Social network analysis on consumers' seeking behavior of health information via the Internet and mobile phones

Journal of Korea Multimedia Society Vol. 17, No. 8, August 2014(pp. 995-1011) http://dx.doi.org/10.9717/kmms.2014.17.8.995

Social network analysis on consumers’ seeking behavior of health information via the Internet and mobile phones

Ji-Young An

, Haeran Jang

, Jinkyung Paik

ABSTRACT

In consideration of the rapid changes in the so-called information society of the 21

century, about 80% of a total population in Korea has used the Internet. However, the social effect of the Internet and related devices has not been yet systematically studied in the literature. In healthcare as well, consumers’

efficient use of the Internet for their positive health outcomes is becoming an issue. The purpose of this study was to analyze the medical subject headings keywords of the selected studies on consumers’ use of Internet and mobile health information. For the analysis, social network analysis was used to provide basic information to present directions for future research on the field of interest.

Key words: e-Health, Internet, Mobile Phone, Health Information, Social Network Analysis

※ Corresponding Author : Haeran Jang, Address: (312- 702) 5TH Floor, R&DB Foundation, 201 Daehak-ro, Chubu-myeon, Geumsan-gun, Chungnam, Korea, TEL : +82-41-750-6264, FAX : +82-41-750-6396, E-mail : hljang

@joongbu.ac.kr

Receipt date : Jul. 8, 2014, Revision date : Jul. 21, 2014 Approval date : Jul. 23, 2014

†

u-Healthcare Design & Healthcare Service Design Development Program, Design Institute, Inje University (E-mail: ajy0130@inje.ac.kr)

†††

Herbal and Health Management, Joongbu University

†††

College of Design, Design Institute, Inje University (E-mail: dejpaik@inje.ac.kr)

※ The part of this paper was presented at MITA2013 which was held in Bali, Indonesia on July 2-6, 2013 (Also winner of MITA2013 best paper award). This work was supported by the Korean Research Foundation Grant funded by the Korean Government (NRF-2011-413- G00006) and (in part) by the Health Promotion Fund, Ministry of Health & Welfare, Republic of Korea (2012-4).

1. INTRODUCTION

Statistics Canada conducts the Canada Internet Use Survey (CIUS) every other year under the ad- vice of the Organisation for Economic Co-oper- ation and Development (OECD). Since 2000, the United States (US) has been conducting a survey called “Pew Internet & American Life Project”

through the private research laboratory Pew Research Center to study the American citizens’

Internet use and its effects. In addition, Pew Research Center has been carrying out “tracking survey” numerous times a year to obtain general information regarding their use of the Internet. In Korea, the Korea Internet & Security Agency (KISA) also conducted home interview surveys through a stratified multi-stage cluster sampling

method and used the ordinary household enumera- tion districts in the 2005 Population and Housing Census by the National Statistical Office (NSO) as the sample frame, extracting 30,000 households throughout Korea and 72,559 family members thereof. From 2004 onwards, KISA has expanded the scope of research subjects from 3 years to older individuals by including wireless mobile phones in their survey on the Internet use.

With the launching of the high-speed Internet in 1998 following the opening of Internet commer- cial services in 1994, the national Internet usage rate reached 70.2% with 3,158 Internet users as of 2004. Since then, the Internet usage rates were 72.8% in 2005, 76.5% in 2008, and 77.8% in 2010, revealing an increase rate to less than 1% [1].

According to KISA [1], as of July 2011, 78% of a

population of more than 3 years of age used the Internet. Approximately 98.7% of the Internet users were online at least once a week, and 85.2% at least once a day. By household, 81.9% had computers, and 81.8% of the total households had Internet ac- cess at home. About 42.9% of the total households had smart devices with Internet connection.

In order to fully understand the social effect of the use of the Internet, which has a significant im- pact on our daily lives, a systematic analysis on the use of Internet should be conducted. Addition- ally, with the advent of various types of Internet communities and media materials, the influence of the Internet has been expanding exponentially throughout our society. Nevertheless, research providing the general characteristics of Internet use or forecasting its future direction is lacking [2].

Even in the field of healthcare, the efficient use of the Internet has become an issue, especially now that a social paradigm shift between patients and their healthcare service providers has transpired [3,4].

Research aiming to analyze the status of using online health information has been conducted in the literature [5-8]. Healthcare consumers no longer regard online information or related services as a simple act of using an application, but recognize it as a medium that affects their daily life. They are advancing from passive healthcare consumers to active users of online health information, partic- ular using such information for autonomous deci- sion-making. Therefore, it is the responsibility of healthcare professionals to set a very solid founda- tion regarding consumers’ use of online health information. In order to obtain a positive outcome from the use of health information via the Internet and mobile phones, the starting point should be scientifically analyzing previous research on the field of interest.

Due to exponentially increasing the number of journal articles, it is not simple to grasp the associ- ations and or relationships among research themes

of the articles [9-11]. Therefore, researchers in the fields of sociology and psychology have developed an interdisciplinary method of social network anal- ysis (SNA). This method has been widely used in academia [5], which is an excellent way of visual- izing links between research themes by converting human relation networks into mathematical models [12-14].

Analysis on co-occurring keywords is one of the methods to visualize the cognitive structures through SNA; it is an appropriate method to ex- press actual contents of the outcomes of research by linking extracted keywords from the research papers [10,14]. Co-occurring keywords in research as well as research purposes and strategies pre- sented by the authors in their papers reflect se- mantic relations, which can be considered as a contents analysis technique [15].

Recently, SNA using keywords extracted from published papers has been applied to the field of healthcare to understand the whole picture of the relationships constituting the cognitive structures of the research themes. For instance, results of network analyses using extracted keywords from the field of preventive medicine [16], epidemiology [17], medical informatics [18], and nursing [19]

have been reported in the literature. In addition, there are a few studies employing SNA by the use of extracted medical subject headings (MeSH), a hierarchically structured medical terminology and uniformly designated by professionals to index pa- pers [13,20-23], which are controlled and main- tained by the U.S. National Library of Medicine (NLM) at the National Institutes of Health (NIH).

The purpose of this study, therefore, were to (1)

analyze MeSH keywords of academically accumu-

lated empirical studies on the use of health in-

formation via the Internet and mobile phones by

employing SNA of co-occurring keywords, and (2)

provide basic data to present directions for future

research on the field of interest. For the analysis,

the concepts of weighted degree centrality and

997 Social network analysis on consumers’ seeking behavior of health information via the Internet and mobile phones

standardized degree centrality were considered.

From this study, it was also expected that the po- tential of applying a social network analysis to the field of healthcare would be presented [22].

2. METHOD

2.1 Data Collection

The MeSH keywords from the PubMed data- base, which is a free database accessing primarily the MEDLINE database of references and ab- stracts on life sciences and biomedical topics, were extracted from the papers necessary for analyzing the use of online health information via the Internet and mobile phones. In order to select the targeted papers for the analysis, search terms with search conditions were applied. After excluding over- lapping papers, keywords of the pre-selected pa- pers were reviewed.

2.1.1 Paper Extraction

A total of 1,229 initially extracted papers regard- ing the use of Internet health information were ex- tracted with the MeSH search terms “Internet”

AND “Consumer Health Information,” “Internet/

utilization” AND “Consumer Health Information,”

and “Internet/utilization” AND “Health Education,”

which yielded 544, 105, and 582 search results, respectively. Finally, 1,001 papers were selected after excluding the overlapping ones. For mobile health information, the MeSH search terms were

“Cellular P hone” AND “Consumer Health Infor- mation” and “Cellular P hone” AND “Health Education.” A total of 98 papers were finally se- lected after excluding the overlapping ones. Since a large number of studies regarding consumers’

use of Internet and mobile health information were conducted along with health education; therefore,

“Health Education” was also included in the search terms as conditions.

2.1.2 Keyword Extraction and Selection of Keywords The extracted keywords from the 1,001 papers

regarding Internet health information were 1,213 in total (accumulated keywords: 8,021). After exclud- ing the overlapping keywords from the five cate- gories of the three-year brackets (before 2000, 2000 –2002, 2003–2005, 2006–2008, and 2009–2011), finally 155 (weighted frequencies: 3,780) keywords were included for the analysis.

For mobile health information, a total of 926 keywords were extracted from the 98 papers re- garding to the use of mobile health information.

After excluding the overlapping keywords, 326 were finally included in this study. To observe the changes in the categories containing these words, the keywords occurring before 2005 that showed very low frequency were grouped as one category.

Therefore, a total of three categories were set in- cluding the 2006–2008 and 2009–2011 brackets.

After excluding the overlapping keywords, 98 key- words (weighted frequencies: 636) were finally in- cluded in this study.

When previous studies in the field of interest are taken into consideration, any change significantly noticed within a certain time frame may be ob- served in the literature. Therefore, it is helpful for researchers to examine the period of “turning point” by analyzing the time sequential appearance of keywords for future studies [23,24].

2.2 Data Analysis

To conduct SNA using the selected keywords, NetMiner v.3 (CYRAM, Korea) was used. Key- words were referred to as nodes in SNA; in case of a same keyword appearing in different papers, a link was considered to exist between these key- words as building a keyword matrix.

If a network is established using all the selected

keywords, it is difficult to identify the links be-

tween these keywords. Thus, the degree and the

density of the links were applied to SNA. The more

links between nodes, the more resources are yield-

ed [25]. Density is defined as the ratio of actual

links to the total possible links; however, it is in-

Fig. 1. Keywords regarding health information via the Internet.

versely proportional to the size of a group.

Therefore, SNA should be conducted after stand- ardizing the size of the group because even if a node has the same number of links, the density de- creases as the group size increases [25,26].

Moreover, the impact coming from the network degree of the nodes can be measured by looking at the center of the network, which is referred to as a degree centrality. It means that a node with several direct links is considered to have a high degree centrality [25,27]. More importantly, a weighted degree centrality index is the connection strength of a central role within the network [28,29]. When weighted inter-node connection strength is strong, it can be placed in the center of a network due to an increased centrality.

However, even if nodes have the same number of links, densities are different depending on the size of the group.

To observe the changes in the core keywords, which was the purpose of this study, the weighted degree centrality index should be used to establish networks [30]. Therefore, in this study, a pruning method was applied to determine the boundary of the networks, which restructures core keywords with a high degree centrality value [13,31] and with a cut off value as a standard [10,24]. In this case, some information loss may occur, but the degree centrality among core keywords showing strong connections can be observed [13]. The cut off value of the pruning method was set at the value of 0.6% as the maximum degree centrality of a keyword.

Finally, two total degree centrality networks, five category-based networks, and three category- based networks were presented in this study. By observing the change of the keywords in each cat- egory, this study was able to track the time-de- pendent changes in the core keywords analyzed in this study.

3. RESULTS

3.1 Analysis of Using Health Information via the Internet

3.1.1 Network Analysis of the Total Keywords In the literature, keywords regarding health in- formation via the Internet started appearing in 1998. The frequency of occurrence has steadily in- creased every year with a rapidly increase ob- served since 2007 (Fig. 1).

According to the degrees among the keywords shown in the networks of 155 keywords (Fig. 2),

“Patient Education as Topic” had the highest de- gree centrality and showed a strong degree of con- nection with “Health Education,” “Information Services,” “Health Knowledge, Attitudes, Practice,”

“Self Care,” and “P atient Participation.”

According to the 2012 MeSH Browser main- tained by NLM, “Patient Education as Topic”

means “the teaching or training of patients con- cerning their own health needs.” “Health Educa- tion” related keywords were linked with “Health Behavior,” “Consumer Satisfaction,” “Health Promotion,” “Information Dissemination,” and

“Attitude towards Health,” which can be regarded as positives outcomes of health education. “Infor- mation Services” was linked to keywords such as

“Attitude towards Computers” and “Information Storage and Retrieval.”

Studies using the computer as a parameter, such

999 Social network analysis on consumers’ seeking behavior of health information via the Internet and mobile phones

Fig. 2. Total keyword network of health information via the Internet (pruning=35, n=31).

as “User-Computer Interface,” “Computer-Assisted Instruction,” and “Computer Literacy,” as well as studies dealing with communication with health- care professionals, such as “Communication” and

“P hysician-P atient Relations,” were connected in the case of providing health information for pa- tients or health consumers.

For patient education, according to the hier- archical structure of MeSH, carcinoma appeared to be the core keyword in the network of neoplasms.

Keywords regarding age, sex, socio-economic fac- tors, and educational status also appeared which are known to affect consumer behavior towards health information.

3.1.2 Category-based Network Analysis

When examining the core keywords regarding health information via the Internet, “Patient Education as Topic” and “Health Education” ap- peared to be the core research topics in every cat- egory (Table 1). It was also reflected in this study

that as consumers started to engage in deci- sion-making regarding the management of their own health. Consumers wanted to be involved in, for example, “Self Care,” “Self-Help Groups,” and

“P atient P articipation.” Therefore, keywords re- garding online “Social Support” function appeared to be one of the important core keywords of the related studies. One of the keywords most fre- quently searched by consumers regarding actual diseases were “Women’s Health” and “Neo- plasms.” In case of “Neoplasms,” its degree cen- trality increased over time, which indicates a con- stant and active search by consumers.

In examining the keywords by the category, 28 keywords out of 155 appeared in the category of

‘before 2000’ (Fig. 3). “Information Services”

seemed to be the keyword with the strongest cen- trality, reflecting the social change in the online society. Furthermore, the keywords such as

“Marketing of Health Services,” “Commerce,” and

“Remote Consultation” were indicative of the fact

Table 1. Health information via the Internet and the keywords listed in top 20 by category

Rank -2000 2000-2002 2003-2005 2006-2008 2009-2011

1 Information

Services Health Education Patient Education

as Topic Patient Education

as Topic Patient Education as Topic

2 Health Education Patient Education

as Topic Information

Services Health Knowledge,

Attitudes, Practice Health Knowledge, Attitudes, Practice 3 Patient Education

as Topic Information

Services Health Education Health Education Social Support 4 Marketing of Health

Services Patient Participation Health Knowledge, Attitudes, Practice Information

Services Health Education 5 Consumer

Participation Attitude to

Computers Socioeconomic

Factors Attitude to

Computers Information Dissemination 6 Data Collection Attitude to Health Information Storage and Retrieval Information

Dissemination Health Literacy 7 Physician-Patient

Relations Physician-Patient

Relations Attitude to Health Physician-Patient Relations Information Seeking Behavior

8 Patient Participation Consumer Participation Age Factors Information Storage

and Retrieval Self Care 9 Quality Control Health Knowledge,

Attitudes, Practice Attitude to

Computers Attitude to Health Attitude to Health 10 Consumer

Satisfaction Decision Making Educational Status Age Factors Age Factors 11 Commerce Communication Patient Participation Patient Acceptance of Health Care Health Behavior 12 Medical Informatics Self Care Sex Factors Computer Literacy Health Promotion 13 Delivery of Health

Care Social Support Health Care

Surveys Communication Socioeconomic Factors 14 Advertising as

Topic User-Computer

Interface Physician-Patient

Relations Educational Status Neoplasms 15 Communication Age Factors Medical Informatics Neoplasms User-Computer

Interface 16 Remote

Consultation Confidentiality Information

Dissemination Patient Participation Communication 17 Attitude to

Computers Health Services

Accessibility Neoplasms Social Support Physician-Patient Relations 18 Confidentiality Consumer

Satisfaction User-Computer

Interface Socioeconomic

Factors Self-Help Groups 19 Physicians Information Storage

and Retrieval Women’s Health Health Care

Surveys Patient Satisfaction 20 Professional-Patien

t Relations Self-Help Groups Computer-Assisted

Instruction Medical Informatics Information Services

that health information via the Internet was per- ceived as one of the commercial means actively used by healthcare service providers rather than consumers. Research conducted during this period showed that “Delivery of Health Care” was related to “Consumer Participation” and “Consumer

Satisfaction.” Additionally, “Quality Control” of health information was seen as an example of a conscious effort to verify the credibility of health information via the Internet.

During 2000–2002, newly emerging research

topics such as “Attitude to Health” and “Health

1001 Social network analysis on consumers’ seeking behavior of health information via the Internet and mobile phones

Fig. 3. Keyword network of before 2000 (pruning=0, n=28).

Fig. 4. Keyword network of 2000–2002 (pruning=5, n=33) Services Accessibility,” “Decision-Making” for

self-care of consumers as well as “Self-Care” and

“Self-Help Groups” were appeared (Fig. 4). “Con-

sumer Participation” and “Patient Participation”

were noted as well. Studies on online “Social

Support” function, which was related to “Adapta-

Fig. 5. Keyword network of 2003–2005 (pruning=9, n=23).

tion, Psychological,” were initiated in this period.

“Self-Care” and “Social Support” are high-ranking keywords in the network even to this date. The health-education-related keywords “Adolescent Behavior” and quality-related keywords “Quality Assurance, Health Care” also appeared in the network.

In the period of 2003–2005, a network regarding

“Patient Education as Topic” related to socio-dem- ographic characteristics were expanded along with the increase of centrality of “Health Knowledge, Attitudes, Practice” (Fig. 5). “Information Storage and Retrieval” and “Information Dissemination”

also appeared likewise. Keywords regarding dis- eases or illnesses were “Neoplasms” and “Women’s Health.” In case of “Neoplasms,” the degree cen- trality increased over time, showing actively stud- ied on the topic.

During 2006–2008, the centrality of “Health Knowledge, Attitudes, Practice” and “Information Dissemination” increased (Fig. 6). The keyword of

“Computer Literacy” newly appeared; it was be- cause understanding computers was importantly

considered to search, locate, and utilize online health information. An independently identified network of the keywords was related to “Smoking”

and “Smoking Cessation.”

In the period of 2009–2011, the centrality of

“Health Education” relatively decreased, while that of “Social Support” increased (Fig. 7). Newly ap- peared keywords in the network were “Information Seeking Behavior” and “Health Literacy,” which were related to high-ranking core keywords such as “Patient Education as Topic.” Moreover, key- words related to a specific illness included

“Chronic Disease.” The genetic-related keywords of “Genetic Testing” and “Genome, Human” also newly appeared and thus expanded the scope of the network.

3.2 Analysis of Using Health Information via Mobile Phones

3.2.1 Network Analysis of Total Keywords

In the literature, keywords regarding health in-

formation via mobile phones started appearing in

1003 Social network analysis on consumers’ seeking behavior of health information via the Internet and mobile phones

Fig. 6. Keyword network of 2006–2008 (pruning=11, n=31).

Fig. 7. Keyword network of 2009–2011 (pruning=17, n=33).

2002. While a constant increase in keywords was observed from 2005 to 2010, the drastic and rapid increase occurred thereafter (Fig. 8).

As examining the network including the 98 key- words, which structured the 43 networks with the high centrality, the degree centrality of “Cellular

Phone” ranked the highest, followed by “Patient

Education as Topic,” “Internet,” “Health Educa-

tion,” then “Self Care” (Fig. 9). From this network,

it was found that using health information via mo-

bile phones brought along the simultaneous use of

health information via the Internet. In addition, the

Fig. 8. Keywords regarding health information via mobile phones.

Fig. 9. Total keyword network regarding health information via mobile phones (pruning=10, n=43).

purpose of searching and locating health in- formation via mobile phones was to provide in- formation regarding “Self Care.” It seemed to have originated from the “Social Support” aspect.

Mainly provided health information via mobile phones was about disease-related information services and specific disorders such as “Diabetes

Mellitus, Type 2,” “Obesity,” “HIV Infection,”

“Asthma,” “Depression,” and “Anxiety.” Ongoing studies were regarding risk factor management to improve the overall health status, including studies on “Smoking Cessation,” “Exercise,” and “Medi- cation Adherence.”

3.2.2 Category-based Network Analysis

For the high-ranking keywords listed in the top 20, a mobile phone was used to obtain health in- formation related to “Diabetes Mellitus, Type 2”

and “Blood Glucose Self-Monitoring.” Keywords related to “HIV Infections” appeared recently al- though they were not high-ranking keywords. It was also noticed that research on “Telemedicine”

was more extensively conducted (Table 2).

Specific keywords in nursing, such as “Nursing

Evaluation Research” and “Nurse’s Role,” were

studied along with the use of health information

1005 Social network analysis on consumers’ seeking behavior of health information via the Internet and mobile phones

Table 2 List of top 20 degree centrality keywords by category

rank 2002-2005 2006-2008 2009-2011

1 Cellular Phone Cellular Phone Cellular Phone

2 Patient Education as Topic Patient Education as Topic Patient Education as Topic

3 Diabetes Mellitus, Type 2 Internet Internet

4 Internet Diabetes Mellitus, Type 2 Self Care

5 Blood Glucose Blood Glucose Health Education

6 Computer-Assisted Instruction Hemoglobin A, Glycosylated Diabetes Mellitus, Type 2

7 Fasting Nursing Evaluation Research Health Promotion

8 Hemoglobin A, Glycosylated Nurse’s Role Health Knowledge, Attitudes, Practice 9 Follow-Up Studies Computer-Assisted Instruction Diabetes Mellitus

10 Patient Satisfaction Blood Glucose Self-Monitoring Patient Compliance

11 Self Care Program Evaluation Blood Glucose Self-Monitoring

12 Program Evaluation Fasting Electronic Mail

13 Health Education Postprandial Period Telemedicine

14 Treatment Outcome Telemedicine Patient Satisfaction

15 Risk Factors Self Care Reminder Systems

16 Patient Satisfaction Information Dissemination

17 Pilot Projects Pilot Projects

18 Communication Hemoglobin A, Glycosylated

19 Feasibility Studies HIV Infections

20 Social Support Exercise

Fig. 10. Keyword network of 2002–2005 (pruning=0, n=15).

via mobile phones. Information on the role of nurses as a mediator as well as an advocate of health used to be provided in clinical settings in the past; however, with the use of the Internet and mobile phones, this result showed that consumers turned to be very active in the management of their health and empirically demonstrated the need for a role shift among nurses with regard to providing health information. It was also noted that studies regarding the outcomes of such phenomenon were intensively conducted.

Although not listed in the high rank, there was an increase in the provision of information through

“Text Message.” Compared with the previous years, many keywords newly appeared during the period of 2009–2011, which reflected a rapid in- crease of keywords on the use of health in- formation via mobile phones.

For the keyword networks by each category, 15 keywords appeared during the period of 2002–2005

(Fig. 10). Studies on mobile-based education in- cluding “Computer-Assisted Instruction” as well as information on “Diabetes Mellitus, Type 2” and

“Self Care,” “Patient Satisfaction,” and “P rogram Evaluation” for such diseases, were conducted.

During the period of 2006–2008, 52 keywords

Fig. 11. Keyword network of 2006–2008 (pruning=5, n=32).

appeared in the network (Fig. 11). The core key- words of “Diabetes Mellitus, Type 2,” “Obesity”

and “HIV Infections” were noticed. Consequently, a gradual expansion of the scope and the topics re- garding health information via mobile phones was noticed. Studies on “Patient Compliance” and pro- viding health information via “Electronic Mail” al- so started to appear in the literature.

During 2009–2011, 76 keywords appeared in the network (Fig. 12). The degree centrality of “Health Education” and “Self Care” increased. It was no- ticed that information on “Obesity” and “HIV Infection” was disseminated via mobile phones.

For the information used to be provided via “Elec- tronic Mail,” alternative communication channels such as “'Telemedicine” or “Text Message” start-

ed being adopted. Although not being presented in the network, the use of health information regard- ing “Asthma,” “Depression,” “Anxiety,” and “Smoking Cessation” were actively studied. “P atient Com- pliance” regarding “Health Knowledge, Attitudes, Practice” also appeared as important keywords.

4. DISCUSSION

This study attempted to grasp the systematic picture of research topics by extracting core key- words in the field of interest. The centrality meas- ure for SNA has a downside of not reflecting the weighted degree of connection between nodes [30].

Therefore, in order to grasp the centrality of re-

search topics, this study applied weight not only

1007 Social network analysis on consumers’ seeking behavior of health information via the Internet and mobile phones

Fig. 12. Keyword network of 2009–2011 (pruning=7, n=32)

to the number of linked keywords but also to the frequency of the links. In this manner, the degree centrality increases with increased weight [29,32].

However, as such weighting does not completely reflect the quality of a research topic; thus, it should be considered to apply relative weight for future research. Additionally, because it is not pos- sible to display all of the keywords extracted from the literature, only rather frequently appearing keywords above a certain level of values were in- cluded by the pruning method. Considering these limitations of SNA in this study, attention should be paid to the presence of the keywords having a high degree of centrality in the networks.

In the keyword networks related to health in- formation via the Internet, the research topic of

“P atient Education as Topic” had the strongest centrality. As it was noticeable in the network,

“Patient Education as Topic” was linked to “Age”

and “Sex” as well as other core keywords.

Researchers in the past actually attempted to iden-

tify influential factors affecting a Technology Acceptance Model (TAM). Based on their findings, an Internet-related Information and Communica- tion Technology Acceptance Model (ICTAM) was devised and proposed [3,4]. In the study demon- strating the validity of ICTAM, An and her col- leagues [3,4] reported that gender differences had a significant impact on the acceptance of Internet- related information and communication technologies.

To date, there are still insufficient amount of theo- retical and empirical studies regarding ICTAM by age. Therefore, socio-demographic factors and the inherent information gap should be further studied in interpreting the results of research on the use of online health information.

In the studies on the ensuing use of health in-

formation via mobile phones, consumer behaviors

were extensively studied with Internet-related

research. However, the keywords shown in the

networks were very different. In addition, the ab-

solute number of the papers on mobile related

health information was certainly not comparable with that on Internet related health information. In case of health information via mobile phones, the keywords extracted from environmental diseases or socio-psychological perspective such as

“Asthma,” “Depression,” and “Anxiety” as well as

“Smoking Cessation,” “Diabetes Mellitus, Type 2,” and “Chronic Disease” appeared in the net- works, indicating consumers used essential health information for the management of their health without delay, whenever necessary. These key- words recently and rapidly appeared in the literature. It was also noticed that the age groups using mobile phones to access health information were expanded from teenagers and young adults to 40 years and older. This trend accelerated stead- ily in the literature. In addition, cutting-edge medi- cal information, such as genetic related information mobile, could be accessible via mobile phones.

The keywords regarding health information via the Internet started to appear in 1998 and increased rapidly after 2007. In case of keywords regarding health information via mobile phones, it started to appear in 2002 and increased rapidly after 2010.

Diffusion of ‘so-called’ innovation of both cases took about 10 years in the academia. Rogers [33]

defined innovation as “an idea, practice, or object that is perceived as new by an individual or other unit of adoption” when discussing diffusion of innovations. With the wide use of smart phones, its convergence contents were dramatically diffused. The early adopters always are using an upgraded level of “innovation” and spreading it at the same time [34]. Therefore, studies regarding consumers’ use of health information can be pre- dicted on smart phone-based digital convergence contents.

Healthcare consumers no longer regard online information or related services as a simple act of using an application. Instead, they recognize it as a medium or channel for communication. In con- sideration of the rapid changes in the information

society of the 21

century, healthcare consumers are turning to very active users using online health information for their autonomous decision-making.

Therefore, it is the responsibility of healthcare pro- fessionals to set an evidence-based solid founda- tion regarding online health information. In order to obtain a positive outcome from the use of online health information, the starting point should be based on scientifically studied empirical research on consumers’ use of online health information.

The results of this study, therefore, can be pro- vided basic information to present directions for future research on this area.

Recommendations were proposed based on the results of this study. First, methods for future re- search should be designed in consideration of in- formation gap and speed of information diffusion, which are dependent on socio-demographic factors of health consumers. In interpreting the results, in- herent information gap should be also considered.

Second, analyzing the time sequential trend of the

increasing number of Internet users in Korea, it

seems that the increase will slow down. However,

along with the expanding social communications

such as emails, messengers, blogs, and social net-

working services and in tune with the current trend

of digital media convergence contents, studies on

the use of smart phone-based health information

by consumers seem to be necessary. Furthermore,

consumers also should be educated on how to ap-

propriately use the obtained health information for

the improvement of their health. A balanced view-

point on information technology, culture, and the

society is also required. Third, when conducting

SNA, it is suggested that applying weight to the

MeSH keywords should be considered [35]. In ad-

dition, to utilize the results of SNA as a basis for

policy making or designing future research, it is

recommended that an in-depth expert review re-

garding its results should be conducted. Lastly, a

previously proposed framework for the visual-

ization of networks could be helpful for readers to

1009 Social network analysis on consumers’ seeking behavior of health information via the Internet and mobile phones

understand the results of the networks. In addition, this approach may be an emerging research theme in information design [36].

5. CONCLUSION

This study attempted to grasp the time sequen- tial appearance of research topics in the field of in- terest by extracting core keywords. Based on the analysis, access to health information via the Internet was made prior to information access via mobile phones. Studies on health information via the Internet seemed to be in a settled stage when it comes to investigating the influence of various socio-demographic factors and the characteristics of users (i.e., consumer behaviors) on finding in- formation for the maintenance and improvement of self-care. In the studies on the ensuing use of health information via mobile phones, consumer behaviors have been extensively studied via Internet-related research. Many keywords not ap- peared in the Internet-related studies appeared in the mobile-related studies even in the initial stage of research. The results of this study using SNA provided basic information to guide research direc- tions for future studies in the field of interest. In this regard, an in-depth review (i.e., an expert re- view) of these results should be performed to es- tablish the objectivity and the validity of this study.

REFERENCE

[ 1 ] Korea Internet & Security Agency, 2011 Internet Usage Survey, 2011.

[ 2 ] J.S. Hwang and J.Y. Yoo, “A Study on Changing Trends and Causes of Internet Usage in Everyday Life–comparison between 2002 and 2005,” Informatization P olicy, Vol.

14, Issue 1, pp. 19-43, 2011.

[ 3 ] J.Y. An, “Theory Development in Health Care Informatics: Information and Communication Technology Acceptance Model (ICTAM) Improves the Explanatory and Predictive

Power of Technology Acceptance Models,”

Studies in Health Technology and Informatics, Vol. 122, pp. 63-67, 2006.

[ 4 ] J.Y. An, L.L. Hayman, T.L. Panniers, and B.

Carty, “Theory Development in Nursing and Healthcare Informatics: A Model Explaining and Predicting Information and Communica- tion Technology Acceptance by Healthcare Consumers,” Advances in Nursing Science, Vol. 30, Issue 3, pp. E37-49, 2007.

[ 5 ] B.S. Jung and Y.K. Kwon, “A Study on the Knowledge Map of the Computer Engineering Field by using A Research Paper Database,”

Annual Proceeding of Korean Information Processing Society, Vol. 18, Issue 1, pp. 1460- 1462, 2011.

[ 6 ] H.S. Lee, K.S. Lee, M.S. Kim, S.W. Whang, D.S. Kim, J.W. Woo, and et al., “A Study on Health Information and Medical Consulting via Internet Focusing on the Age Group of 20s,” J ournal of Digital P olicy & Manage- ment, Vol. 10, Issue 2, pp. 255-267, 2012.

[ 7 ] J.H. Park, J.S. Lee, H. Jang, and Y. Kim,

“Factors Affecting Consumer's Usage of Health Information on the Internet,” J ournal of P reventive Medicine, Vol. 41, Issue 4, pp.

241-248, 2008.

[ 8 ] S.W. Ryu and Y.J. HA, “Usage of Health Information on the Internet,” Health and Welfare Policy Forum, Vol. 97, pp. 71-87, 2004.

[ 9 ] Q. He, “Knowledge Discovery Through Co-word Analysis,” Library Trends, Vol. 48, Issue 1, pp. 133-159, 1999.

[10] K.K. Mane and K. Borner, “Mapping Topics and Topic Bursts in PNAS,” P roceedings of the National Academy of Sciences, Vol. 101, pp. 5287-5290, 2004.

[11] E.V. Bernstam, J.R. Herskovic, Y. Aphinya-

naphongs, C.F. Aliferis, M.G. Sriram, and

W.R. Hersh, “Using Citation Data to Improve

Retrieval from MEDLINE,” J ournal of

American Medical Informatics Association,

Vol. 13, Issue 1, pp. 96-105, 2006.

[12] X. Polanco, “Co-word Analysis Revisited:

Modeling Co-word Cluster in Terms of Graph Theory,” Proceeding of ISSI, Vol. 2, pp. 662- 663, 2005.

[13] D.W. Sohn, Social Network Analysi, Kyung- mun Publsher, Seoul, 2010.

[14] J. Zhang, J. Xie, W. Hou, X. Tu, J. Xu, F. Song, and et al., “Mapping the Knowledge Structure of Research on Patient Adherence: Knowledge Domain Visualization Based Co-word Analysis and Social Network Analysis,” PLoS One, Vol. 7, Issue 4, pp. e34497, 2012.

[15] X. Wang, J. Wang, F. Ma, and C. Hu, “The

“Small-world” Characteristic of Author Co- words Network,” P roceeding of 2007 Inter- national Conference on Wireless Communi- cations, Networking and Mobile Computing, pp. 3717-3720, 2007.

[16] M.S. Jung and D.J. Chung, “Co-author Key- word Network and Their Clustering Appearance in Preventive Medicine Fields in Korea:

Analysis of Paper in the Journal of Preventive Medicine and Public Health,” J ournal of P reventive Medicine & P ublic Health, Vol.

41, Issue 1, pp. 1-9, 2008.

[17] M.S. Jung, “Academic Research Activities and Their Co-author and Keyword Network in Epidemiology Fields: Analysis of Papers in Korean Journal of Epidemiology,” J ournal of Korean Epidemiology, Vol. 30, Issue 1, pp.

60-72, 2008.

[18] S. Jeong, S.K. Lee, and H.G. Kim, “Knowledge Structure of Korean Medical Informatics: A Social Network Analysis of Articles in Journal and Proceedings,” Healthcare Informatics Research, Vol. 16, Issue 1, pp. 52-59, 2010.

[19] S.K. Lee, S. Jeong, H.G. Kim, and Y.H. Yeom,

“A Social Network Analysis of Research Topics in Korean Nursing Science,” J ournal of Korean Academy of Nursing, Vol. 41, Issue 5, pp. 623-632, 2011.

[20] A. Albert, B. Granadino, and L.M. Plaza,

“Scientific and Technological Performance Evaluation of the Spanish Council for Scientific Research in the Field of Biotechnology,”

Scientometrics, Vol. 70, Issue 1, pp. 41-51, 2007.

[21] X.Y. An and Q.Q. Wu, “Co-word Analysis of the Trends in Stem Cells Field based on Subject Heading Weighting,” Scientometrics, Vol. 88, Issue 1, pp. 133-144, 2011.

[22] H.R. Jang, Y.S. Lee, and J.Y. An, “Application of Social Network Analysis to Health Care Sectors,” Healthcare Informatics Research, Vol. 18, Issue 1, pp. 44-56, 2012.

[23] H.C. Zheng, L. Yan, L. Cui, Y.F. Guan, and Y. Takano, “Mapping the History and Current Situation of Research on John Cunningham Virus-a Bibliometric Analysis,” BMC Infec- tious Disease, Vol. 9, Issue 28, pp. 1-12, 2009.

[24] C. Chen, “Searching for Intellectual Turning Points: Progressive Knowledge Domain Visualization,” PNAS, Vol. 101, Issue Suppl.

1, pp. 5303-5310, 2004.

[25] Y.H. Kim, Social Network Analysis, Pak- youngsa, Seoul, 2007.

[26] K. Faust, “Comparing Social Networks: Size, Density, and Local Structure,” Metodoloskiz- vezki, Vol. 3, Issue 2, pp. 185-216, 2006.

[27] S.P. Bogatti, “Power and Network Flow,”

Social Networks, Vol. 27, Issue 1, pp. 55-71, 2005.

[28] L.C. Freeman, “Centrality in Social Networks:

Conceptual Clarification,” Social Networks, Vol. 1, Issue 3, pp. 215-239, 1979.

[29] T. Osphal, F. Agneessens, and J. Skvoretz,

“Node Centrality in Weighted Networks:

Generalizing Degree and Shortest Paths,”

Social Network, Vol. 32, Issue 3, pp. 245-251, 2010.

[30] J.Y. Lee, “Centrality Measure for Bibliomatric

Network Analysis,” Journal of Korean Society

for Information Management, Vol. 40, Issue

3, pp. 191-214, 2006.

1011 Social network analysis on consumers’ seeking behavior of health information via the Internet and mobile phones

[31] L.A. Adamic, R.M. Lukose, A.R. Puniyani, and B.A. Huberman, “Search in Power-law Network,” Physical Review, Vol. 64, Issue 4, pp. E0461351-8, 2001.

[32] X.Y. An and Q.Q. Wu, “Co-word Analysis of the Trends in Stem Cells Field based on Subject Heading Weighting,” Scientometrics, Vol. 88, Issue 1, pp. 133-144, 2011.

[33] E.M. Rogers, Diffusion of Innovations, Free Press: New York, NY., 1983.

[34] J.C. Fisher and R.H. Pry, “A Simple Substitution Model of Technological Change,”

Technological Forecasting & Social Change, Vol. 3, Issue 1, pp. 75-88, 1971.

[35] A. Barrat, M. Barthelemy, R. Pastor-Satorras, and A. Vespignani, “The Architecture of Complex Weighted Network,” PNAS, Vol.

101, Issue 11, pp. 3747-3752, 2004.

[36] S.H. Jang and S.H. Jang, “A Framework for Visualizing Social Network Influence,” Journal of Korea Multimedia Society, Vol. 12, Issue 1, pp. 139-146, 2009.

Ji-Young An

She received her Ph.D. degree in Nursing (Healthcare Informa- tics) from New York University (NYU), New York, NY, USA in 2005. She received her M.P.H. in Public Health (Biostatistics) from Seoul National University, Seoul, Korea in 2000 and her B.S. in Nursing from Korea University, Seoul, Korea in 1998. Formerly, she was a full time faculty at NYU, New York, NY, USA and an assistant professor at Rutgers, The State University of New Jersey, Newark and New Bruns- wick, NJ, USA. She currently serves as a professor at the Design Institute of Inje University, Seoul, Korea.

Her research interests are consumers’ online health in- formation seeking behaviors, social network analysis, and big data analysis.

Haeran Jang

She received her Ph.D. degree in Medical Informatics and Mana- gement in 2012 and M.S. degree in Public Health Management in 1999 from Chungbuk National University, Cheongju, Korea.

She received her B.S. degree in Nursing from Seoul National University, Seoul, Korea in 1985. She currently works as Assistant Professor at Joongbu University, Chungnam, Korea. Her re- search interests are biomedical research trend analy- sis, MeSH Indexing, and nursing informatics.

Jinkyung Paik

She received her doctoral degree in Design from Sejong Univer- sity, Seoul, Korea in 2004. She received her M.F.A. in Graphic Design from the University of Michigan, Ann Arbor, MI, USA in 1986. She received her B.F.A.

in Visual Design from Seoul National University,

Seoul, Korea in 1982. Currently, she serves as a pro-

fessor at the College of Design, Kyeongnam, Korea

and a director of the Design Institute, Inje University,

Seoul, Korea. Her research interests are typography,

infographics, sign design, and healthcare design.