, including definition, year of introduction, number and type of publications is supplied in Multimedia Appendix 1.

Using eHealth as a Search Term

As mentioned previously, there are currently no MeSH or equivalent coding categories in any of the databases searched which explicitly incorporate the term eHealth or its variants in their thesauri. This suggests that articles making reference to eHealth are being absorbed within existing classification schemes, such as Medline's Medical Informatics taxonomy.

When duplicates across databases were discarded we identified a total of 392 publications which explicitly referred to eHealth in the title, abstract, or journal title. Of these, most were represented in Medline. Appearing only in the Medline database were 283 (72%) articles, 54 (14%) only on the CINAHL database, and 55 (14%) only on the SCI, SSCI and ISTP databases.

Figure 5 illustrates trends in the volume of eHealth publications appearing across databases over time. This shows that the term did not start to be used in the research literature until 2000. References to eHealth showed a dramatic rise in 2000 to 2001 and, despite a small dip in 2002 a general upward trend persists. Note that we also retrieved publications from the Journal of Telemedicine and E-health which were picked up due to the journal name, not necessarily because they dealt with eHealth.

In Which Journals Do Publications Using the Term eHealth Appear?

In our study, publications containing the term eHealth were found in 154 different journals. A research fellow classified these by type, using a scheme agreed by the research team. The number of articles appearing within each journal were documented. Of the 387 publications found across multiple databases (after eliminating 5 that were clearly irrelevant), 77 appeared in clinical journals, 61 in health-services - related journals, 7 in finance-related journals, 4 in legal journals, 3 in journals related to medical education, and 28 in other journals not easily categorized. The journal titles with the most articles containing the term eHealth (n=9 for each journal) were the Journal of Medical Internet Research, Managed Care Interface, and Journal of AHIMA / American Health Information Management Association. The majority of publications were IT-related (207): however, among these, 116 articles were published in the Journal of Telemedicine and E-health, which were mainly picked up due to the journal name: only 4 articles actually contained the term eHealthin the abstract or title. Further details are provided in Table 3 and a detailed breakdown of journal titles is given in Multimedia Appendix 2.

What Topics are Covered in the Literature Using the Term eHealth?

In our study, in order to identify the topics dealt with in papers explicitly referring to eHealth, article titles and abstracts were examined by a research fellow and classified using narrative descriptors. This indicated that the most common topics are related to telemedicine (25% of publications) or the Internet (13%), while some (6%) are concerned with issues such as the scope of eHealth, future trends, or progress and challenges in the field. Note that this view is possibly biased towards the telemedicine field, as all articles published in the Journal of Telemedicine and E-health were retrieved, even if they did not mention eHealth specifically. Other topics are distributed across a range of diffuse areas such as antiterrorism and medical errors, none of which is represented by more than 4 papers (hence relevant percentages have not been calculated). A heuristic summary is provided in Figure 6, which highlights the key topics and subtopics identified. These results are based on preliminary analysis; further validation work is underway.

Definitions of eHealth

We identified 36 definitions of eHealth [13-52] appearing in published scientific abstracts and Web-based information sources (Table 4). As stated previously, our aim was not to perform an exhaustive and systematic review of definitions (which would have necessitated hand searching of full-text articles and reference lists), but to aggregate the most salient and easily accessible examples. Since many research databases are Internet accessible, there was some overlap between the definitions obtained by the two methods; however, they did yield largely unique results. In total, 36 definitions were identified. Definitions 1 to 15 were accessed via the research literature and 16 to 36 via the independent online searches, while 1, 5, 6, 7, 15 and 28 emerged from both searches.

Definitions were analyzed thematically in order to highlight specific technologies, applications or stakeholders referred to, and other theoretical concepts addressed, as detailed in Table 4. Analysis was initially performed by one investigator and the results checked by two others, thereby establishing agreement.

Our analysis suggests that there is significant variability in the scope and focus of existing definitions of eHealth both within the research literature and relevant sources on the World Wide Web. In terms of its functional scope, most definitions conceptualize eHealth as a broad range of medical informatics applications for facilitating the management and delivery of health care. Purported applications include dissemination of health-related information, storage and exchange of clinical data, interprofessional communication, computer-based support, patient-provider interaction and service delivery, education, health service management, health communities, and telemedicine, among others. A few narrow the concept down to specific applications, such as telemedicine or e-business, but these are the exceptions. While the range of applications is broad, a general theme relates to communication. One example is “E-health is connectivity; it is transactional; it is clinical. It is informational, interactive and interventional.”[43]

The majority of definitions (n=24) specify the use of networked information and communications technologies, primarily the Internet, and digital data, thus differentiating eHealth from the broader field of medical informatics, which incorporates “harder” technologies, such as scanning equipment, and bioinformatics research which tends to take place in isolation and is less directly applicable to health care service delivery. It is acknowledged that the Internet “…has the reach, the infrastructure, and the acceptance to achieve widespread change” [17] and it is envisaged that “Internet technology may rank with antibiotics, genetics and computers as among the most important changes for medical care delivery.”[16] Only 1 definition makes specific reference to harder technologies such as nanotechnology, robotics and laboratory tools [27], although another refers to Internet-compatible ICTs such as digital TV [40]. Of the 36 definitions identified, a sizable proportion make reference to telemedicine or telecare, either explicitly (7 examples) or in terms commonly used to describe these areas, such as delivery of care over distances. In most cases this is presented as part of a wider sphere of applications, although the definition from NHS Wales clearly identifies eHealth with telemedicine and telecare [45]. We identified 6 definitions that make explicit reference to business or e-business, although others contain related ideas such as the online trading of goods and services. In the majority of cases, such commercial applications are presented as merely one expression of eHealth.

In terms of the stakeholders considered to be the users or targets of eHealth, many definitions emphasize applications for providers and organizations–particularly those stressing electronic data exchange for clinical and administrative purposes. Others emphasize provision of information, education and services to consumers, including patients and “citizens”, with a small number clearly identifying eHealth with consumer health informatics [14, 46, 50]. Nevertheless the majority appear to encompass applications for all stakeholder groups, whether specified or implied by the breadth of the definition.

There is also variation in the degree to which alternative definitions consider wider theoretical issues, such as the influence of eHealth on society or on professional behaviour. Several highlight the changing cultural environment of health care; particularly growing patient empowerment (access to information and ability to use it), and point to the potential of eHealth to facilitate doctor-patient communication, partnership and shared decision making. Others emphasize the changes required to ensure that eHealth reaches its full potential, recognising that it requires new ways of working and attitudes and must take account of human and organizational influences affecting technology adoption and change. More broadly, eHealth is said to require a fundamental rethinking of health care processes and a commitment for networked global thinking to improve health care [22]. Overall, the definitions suggest a general excitement and optimism about the potential of this rapidly evolving field to improve health care processes and patient outcomes, and many clearly identify projected benefits such as improved clinical decision making, efficiency and safety.