To search is to believe? A comparative study of health information use by internet users
This study was designed to characterize health information seekers on the Internet using relatively large-scale survey data. The primary goals are to (1) ascertain whether the information seekers' credibility assessment of online health information varies by levels of Web searching activities, and (2) identify specific impacts of online health information on decision-making. Using a national survey by Pew Internet & American Life Project (2006) (N = 2,928), we conducted TwoStep cluster analysis focusing on the health-related topics. We successfully identified and labeled the two clusters of health information users as ‘active’ and ‘less active’ users. The data suggests that active users were more likely to evaluate the credibility of online health information resources than less active users; types of users did not make a difference in searches on behalf of others. More importantly, between active and less active users online health information had strong impacts on three specific aspects of decision-making: (1) the treatment on an illness or condition, (2) the overall approach and (3) asking new questions. The results advance our understanding of users' credibility assessment of online health information sources. Given these findings we provide avenues for future research.
The Internet has become a prevalent source for people to obtain health information (Fox, 2006; Johnson, et al. 2006). From perspectives of consumer health information, one of the common themes emerging from studies of health information seeking is the importance of taking into account the larger environmental and role-related contexts (e.g., Dey, 2004). For instance, user characteristics, such as self-reported health status, have been examined in relation to information environment (Goldner, 2006; Nicholas, 2001) and user perceptions about information resources (Navarro & Wilkins, 2001; Rains, 2007). For the health information seekers, Lorence and Park (2006) proposed to classify them according to the topics they have searched; Navarro and Wilkin (2001) suggest that there are both ‘active’ and ‘less active’ seekers. The issues pertaining to credibility assessment of health information on the Web have been extensively studied, since the quality of health information is critical to the user's information acquisition and subsequent decision-making (see e.g., Metzger, 2007; Rieh & Danielson, 2007).
Drawing from a relatively large scale national survey data (Pew Internet & American Life Project, 2006), this study was designed to (1) ascertain whether the information seekers' credibility assessment of online health information varies by levels of Web searching activities, and (2) identify specific impacts of online health information on decision-making.
We conducted a secondary quantitative data analysis of the 2006 Health Survey Data Set (Pew Internet & American Life Project, 2006). It consisted of a sample of 2,928 adults aged 18 and older. A large proportion of respondents (68.0%, N = 1,990) reported that they ever went online to access the Internet. Among the Internet users, those who have searched health information on the Internet (N = 1,594) accounted for 80.1%. Using this dataset, we first conducted a TwoStep cluster analysis to identify two groups of Internet health information searchers according to their search activity, and then used Chi-square tests to compare the differences between these two groups.
Data analysis and results
We conducted TwoStep Cluster Analysis (from SPSS, version 16.0) of the Internet users based on the 17 types of health-related topics in the survey. Our results revealed 3 clusters: non-searchers, active searchers and less active searchers. For purposes of this study we focused on ‘active’ and ‘less active’ searchers. Overall, there was statistically significant difference between these two groups in terms of the number of topics they have searched. In average, active searchers searched for about eight topics, while less active searchers searched for only two to three topics. For both groups of searchers, information about “a specific disease or medical problem” and about “a certain medical treatment or procedure” is the most popular topic they have searched; and information about “how to quit smoking” and about “problems with drugs and alcohol” has seldom been searched. This trend is consistent with results in previous studies (e.g. Lorence & Park, 2006; Rice, 2006).
Even though Goldner (2006) found that individuals who have a medical condition are more likely to search most health topics online than healthy individuals, we did not find significant difference between active and less active searchers in terms of their self-reported health status (χ2 (3, N=1585) =.110, p=.991, V=.008). In addition, most of searchers (86.6% of ‘active searchers’; 86.4% of ‘less active searchers’) rated their overall health condition as good or excellent. It suggested that self-reported health status may not be the primary motivational factor for people's health information seeking on the Internet, because people often search health information because of or on behalf of others (Harbour & Chowdhury, 2007; Morey, 2007; Fisher & Julien, 2008). Our result from a comparison of the question of whom they were searching for also was not significant (chi-square test, χ2 (2, N=1463) = 8.793, p=.012, V=.078). Similarly, we found over half of both active (51.1%) and less active (53.5%) searchers searched for someone else.
Information credibility is about whether people believe what they find on the Web. In order to judge the credibility, the searchers for the medical task may check the source reputation, author/creator credentials and type of source (Rieh, 2002). We then compared the information credibility behavior between the two groups according to whether they asserted to check the information providers and whether they checked the date and the reviewer of the information. The results revealed that active searchers check the information providers significantly more than less active searchers (χ2 (4, N=1567) = 231.96, p<0.0001, V=.385), and also check the date and the reviewer significantly more often (χ2 (4, N=1581) = 210.43, p<0.0001, V=.365). Thus, based on Rieh's (2002) results, we can conclude that active searchers were more likely to evaluate the credibility of online health information resources than less active searchers.
As to the perceived usefulness, significantly more active searchers (60.5%) claimed that the information they found had major or minor impact on themselves or someone else than less active searchers (39.5%). Especially, online health information had strong impact on three specific aspects of decision-making of active searchers: (1) the treatment on an illness or condition, (2) the overall approach and (3) asking new questions.
These results advance our understanding of users' credibility assessment of online health information sources, and of reasons for success of searching. The differences between active and less active searchers in their activities of credibility assessment and the impact of acquired health information on decision-making suggest that interventions that encourage credibility assessment might lead to more useful search results for less active searchers. Other characteristics of information seekers, such as health-related domain knowledge, gender, and motivational characteristics (see e.g., Abrahamson, 2008; Keselman, Browne & Kaufman, 2008; Rieh & Danielson, 2007), in lieu of levels of Web searching activities, are promising avenues for future research.
The authors would like to thank the Institute of Museum and Library Services (IMLS) for supporting this study, and two anonymous reviewers for their comments.