Journal of Health Communication, 20:285–296, 2015 Copyright # Taylor & Francis Group, LLC ISSN: 1081-0730 print/1087-0415 online DOI: 10.1080/10810730.2014.927035
Linking Health Information Seeking to Behavioral Outcomes: Antecedents and Outcomes of Childhood Vaccination Information Seeking in South Korea HYUN OU LEE1 and SOYOON KIM2 1
Department of Advertising and Public Relations, College of Communication and Social Science, Hanyang University, Ansan, Republic of Korea 2 Department of Communication Studies, School of Communication, University of Miami, Coral Gables, Florida, USA
Although research on health information has made significant progress in identifying the antecedents of individuals’ information-seeking behavior in the context of the United States, the results have not been generalizable to the contexts of many other countries. Moreover, little is known about how one’s information-seeking behavior is connected to actual behavioral outcomes relevant to the search action. The authors conducted an online survey with a stratified random sample of 1,004 mothers to examine the applicability of the comprehensive model of health information seeking in predicting the use of diverse childhood vaccination information sources in South Korea, and to investigate associations between the mothers’ engagement with specific vaccine information sources and behavioral intention to immunize their children. Findings indicated that the hierarchical structure and the role of predictors within the comprehensive model of health information seeking provided a valid framework in the context of vaccine information seeking in Korea. In addition, the authors found negative associations between the use of certain types of information sources and mothers’ intention to vaccinate. This suggests that the dissemination of critical health information through a variety of available sources does not automatically lead to prudent behavioral decisions when the specific characteristics of the different sources are not considered.
Research concerning health information has become an important branch of communication research, both because of its critical implications for public health outcomes (Niederdeppe, 2008) and because of the wealth of information sources currently available (Cline & Haynes, 2001; Dutta-Bergman, 2004). While a considerable body of research has examined various factors of health information seeking in diverse contexts (e.g., Cotton & Gupta, 2004; DeLorme, Huh & Reid, 2011; Rains, 2008) few studies have systematically examined health information-seeking behavior outside the United States, limiting the external validity of the findings. Moreover, while much attention in the health information-seeking literature has been focused on identifying who actively seeks health information, far less attention has been paid to how specific health information-seeking actions influence behavioral outcomes (Anker, Reinhart, & Feeley, 2011; Rains, 2007). We, thus, had two goals for the present study. First, to test the applicability of the prediction of health informationseeking behavior found in the United States, we examine various factors that influence mothers’ use of vaccine Address correspondence to Hyun Ou Lee, Department of Advertising and Public Relations, College of Communication and Social Science, Hanyang University, 55 Hanyangdaehakro, Sangnok-gu, Kyeonggi-do, 426-791, Ansan, Republic of Korea. E-mail: [email protected]
information sources with regard to their children in the context of South Korea. The comprehensive model of information seeking (CMIS; Johnson, 1997; Johnson & Meischke, 1993), which includes a wide range of factors associated with information-seeking actions, has provided the theoretical framework for approaching this problem. Second, we investigate how the mothers’ engagement with specific vaccine information sources is connected to their behavioral intention to have their children vaccinated. In its identification of the behavioral consequences as well as the predictors of vaccine information seeking in the South Korean context, the current study offers a theoretical contribution to the extension of an existing health information-seeking model and provides practical implications for the implementation of communication strategies to disseminate critical health information. Childhood Vaccination and Parental Concerns The success of national immunization programs throughout the twentieth century has contributed to the global eradication of smallpox, the elimination of the polio virus in the United States, and a substantial reduction in the occurrence of vaccine-preventable diseases, including measles, mumps, rubella, diphtheria, pertussis, tetanus, and Haemophilus influenza type b (Centers for Disease Control and Prevention, 1999, 2008). While variations exist in the success of
286 increasing vaccination coverage among young children across different countries (Kee et al., 2007; Zhao, Smith, & Luman, 2009), South Korea is among several countries that have made significant progress in controlling infectious diseases by implementing active vaccine surveillance systems. Since the vaccination was first introduced to Korea in the 1960s, most conventional vaccines have been included in the national immunization program, and the estimated coverage rates of recommended vaccines, including hepatitis B vaccine (HepB), diphtheria-tetanus-pertussis (DTP), and measles-mumps-rubella (MMR), have exceeded 90% in 2012 (Lu & Santosham, 2012; Shin et al., 2009). Despite universal efforts to increase vaccination coverage, a recent drop in vaccination coverage, especially for MMR, has been observed in several countries (Danis, Georgakopoulou, Stavrou, Laggas, & Panagiotopoulos, 2010; Serpell & Green, 2006). Scholars have indicated that parental concerns about vaccine safety, which have been reported as one of the most important barriers to childhood vaccination (Brown et al., 2010; Gust et al., 2004), largely derive from a lack of information or knowledge concerning childhood vaccination (Ledford, Willett, & Kreps, 2012; Ramsay, Yarwood, Lewis, Campbell, & White, 2002). The average parents are unlikely to have firsthand knowledge of the seriousness of vaccinepreventable disease, nor are they likely to be aware of the true risks of vaccine-adverse events (Serpell & Green, 2006). When uncertainty around a health issue prevails, people are primarily motivated to seek information to reduce that uncertainty (Babrow & Kline, 2000). Although the contemporary view1 of uncertainty management highlights the complex nature of uncertainty that results in diverse patterns of information behaviors, including information avoiding as well as seeking, seeking health information is regarded as an essential first step in coping with health-related uncertainty and health behaviors (Tardy & Hale, 1998; Wanzer, Booth-Butterfield, & Gruber, 2004). Prediction of Health (Vaccine) Information Seeking: CMIS Health information seeking is defined as the purposive acquisition of health information from selected sources for determining one’s own health behaviors (Johnson & Meischke, 1993; Johnson, Donohue, Atkin, & Johnson, 1995). The use of various health information sources is influenced by many factors, including individuals’ socio-demographics (Freimuth, 1990; Lenz, 1984; Zarcadoolas, Blanco, Boyer, & Pleasant, 2002), health-related motivations (Cummings, Becker, & Maile, 1980), and the perception of information sources (DeLorme, Huh, & Reid, 2011; Rains, 2007). A theoretical framework proposed by Johnson and Meischke (1993), the CMIS, provides a useful theoretical framework for investigating the effects of such variables related to health information seeking. The CMIS is an integration of three theoretical perspectives. First, the uses and gratification perspective 1
Brashers and colleagues pointed out that people are not always motivated to gain more information to decrease uncertainty surrounding a health issue, as they tend to adjust their information seeking to the level of uncertainty they can tolerate (Brashers, 2001; Brashers, Goldsmith, & Hsieh, 2002; Brashers, Neidig, & Goldsmith, 2004).
H. O. Lee and S. Kim (Blumler & Katz, 1974; Rubin, 1986) imparts to the CMIS the concept of active consumers and their goal-directed use of various media to fulfill their various needs. Second, drawing on the health belief model (Rosenstock, 1974; Rosenstock, Strecher & Becker, 1988), the CMIS incorporates health-related factors, such as health concerns and perceived susceptibility, as the driving forces behind information seeking. Last, the CMIS emphasizes the link between the evaluation and use of communication channels based on the model of media exposure and appraisal (Johnson, 1983). As a result, the CMIS includes three classes of variables: (1) antecedents to information seeking (i.e., background factors [demographics and personal experiences] and personal relevance factors [salience and efficacy belief]), which determine the perception of an information source; (2) information carrier factors (i.e., the characteristics and the utilities of the media), which, in turn, shape the specific intentions related to seeking information from a particular source; and (3) information-seeking actions, which are the outcomes of these classes of variables (Johnson, 1997). Since Johnson and Meischke (1993) demonstrated the superiority of information-carrier factors over background and personal relevance factors in anticipating the search for cancer information through magazines, the CMIS has been empirically tested in, or at least applied across, a variety of contexts in the United States (e.g., DeLorme, Huh, & Reid, 2011; Johnson, Andrews, & Allard, 2001; Rains, 2007). The findings of subsequent studies generally support the hierarchical structure of the CMIS in explaining information-seeking behaviors, but the specific patterns of relationships within the model have suggested some contingent effects (Case, Andrews, Johnson, & Allard, 2005). Also, an individual’s selection and use of sources of health information are dependent on their match with the seeker’s needs and expectations (Johnson, 1997). For example, while health care professionals are generally regarded as the most preferred sources of health information among patients or their family members (Andreassen, Randers, Naslund, Stockeld, & Mattiasson, 2005), friends and family are viewed as more effective than other sources when individuals need more tailored emotional support in obtaining complex and serious health information (Burleson & MacGeorge, 2002; Johnson, 1997). In addition to traditional mass media that plays an important role in providing detailed information about products and services related to health and illness, the Internet has become one of the most popular health information channels, primarily based on its accessibility (Cotton & Gupta, 2004; Johnson & Meischke, 1993; Nguyen & Bellamy, 2006). In particular, the Internet is one of the frequently used sources of health information in South Korea, as nearly 98% of households in the country have broadband connections (Organisation for Economic Co-operation and Development, 2010). A special medium also exists for childhood vaccine information in South Korea: the Child Pocket Book. A Child Pocket Book is a booklet issued by public health centers or private hospitals that enables parents to keep track of their children’s immunization schedules and records. Along with the Internet and interpersonal sources,
Linking Health Information Seeking to Behavior it is also one of the primary sources of vaccine information in South Korea (Korean Centers for Disease Control and Prevention, 2006). Given that the prediction of information-seeking behavior is specific to the type of disease and=or the cultural background of the seeker, it is necessary to ask whether the CMIS is a valid model for explaining information-seeking behavior across a variety of diseases, even outside the United States. Although the model of media exposure and appraisal, from which the CMIS’s hypothesized link between information carrier factors and information-seeking actions was drawn, has been tested in international settings (e.g., Johnson, 1983; Johnson & Oliveira, 1988), no research has systematically examined the CMIS framework as a whole to predict information-seeking behaviors related to childhood vaccination outside the United States, especially across various information channels. In light of this gap, the present study forwards the following research question to examine the general validity of the model for predicting the use of diverse sources of vaccine information, including (a) expert sources (i.e., health professionals); (b) nonexpert sources (i.e., family and friends); (c) entertainment-oriented media (i.e., television and radio); (d) information-oriented media (i.e., newspapers and magazines); (e) the Internet; and (f) the Child Pocket Book, in the South Korean context. Research Question 1: Is the CMIS an appropriate model to predict vaccine information-seeking actions in South Korea regardless of the types of sources examined?
In addition, we investigate the specific roles of demographic backgrounds, health-related motivational factors, and information carrier factors with following hypotheses that are drawn from relations among the variables in the CMIS. Hypotheses 1 through 6: The set of antecedents (i.e., demographic characteristics and health-related factors) and information carrier factors will influence health information seeking related to the extent of mothers’ use of childhood vaccine information from diverse sources, including expert sources (Hypothesis 1); nonexpert sources (Hypothesis 2); entertainment-oriented media (Hypothesis 3); information-oriented media (Hypothesis 4); the Internet (Hypothesis 5); and the Child Pocket Book (Hypothesis 6).
Outcomes of Information Seeking by Source As scholars have underscored the importance of obtaining health information as a first step in health behavior change (Freimuth, Stein, & Kean, 1989; Tardy & Hale, 1998), some information-seeking models, such as the Expanded Conceptual Model of Information Seeking Behaviors (Longo, 2005), assumed a positive link between obtaining health information and health-related decision-making processes and health outcomes. The CMIS also postulates that health-related behaviors are ultimately affected by the scope and depth of information-seeking actions. Johnson and
287 colleagues, however, indicated that the prediction of specific health-related behaviors is beyond the scope of the CMIS, as the model focuses on critical factors that influence information-seeking actions (Johnson, 1997; Johnson & Case, 2012). Notably, they highlighted potential negative outcomes of information-seeking actions (e.g., erroneous conclusions), stating that ‘‘It is not that people do not gather information or learn things, but often they gather the wrong information for the wrong reasons from the wrong source’’ (p. 133). Studies have continuously reported that not all sources of vaccine information affect childhood vaccination rates in the same way. While vaccine information obtained from expert sources or booklets provided by health care organizations can increase vaccine uptake (e.g., Gamertsfelder, Zimmerman & DeSensi, 1994; Korean Centers for Disease Control and Prevention, 2006), some widely used information sources have contributed significantly to increased parental concerns regarding childhood vaccination (Clarke, 2008; Cooper, Larson, & Katz, 2008). For example, mass media reports in the 1990s contributed to the distribution of the allegation of a possible link between MMR vaccine and autism, thus instigating the antivaccination movement (Ledford, Willett, & Kreps, 2012). Similarly, the advent of the Internet has increased access not only to accurate information, but also to misinformation about vaccines, thus providing additional channels of communication for antivaccine activists (Cooper, Larson, & Katz, 2008). As several scholars (Anker, Reinhart, & Feeley, 2011; Case, Andrews, Johnson, & Allard, 2005) have pointed out, most previous health information-seeking studies have identified the influencing factors of health informationseeking behavior; however, far fewer have focused on examining relevant outcomes associated with search actions. Thus, additional investigation regarding outcomes of health information seeking will be an important addition to current theoretical frameworks, including the CMIS. Accordingly, rather than viewing health information-seeking behaviors as a final outcome, we propose another research question to identify the effects of health information-seeking behaviors on the outcomes associated with actual health behaviors. Behavioral theories suggest that behavioral intention, which refers to one’s intention to perform a behavior, is the most direct indicator of an actual behavior (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975; Fishbein & Yzer, 2003). Given that the actual vaccination behavior will only be available from the same sample at a point in the future, we examine vaccination intention for children as the outcome of engagement with various vaccine information sources as the proxy of the performance of childhood vaccination. Research Question 2: How is the use of childhood vaccine information from nonexpert interpersonal sources, expert sources, entertainment-oriented media (TV=radio), information-oriented media (newspapers=magazines), the Internet, and the Child Pocket Book related to mothers’ behavioral intention to have their children vaccinated?
288 Method Data Collection A nationwide, cross-sectional web-survey of individuals’ perception on childhood vaccination was conducted through a professional research agency in South Korea during August 2010. The agency retains its own panel that consists of a stratified sample of 22,146 people from across the country based on 2005 Census data. To be eligible for participating in the survey, a woman had to have children younger than 12 years of age.2 Among the potential respondents who received the web-survey questionnaire, which was administered in the Korean language, only those who met the eligibility (n ¼ 2,459) and sent back complete survey responses (n ¼ 1,004) were included in the final data set. Mothers analyzed in this study were 23–59 (M ¼ 37.81, SD ¼ 1.49) years of age. The monthly household income of 48.2% of the sample ranged from US$2,500 to US$4,500, indicating that the profile of the study sample was comparable with national figures, as the average monthly household income in South Korea was about US$3,500 in 2010 (Korean Statistical Information Service, 2010). However, it was also found that those in the study sample were slightly better educated than the national population as a whole. Thus, caution should be exercised in generalizing the study’s results related to education level. Measures Antecedents We assessed respondents’ age, education level and monthly household income as indicators of demographics. The first personal relevance factor in the CMIS, salience, is operationalized as perceived severity, which refers to a person’s beliefs about the medical or social seriousness of a health threat or leaving it untreated (Janz & Becker, 1984; Rosenstock, Strecher, & Becker, 1994). Perceived severity of vaccinepreventable infectious disease was assessed by three items, including ‘‘Infectious diseases that may be caused by lack of childhood vaccination are very serious.’’ All items were scored on a five-point scale with anchors 1 (strongly disagree) to 5 (strongly agree). The reliability across the three items was a ¼ .89. The second personal relevance factor, efficacy belief, is defined as an individual’s belief in the efficacy of various medical procedures associated with a disease (Johnson, 1997). We measured respondents’ efficacy belief with three items (e.g., ‘‘I am confident in my ability to get my children vaccinated against preventable infectious diseases.’’) using a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). The reliability across the three items was a ¼ .78. (Table 2 describes measurement items for the personal relevance factors included in this study.) 2
The Korean Centers for Disease Control and Prevention’s standard immunization schedule recommends that the sixth dose of DTap and the fifth dose of Japanese encephalitis vaccine be administered before the age of twelve, whereas most of other recommended vaccinations—including four doses of diphtheria, tetanus toxoids, and acellular pertussis vaccine (DTaP); three doses of HepB; three doses of Polio; and one dose of MMR—are expected to be complete before the age of six.
H. O. Lee and S. Kim Information Carrier Factors Johnson and Meischke (1993) defined the information carrier factor as a perceived utility of the information carrier. In this study, we focused on six vaccine information sources widely used in South Korea, including (a) physicians=nurses, (b) family=friends, (c) TV=radio, (d) newspapers=magazines, (e) the Internet, and (f) the Child Pocket Book. The utility of a vaccine information source was indicated by two questions asking respondents’ level of trust in a given source and their perceived usefulness of that source. Respondents marked on a 5-point Likert-type scale with the anchors 1 (not at all) to 5 (very much) for each of the six sources, and the reliabilities of the two items across all sources were satisfactory, ranging from a ¼ .85 (TV= radio) to a ¼ .90 (expert sources). Use of Vaccine Information From Different Sources Information-seeking actions, the eventual outcome of the preceding factors under the CMIS, were operationalized as the extent of mothers’ use of various sources to seek childhood vaccine information. A single item (‘‘How much vaccination-related information do you obtain from the following information sources?’’) was used to assess the information-seeking behavior. The means and standard deviations for the use of each source were reported in Table 1. Behavioral Intention to Childhood Vaccination One of the main interests of the present study is to examine the consequence of the use of various health information sources, which the CMIS does not explicate under the current frame. As the behavioral outcome of information-seeking actions, respondents’ intention to have their children vaccinated was assessed by three items (e.g., ‘‘I will have my children vaccinated on time regardless of my busy schedule.’’) followed by a 5-point Likert-type scale ranging from 1 (I definitely will not) to 5 (I definitely will). The reliability across the three items was a ¼ .79. Correlations, means, and standard deviations for key variables across six different sources are presented in Table 1.
Results We first estimated a measurement model by conducting a confirmatory factor analysis (CFA) to assess the validity of the factor structure of each construct measured in this study, and then examined structural models which specify the theoretical relations among the variables. A minimum factor loading of .50 was used as a criterion for including items to be a part of a construction. As criteria for a good model fit, we used v2, CFI (comparative fit index), GFI (goodness-of-fit index), NFI (normed fit index), and RMSEA (root mean square error of approximation) following recommendations in the literature (e.g., Bentler, 1980; Browne & Cudeck, 1993). For CFI, GFI, and NFI, the desired threshold is above .90 (Byrne, 2001; Hu & Bentler, 1999); for RMSEA, values less than or equal to .08 indicate ‘‘reasonable’’ fit, though values of .06 or below should be preferred (Browne & Cudeck,1993; Jo¨reskog & So¨rbom, 2003). The measurement and structural models were examined using AMOS 16.0. The result of the CFAs with maximum likelihood estimation
289
M
SD
1
1. Age 37.81 4.70 — 2. Education 3.03 0.78 –.02 3. Income 3.80 1.81 .13 4. Salience 3.83 0.62 .09 5. Efficacy 3.51 0.61 .05 6. Utility_expert 3.78 0.63 .02 7. Utility_nonexpert 3.42 0.62 –.05 8. Utility_tv=radio 3.25 0.68 .10 9. Utility_news=mag 3.13 0.66 .10 10. Utility_Internet 3.26 0.63 –.10 11. Utility_Pocket Book 3.80 0.65 .09 12. Use_expert 3.31 0.85 –.03 13. Use_nonexpert 3.31 0.80 –.12 14. Use_tv=radio 2.09 0.76 .06 15. Use_news=mag 2.48 0.86 .04 16. Use_Internet 3.26 0.91 –.25 17. Use_Pocket Book 3.82 0.77 .04 18. Vaccination intention 3.87 0.65 –.01
Variable
3
4
5
6
7
— .25 — –.04 .03 — .07 .14 .25 — .03 .10 .22 .19 — –.02 –.02 .12 .11 .33 — .06 .09 .14 .08 .30 .22 .05 .14 .13 .11 .33 .23 –.03 .02 .12 .07 .26 .36 .02 .00 .19 .17 .43 .29 .06 .04 .13 .17 .49 .25 .01 –.02 .07 .07 .23 .63 –.02 .01 –.01 .02 –.02 –.01 .00 .08 –.03 –.01 .06 .07 .01 .01 .01 .02 .14 .18 .05 .01 .14 .17 .30 .23 .03 .02 .46 .37 .26 .14
2
Table 1. Correlations, means, and standard deviations for indicants (N ¼ 1,004)
— .65 .29 .24 .03 .10 .21 .16 .08 .14 .09
8
— .40 .24 .11 .12 .18 .44 .19 .17 .12
9
11
— .19 — .17 .23 .21 .13 .07 –.03 .19 .00 .60 .02 .15 .75 .15 .26
10
— .31 .11 .19 .20 .25 .15
12
14
15
16
17
18
— .04 — .14 .41 — .25 .10 .28 — .15 –.03 .01 .07 — .07 –.07 –.07 .06 .21 —
13
290
Goodness-of-fit indices of measurement models for six information sources
Utility of an information carrier
3 1 2 3 1 2 1 2 1 2 1 2 1 2 1 2 1 2
1 2
Salience
Efficacy
Item
Variable
Pocket Book
Internet
Newspapers=magazines
TV=radio
Family=friend
Physicians=nurses
Pocket Book
Internet
Newspapers=magazines
TV=radio
Family=friend
.83 .62 .72 .80
.80 .93
Factor loading
.80 .91 .82 .90 .72 .86 .89 .79 .74 .95 .78 .98 v2(17, N ¼ 1,004) ¼ 47.32, p < .001; CFI ¼ .99; GFI ¼ .99; NFI ¼ .99; RMSEA ¼ .04 v2(17, N ¼ 1,004) ¼ 34.48, p < .005; CFI ¼ .99; GFI ¼ .99; NFI ¼ .99; RMSEA ¼ .03 v2(17, N ¼ 1,004) ¼ 51.50, p < .001; CFI ¼ .99; GFI ¼ .99; NFI ¼ .98; RMSEA ¼ .05 v2(17, N ¼ 1,004) ¼ 51.86, p < .001; CFI ¼ .99; GFI ¼ .99; NFI ¼ .98; RMSEA ¼ .05 v2(17, N ¼ 1,004) ¼ 36.83, p < .005; CFI ¼ .99; GFI ¼ .99; NFI ¼ .99; RMSEA ¼ .03 v2(17, N ¼ 1,004) ¼ 41.70, p < .001; CFI ¼ .99; GFI ¼ .99; NFI ¼ .99; RMSEA ¼ .04
Infectious diseases that may be caused by lack of childhood vaccination are very serious. Infectious diseases that may be caused by lack of childhood vaccination can lead to other life-threatening diseases. Infectious diseases that may be caused by lack of childhood vaccination can lead to death. I believe that I have time to get my child(ren) vaccinated. I believe that I have no problem affording vaccinations for my child(ren). I am confident in my ability to get my child(ren) vaccinated against preventable infectious diseases. When obtaining information regarding childhood vaccination, how much trust do you have in __________? How useful is information from ___________ when making decisions about vaccination for your children? Physicians=nurses
Table 2. Results of confirmatory factor analysis: Factor loadings for measurement items
291
Linking Health Information Seeking to Behavior showed excellent model fits to the data (CFI, GFI, and NFI >.98; RMSEA
Linking Health Information Seeking to Behavioral Outcomes: Antecedents and Outcomes of Childhood Vaccination Information Seeking in South Korea HYUN OU LEE1 and SOYOON KIM2 1
Department of Advertising and Public Relations, College of Communication and Social Science, Hanyang University, Ansan, Republic of Korea 2 Department of Communication Studies, School of Communication, University of Miami, Coral Gables, Florida, USA
Although research on health information has made significant progress in identifying the antecedents of individuals’ information-seeking behavior in the context of the United States, the results have not been generalizable to the contexts of many other countries. Moreover, little is known about how one’s information-seeking behavior is connected to actual behavioral outcomes relevant to the search action. The authors conducted an online survey with a stratified random sample of 1,004 mothers to examine the applicability of the comprehensive model of health information seeking in predicting the use of diverse childhood vaccination information sources in South Korea, and to investigate associations between the mothers’ engagement with specific vaccine information sources and behavioral intention to immunize their children. Findings indicated that the hierarchical structure and the role of predictors within the comprehensive model of health information seeking provided a valid framework in the context of vaccine information seeking in Korea. In addition, the authors found negative associations between the use of certain types of information sources and mothers’ intention to vaccinate. This suggests that the dissemination of critical health information through a variety of available sources does not automatically lead to prudent behavioral decisions when the specific characteristics of the different sources are not considered.
Research concerning health information has become an important branch of communication research, both because of its critical implications for public health outcomes (Niederdeppe, 2008) and because of the wealth of information sources currently available (Cline & Haynes, 2001; Dutta-Bergman, 2004). While a considerable body of research has examined various factors of health information seeking in diverse contexts (e.g., Cotton & Gupta, 2004; DeLorme, Huh & Reid, 2011; Rains, 2008) few studies have systematically examined health information-seeking behavior outside the United States, limiting the external validity of the findings. Moreover, while much attention in the health information-seeking literature has been focused on identifying who actively seeks health information, far less attention has been paid to how specific health information-seeking actions influence behavioral outcomes (Anker, Reinhart, & Feeley, 2011; Rains, 2007). We, thus, had two goals for the present study. First, to test the applicability of the prediction of health informationseeking behavior found in the United States, we examine various factors that influence mothers’ use of vaccine Address correspondence to Hyun Ou Lee, Department of Advertising and Public Relations, College of Communication and Social Science, Hanyang University, 55 Hanyangdaehakro, Sangnok-gu, Kyeonggi-do, 426-791, Ansan, Republic of Korea. E-mail: [email protected]
information sources with regard to their children in the context of South Korea. The comprehensive model of information seeking (CMIS; Johnson, 1997; Johnson & Meischke, 1993), which includes a wide range of factors associated with information-seeking actions, has provided the theoretical framework for approaching this problem. Second, we investigate how the mothers’ engagement with specific vaccine information sources is connected to their behavioral intention to have their children vaccinated. In its identification of the behavioral consequences as well as the predictors of vaccine information seeking in the South Korean context, the current study offers a theoretical contribution to the extension of an existing health information-seeking model and provides practical implications for the implementation of communication strategies to disseminate critical health information. Childhood Vaccination and Parental Concerns The success of national immunization programs throughout the twentieth century has contributed to the global eradication of smallpox, the elimination of the polio virus in the United States, and a substantial reduction in the occurrence of vaccine-preventable diseases, including measles, mumps, rubella, diphtheria, pertussis, tetanus, and Haemophilus influenza type b (Centers for Disease Control and Prevention, 1999, 2008). While variations exist in the success of
286 increasing vaccination coverage among young children across different countries (Kee et al., 2007; Zhao, Smith, & Luman, 2009), South Korea is among several countries that have made significant progress in controlling infectious diseases by implementing active vaccine surveillance systems. Since the vaccination was first introduced to Korea in the 1960s, most conventional vaccines have been included in the national immunization program, and the estimated coverage rates of recommended vaccines, including hepatitis B vaccine (HepB), diphtheria-tetanus-pertussis (DTP), and measles-mumps-rubella (MMR), have exceeded 90% in 2012 (Lu & Santosham, 2012; Shin et al., 2009). Despite universal efforts to increase vaccination coverage, a recent drop in vaccination coverage, especially for MMR, has been observed in several countries (Danis, Georgakopoulou, Stavrou, Laggas, & Panagiotopoulos, 2010; Serpell & Green, 2006). Scholars have indicated that parental concerns about vaccine safety, which have been reported as one of the most important barriers to childhood vaccination (Brown et al., 2010; Gust et al., 2004), largely derive from a lack of information or knowledge concerning childhood vaccination (Ledford, Willett, & Kreps, 2012; Ramsay, Yarwood, Lewis, Campbell, & White, 2002). The average parents are unlikely to have firsthand knowledge of the seriousness of vaccinepreventable disease, nor are they likely to be aware of the true risks of vaccine-adverse events (Serpell & Green, 2006). When uncertainty around a health issue prevails, people are primarily motivated to seek information to reduce that uncertainty (Babrow & Kline, 2000). Although the contemporary view1 of uncertainty management highlights the complex nature of uncertainty that results in diverse patterns of information behaviors, including information avoiding as well as seeking, seeking health information is regarded as an essential first step in coping with health-related uncertainty and health behaviors (Tardy & Hale, 1998; Wanzer, Booth-Butterfield, & Gruber, 2004). Prediction of Health (Vaccine) Information Seeking: CMIS Health information seeking is defined as the purposive acquisition of health information from selected sources for determining one’s own health behaviors (Johnson & Meischke, 1993; Johnson, Donohue, Atkin, & Johnson, 1995). The use of various health information sources is influenced by many factors, including individuals’ socio-demographics (Freimuth, 1990; Lenz, 1984; Zarcadoolas, Blanco, Boyer, & Pleasant, 2002), health-related motivations (Cummings, Becker, & Maile, 1980), and the perception of information sources (DeLorme, Huh, & Reid, 2011; Rains, 2007). A theoretical framework proposed by Johnson and Meischke (1993), the CMIS, provides a useful theoretical framework for investigating the effects of such variables related to health information seeking. The CMIS is an integration of three theoretical perspectives. First, the uses and gratification perspective 1
Brashers and colleagues pointed out that people are not always motivated to gain more information to decrease uncertainty surrounding a health issue, as they tend to adjust their information seeking to the level of uncertainty they can tolerate (Brashers, 2001; Brashers, Goldsmith, & Hsieh, 2002; Brashers, Neidig, & Goldsmith, 2004).
H. O. Lee and S. Kim (Blumler & Katz, 1974; Rubin, 1986) imparts to the CMIS the concept of active consumers and their goal-directed use of various media to fulfill their various needs. Second, drawing on the health belief model (Rosenstock, 1974; Rosenstock, Strecher & Becker, 1988), the CMIS incorporates health-related factors, such as health concerns and perceived susceptibility, as the driving forces behind information seeking. Last, the CMIS emphasizes the link between the evaluation and use of communication channels based on the model of media exposure and appraisal (Johnson, 1983). As a result, the CMIS includes three classes of variables: (1) antecedents to information seeking (i.e., background factors [demographics and personal experiences] and personal relevance factors [salience and efficacy belief]), which determine the perception of an information source; (2) information carrier factors (i.e., the characteristics and the utilities of the media), which, in turn, shape the specific intentions related to seeking information from a particular source; and (3) information-seeking actions, which are the outcomes of these classes of variables (Johnson, 1997). Since Johnson and Meischke (1993) demonstrated the superiority of information-carrier factors over background and personal relevance factors in anticipating the search for cancer information through magazines, the CMIS has been empirically tested in, or at least applied across, a variety of contexts in the United States (e.g., DeLorme, Huh, & Reid, 2011; Johnson, Andrews, & Allard, 2001; Rains, 2007). The findings of subsequent studies generally support the hierarchical structure of the CMIS in explaining information-seeking behaviors, but the specific patterns of relationships within the model have suggested some contingent effects (Case, Andrews, Johnson, & Allard, 2005). Also, an individual’s selection and use of sources of health information are dependent on their match with the seeker’s needs and expectations (Johnson, 1997). For example, while health care professionals are generally regarded as the most preferred sources of health information among patients or their family members (Andreassen, Randers, Naslund, Stockeld, & Mattiasson, 2005), friends and family are viewed as more effective than other sources when individuals need more tailored emotional support in obtaining complex and serious health information (Burleson & MacGeorge, 2002; Johnson, 1997). In addition to traditional mass media that plays an important role in providing detailed information about products and services related to health and illness, the Internet has become one of the most popular health information channels, primarily based on its accessibility (Cotton & Gupta, 2004; Johnson & Meischke, 1993; Nguyen & Bellamy, 2006). In particular, the Internet is one of the frequently used sources of health information in South Korea, as nearly 98% of households in the country have broadband connections (Organisation for Economic Co-operation and Development, 2010). A special medium also exists for childhood vaccine information in South Korea: the Child Pocket Book. A Child Pocket Book is a booklet issued by public health centers or private hospitals that enables parents to keep track of their children’s immunization schedules and records. Along with the Internet and interpersonal sources,
Linking Health Information Seeking to Behavior it is also one of the primary sources of vaccine information in South Korea (Korean Centers for Disease Control and Prevention, 2006). Given that the prediction of information-seeking behavior is specific to the type of disease and=or the cultural background of the seeker, it is necessary to ask whether the CMIS is a valid model for explaining information-seeking behavior across a variety of diseases, even outside the United States. Although the model of media exposure and appraisal, from which the CMIS’s hypothesized link between information carrier factors and information-seeking actions was drawn, has been tested in international settings (e.g., Johnson, 1983; Johnson & Oliveira, 1988), no research has systematically examined the CMIS framework as a whole to predict information-seeking behaviors related to childhood vaccination outside the United States, especially across various information channels. In light of this gap, the present study forwards the following research question to examine the general validity of the model for predicting the use of diverse sources of vaccine information, including (a) expert sources (i.e., health professionals); (b) nonexpert sources (i.e., family and friends); (c) entertainment-oriented media (i.e., television and radio); (d) information-oriented media (i.e., newspapers and magazines); (e) the Internet; and (f) the Child Pocket Book, in the South Korean context. Research Question 1: Is the CMIS an appropriate model to predict vaccine information-seeking actions in South Korea regardless of the types of sources examined?
In addition, we investigate the specific roles of demographic backgrounds, health-related motivational factors, and information carrier factors with following hypotheses that are drawn from relations among the variables in the CMIS. Hypotheses 1 through 6: The set of antecedents (i.e., demographic characteristics and health-related factors) and information carrier factors will influence health information seeking related to the extent of mothers’ use of childhood vaccine information from diverse sources, including expert sources (Hypothesis 1); nonexpert sources (Hypothesis 2); entertainment-oriented media (Hypothesis 3); information-oriented media (Hypothesis 4); the Internet (Hypothesis 5); and the Child Pocket Book (Hypothesis 6).
Outcomes of Information Seeking by Source As scholars have underscored the importance of obtaining health information as a first step in health behavior change (Freimuth, Stein, & Kean, 1989; Tardy & Hale, 1998), some information-seeking models, such as the Expanded Conceptual Model of Information Seeking Behaviors (Longo, 2005), assumed a positive link between obtaining health information and health-related decision-making processes and health outcomes. The CMIS also postulates that health-related behaviors are ultimately affected by the scope and depth of information-seeking actions. Johnson and
287 colleagues, however, indicated that the prediction of specific health-related behaviors is beyond the scope of the CMIS, as the model focuses on critical factors that influence information-seeking actions (Johnson, 1997; Johnson & Case, 2012). Notably, they highlighted potential negative outcomes of information-seeking actions (e.g., erroneous conclusions), stating that ‘‘It is not that people do not gather information or learn things, but often they gather the wrong information for the wrong reasons from the wrong source’’ (p. 133). Studies have continuously reported that not all sources of vaccine information affect childhood vaccination rates in the same way. While vaccine information obtained from expert sources or booklets provided by health care organizations can increase vaccine uptake (e.g., Gamertsfelder, Zimmerman & DeSensi, 1994; Korean Centers for Disease Control and Prevention, 2006), some widely used information sources have contributed significantly to increased parental concerns regarding childhood vaccination (Clarke, 2008; Cooper, Larson, & Katz, 2008). For example, mass media reports in the 1990s contributed to the distribution of the allegation of a possible link between MMR vaccine and autism, thus instigating the antivaccination movement (Ledford, Willett, & Kreps, 2012). Similarly, the advent of the Internet has increased access not only to accurate information, but also to misinformation about vaccines, thus providing additional channels of communication for antivaccine activists (Cooper, Larson, & Katz, 2008). As several scholars (Anker, Reinhart, & Feeley, 2011; Case, Andrews, Johnson, & Allard, 2005) have pointed out, most previous health information-seeking studies have identified the influencing factors of health informationseeking behavior; however, far fewer have focused on examining relevant outcomes associated with search actions. Thus, additional investigation regarding outcomes of health information seeking will be an important addition to current theoretical frameworks, including the CMIS. Accordingly, rather than viewing health information-seeking behaviors as a final outcome, we propose another research question to identify the effects of health information-seeking behaviors on the outcomes associated with actual health behaviors. Behavioral theories suggest that behavioral intention, which refers to one’s intention to perform a behavior, is the most direct indicator of an actual behavior (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975; Fishbein & Yzer, 2003). Given that the actual vaccination behavior will only be available from the same sample at a point in the future, we examine vaccination intention for children as the outcome of engagement with various vaccine information sources as the proxy of the performance of childhood vaccination. Research Question 2: How is the use of childhood vaccine information from nonexpert interpersonal sources, expert sources, entertainment-oriented media (TV=radio), information-oriented media (newspapers=magazines), the Internet, and the Child Pocket Book related to mothers’ behavioral intention to have their children vaccinated?
288 Method Data Collection A nationwide, cross-sectional web-survey of individuals’ perception on childhood vaccination was conducted through a professional research agency in South Korea during August 2010. The agency retains its own panel that consists of a stratified sample of 22,146 people from across the country based on 2005 Census data. To be eligible for participating in the survey, a woman had to have children younger than 12 years of age.2 Among the potential respondents who received the web-survey questionnaire, which was administered in the Korean language, only those who met the eligibility (n ¼ 2,459) and sent back complete survey responses (n ¼ 1,004) were included in the final data set. Mothers analyzed in this study were 23–59 (M ¼ 37.81, SD ¼ 1.49) years of age. The monthly household income of 48.2% of the sample ranged from US$2,500 to US$4,500, indicating that the profile of the study sample was comparable with national figures, as the average monthly household income in South Korea was about US$3,500 in 2010 (Korean Statistical Information Service, 2010). However, it was also found that those in the study sample were slightly better educated than the national population as a whole. Thus, caution should be exercised in generalizing the study’s results related to education level. Measures Antecedents We assessed respondents’ age, education level and monthly household income as indicators of demographics. The first personal relevance factor in the CMIS, salience, is operationalized as perceived severity, which refers to a person’s beliefs about the medical or social seriousness of a health threat or leaving it untreated (Janz & Becker, 1984; Rosenstock, Strecher, & Becker, 1994). Perceived severity of vaccinepreventable infectious disease was assessed by three items, including ‘‘Infectious diseases that may be caused by lack of childhood vaccination are very serious.’’ All items were scored on a five-point scale with anchors 1 (strongly disagree) to 5 (strongly agree). The reliability across the three items was a ¼ .89. The second personal relevance factor, efficacy belief, is defined as an individual’s belief in the efficacy of various medical procedures associated with a disease (Johnson, 1997). We measured respondents’ efficacy belief with three items (e.g., ‘‘I am confident in my ability to get my children vaccinated against preventable infectious diseases.’’) using a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). The reliability across the three items was a ¼ .78. (Table 2 describes measurement items for the personal relevance factors included in this study.) 2
The Korean Centers for Disease Control and Prevention’s standard immunization schedule recommends that the sixth dose of DTap and the fifth dose of Japanese encephalitis vaccine be administered before the age of twelve, whereas most of other recommended vaccinations—including four doses of diphtheria, tetanus toxoids, and acellular pertussis vaccine (DTaP); three doses of HepB; three doses of Polio; and one dose of MMR—are expected to be complete before the age of six.
H. O. Lee and S. Kim Information Carrier Factors Johnson and Meischke (1993) defined the information carrier factor as a perceived utility of the information carrier. In this study, we focused on six vaccine information sources widely used in South Korea, including (a) physicians=nurses, (b) family=friends, (c) TV=radio, (d) newspapers=magazines, (e) the Internet, and (f) the Child Pocket Book. The utility of a vaccine information source was indicated by two questions asking respondents’ level of trust in a given source and their perceived usefulness of that source. Respondents marked on a 5-point Likert-type scale with the anchors 1 (not at all) to 5 (very much) for each of the six sources, and the reliabilities of the two items across all sources were satisfactory, ranging from a ¼ .85 (TV= radio) to a ¼ .90 (expert sources). Use of Vaccine Information From Different Sources Information-seeking actions, the eventual outcome of the preceding factors under the CMIS, were operationalized as the extent of mothers’ use of various sources to seek childhood vaccine information. A single item (‘‘How much vaccination-related information do you obtain from the following information sources?’’) was used to assess the information-seeking behavior. The means and standard deviations for the use of each source were reported in Table 1. Behavioral Intention to Childhood Vaccination One of the main interests of the present study is to examine the consequence of the use of various health information sources, which the CMIS does not explicate under the current frame. As the behavioral outcome of information-seeking actions, respondents’ intention to have their children vaccinated was assessed by three items (e.g., ‘‘I will have my children vaccinated on time regardless of my busy schedule.’’) followed by a 5-point Likert-type scale ranging from 1 (I definitely will not) to 5 (I definitely will). The reliability across the three items was a ¼ .79. Correlations, means, and standard deviations for key variables across six different sources are presented in Table 1.
Results We first estimated a measurement model by conducting a confirmatory factor analysis (CFA) to assess the validity of the factor structure of each construct measured in this study, and then examined structural models which specify the theoretical relations among the variables. A minimum factor loading of .50 was used as a criterion for including items to be a part of a construction. As criteria for a good model fit, we used v2, CFI (comparative fit index), GFI (goodness-of-fit index), NFI (normed fit index), and RMSEA (root mean square error of approximation) following recommendations in the literature (e.g., Bentler, 1980; Browne & Cudeck, 1993). For CFI, GFI, and NFI, the desired threshold is above .90 (Byrne, 2001; Hu & Bentler, 1999); for RMSEA, values less than or equal to .08 indicate ‘‘reasonable’’ fit, though values of .06 or below should be preferred (Browne & Cudeck,1993; Jo¨reskog & So¨rbom, 2003). The measurement and structural models were examined using AMOS 16.0. The result of the CFAs with maximum likelihood estimation
289
M
SD
1
1. Age 37.81 4.70 — 2. Education 3.03 0.78 –.02 3. Income 3.80 1.81 .13 4. Salience 3.83 0.62 .09 5. Efficacy 3.51 0.61 .05 6. Utility_expert 3.78 0.63 .02 7. Utility_nonexpert 3.42 0.62 –.05 8. Utility_tv=radio 3.25 0.68 .10 9. Utility_news=mag 3.13 0.66 .10 10. Utility_Internet 3.26 0.63 –.10 11. Utility_Pocket Book 3.80 0.65 .09 12. Use_expert 3.31 0.85 –.03 13. Use_nonexpert 3.31 0.80 –.12 14. Use_tv=radio 2.09 0.76 .06 15. Use_news=mag 2.48 0.86 .04 16. Use_Internet 3.26 0.91 –.25 17. Use_Pocket Book 3.82 0.77 .04 18. Vaccination intention 3.87 0.65 –.01
Variable
3
4
5
6
7
— .25 — –.04 .03 — .07 .14 .25 — .03 .10 .22 .19 — –.02 –.02 .12 .11 .33 — .06 .09 .14 .08 .30 .22 .05 .14 .13 .11 .33 .23 –.03 .02 .12 .07 .26 .36 .02 .00 .19 .17 .43 .29 .06 .04 .13 .17 .49 .25 .01 –.02 .07 .07 .23 .63 –.02 .01 –.01 .02 –.02 –.01 .00 .08 –.03 –.01 .06 .07 .01 .01 .01 .02 .14 .18 .05 .01 .14 .17 .30 .23 .03 .02 .46 .37 .26 .14
2
Table 1. Correlations, means, and standard deviations for indicants (N ¼ 1,004)
— .65 .29 .24 .03 .10 .21 .16 .08 .14 .09
8
— .40 .24 .11 .12 .18 .44 .19 .17 .12
9
11
— .19 — .17 .23 .21 .13 .07 –.03 .19 .00 .60 .02 .15 .75 .15 .26
10
— .31 .11 .19 .20 .25 .15
12
14
15
16
17
18
— .04 — .14 .41 — .25 .10 .28 — .15 –.03 .01 .07 — .07 –.07 –.07 .06 .21 —
13
290
Goodness-of-fit indices of measurement models for six information sources
Utility of an information carrier
3 1 2 3 1 2 1 2 1 2 1 2 1 2 1 2 1 2
1 2
Salience
Efficacy
Item
Variable
Pocket Book
Internet
Newspapers=magazines
TV=radio
Family=friend
Physicians=nurses
Pocket Book
Internet
Newspapers=magazines
TV=radio
Family=friend
.83 .62 .72 .80
.80 .93
Factor loading
.80 .91 .82 .90 .72 .86 .89 .79 .74 .95 .78 .98 v2(17, N ¼ 1,004) ¼ 47.32, p < .001; CFI ¼ .99; GFI ¼ .99; NFI ¼ .99; RMSEA ¼ .04 v2(17, N ¼ 1,004) ¼ 34.48, p < .005; CFI ¼ .99; GFI ¼ .99; NFI ¼ .99; RMSEA ¼ .03 v2(17, N ¼ 1,004) ¼ 51.50, p < .001; CFI ¼ .99; GFI ¼ .99; NFI ¼ .98; RMSEA ¼ .05 v2(17, N ¼ 1,004) ¼ 51.86, p < .001; CFI ¼ .99; GFI ¼ .99; NFI ¼ .98; RMSEA ¼ .05 v2(17, N ¼ 1,004) ¼ 36.83, p < .005; CFI ¼ .99; GFI ¼ .99; NFI ¼ .99; RMSEA ¼ .03 v2(17, N ¼ 1,004) ¼ 41.70, p < .001; CFI ¼ .99; GFI ¼ .99; NFI ¼ .99; RMSEA ¼ .04
Infectious diseases that may be caused by lack of childhood vaccination are very serious. Infectious diseases that may be caused by lack of childhood vaccination can lead to other life-threatening diseases. Infectious diseases that may be caused by lack of childhood vaccination can lead to death. I believe that I have time to get my child(ren) vaccinated. I believe that I have no problem affording vaccinations for my child(ren). I am confident in my ability to get my child(ren) vaccinated against preventable infectious diseases. When obtaining information regarding childhood vaccination, how much trust do you have in __________? How useful is information from ___________ when making decisions about vaccination for your children? Physicians=nurses
Table 2. Results of confirmatory factor analysis: Factor loadings for measurement items
291
Linking Health Information Seeking to Behavior showed excellent model fits to the data (CFI, GFI, and NFI >.98; RMSEA
