Maintenance Notice

Due to necessary scheduled maintenance, the JMIR Publications website will be unavailable from Wednesday, July 01, 2020 at 8:00 PM to 10:00 PM EST. We apologize in advance for any inconvenience this may cause you.

Who will be affected?


Citing this Article

Right click to copy or hit: ctrl+c (cmd+c on mac)

Published on 31.07.09 in Vol 11, No 3 (2009): Jul-Sep

This paper is in the following e-collection/theme issue:

    Original Paper

    What Is My Cancer Risk? How Internet-Based Cancer Risk Assessment Tools Communicate Individualized Risk Estimates to the Public: Content Analysis

    1Health Communication and Informatics Research Branch, National Cancer Institute

    2Basic and Biobehavioral Research Branch, National Cancer Institute

    Corresponding Author:

    Erika A Waters, PhD, MPH

    National Cancer Institute

    Health Communication and Informatics Research Branch

    Executive Plaza North Room 4051B

    6130 Executive Blvd MSC 7365

    Bethesda, MD 20892-7365

    Tel: +1 301 496 8586

    Fax:+1 301 480 2198



    Background: Internet-based cancer risk assessment tools have the potential to inform the public about cancer risk and promote risk-reducing behaviors. However, poorly communicated information on these websites may result in unintended adverse health outcomes.

    Objective: This study examined whether: (1) Internet-based cancer risk assessment tools use risk communication formats that facilitate comprehension and reduce bias (as identified by the empirical literature); (2) the use of these formats varies by website affiliation; and (3) the websites provided information necessary to evaluate the quality of the risk estimate.

    Methods: A content analysis of Internet-based cancer risk assessment tools was conducted. The terms calculate cancer risk, cancer risk calculator, estimate cancer risk, assess cancer risk, and cancer risk assessment were searched using three search engines. We identified 47 risk assessment tools and coded each according to standardized criteria. We calculated simple frequencies on all coding categories and performed crosstabulations but did not conduct formal statistical analysis due to small cell sizes.

    Results: Use of risk communication formats that facilitate comprehension and reduce bias varied widely (eg, 30% of websites [14/47] provided absolute and comparative risk information but 83% [39/47] provided safety messages). Use of formats that facilitate comprehension varied by website affiliation and communication strategy (eg, only 8.3% [1/12] websites affiliated with the health care industry provided absolute and comparative risk information, but 83% [5/6] of websites affiliated with a governmental organization did so). Only 53% (25/47) of websites provided information about the statistical model or the peer-reviewed literature that was used to calculate the risk estimate.

    Conclusion: Internet-based cancer risk assessment tools varied in their use of risk communication formats that facilitate comprehension and reduce bias. Formats that are difficult to understand may cause people to misperceive their cancer risk and consequently take inappropriate action.

    J Med Internet Res 2009;11(3):e33




    Many laypeople obtain individualized cancer risk estimates from Internet-based risk assessment tools. Data from the 2005 Health Information National Trends Survey (HINTS), which comprises a representative sample of the United States’ adult population, reveal that approximately 25% of people have used the Internet to seek information about cancer [1]. Although the proportion of people who sought information about their cancer risk is not known, use of online cancer risk assessment tools is common. For example, the Your Disease Risk website averaged nearly 2000 visitors per day in 2006 (personal communication by Graham Colditz, 2006), and the National Cancer Institute’s Breast Cancer Risk Assessment Tool averaged over 1200 hits per day in 2007 (personal communication by Rick Manrow, 2008). The purpose of providing people with individualized risk estimates is to encourage them to engage in health-promoting behaviors [2,3], such as using the health care system appropriately, making good medical decisions, engaging in health screening, avoiding tobacco use, engaging in physical activity, and eating a healthy diet. To achieve this, it is crucial that the information be presented in a way that facilitates comprehension and does not bias risk perceptions [4].

    Gurmankin Levy and colleagues [5] examined the quality of risk estimates for 13 breast cancer risk assessment websites. Many of these risk assessment tools used different risk factors to calculate risk, and in some cases excluded well-established risk factors, such as age at first live birth. Consequently, it is not surprising that the risk estimates provided for a particular risk profile varied across assessment tools. These authors [5] also found that many Internet-based risk assessment tools did not provide sufficient information to evaluate website quality. For example, some websites did not contain information about institutional affiliations or identify the statistical model used to calculate risk [5,6]. Because millions of people seek cancer risk information from Internet sources [1], the public health implications of inaccurate or inadequate risk communication are clear. Risk assessment tools that provide incorrect information may lead people to experience negative outcomes, such as seeking too much or too little medical care.

    Even when risk assessment tools provide accurate risk estimates, the information needs to be presented clearly. People do not always understand probabilistic information [7-11], even though they often report wanting to use it to make medical decisions [12]. The dilemma is exacerbated by the fact that risk perceptions, risk comprehension, and decision making can be affected by the ways in which probability information is presented [4,13-18]. For example, arrays of stick figures (also referred to as “pictographs” and “icon arrays”) can make it easier for people to understand the effects of hypothetical medical treatments and can reduce the undue influence that side effects have on hypothetical treatment decisions [19]. Misperceiving one’s risk of illness may result in poor health decisions. For example, a woman who overestimates her risk of developing cervical cancer might undergo excessive screening, but she might also be so overwhelmed with anxiety that she avoids screening entirely.

    A number of researchers have tested various ways to communicate probabilistic risk information [9,14,16,20-22]. According to these studies, the optimal risk communication format depends on what the communication is expected to accomplish. For example, presenting a drug’s benefit only as a relative risk reduction (“reduce the overall mortality by 20.3%”) can be more effective in persuading physicians to prescribe a drug than presenting only absolute risk reduction information (“reduce overall mortality from”) [23] (page 123). However, if the goal is education rather than persuasion, providing both absolute and relative information is most effective [24]. Thus, no single risk communication format will be appropriate in all situations and under all conditions.

    If the goal of Internet-based risk assessment tools is to assist people in making good decisions about their health, they must communicate probabilistic risk information in a way that facilitates comprehension and reduces biased interpretations [4]. A large and growing literature on risk communication permits us to infer which formats are likely to be most useful in helping people understand cancer risk estimates (Table 1). However, it is unknown how these formats interact with each other to influence perceptions of risk. Consequently, Table 1 should not be viewed as a checklist of required criteria, but instead as a list of important factors to take into consideration when developing a tool. Just as readers would not use a statistical test without understanding its underlying principles, they should not develop a risk assessment tool without having some rudimentary understanding of how different risk communication formats might affect risk perceptions and comprehension. Comprehensive reviews of the complex issues surrounding probabilistic risk communication can be found at [21,22,24-27].

    Table 1. Communication formats that reduce bias and facilitate comprehension of probabilistic risk estimatesa
    View this table

    This study describes the risk communication formats that Internet-based cancer risk assessment tools use to convey individualized risk information to the public. This study asks two questions that, to our knowledge, have not been addressed previously: (1) Do the tools use risk communication formats that have been empirically shown to facilitate comprehension, and (2) Does the use of these formats vary by website affiliation? Following Gurmankin Levy and colleagues [5], we also examined whether the websites provided the basic information necessary to evaluate the quality of the risk estimate (ie, the statistical model or peer-reviewed citations).



    During October 2007 we conducted an Internet search to identify websites that provided individualized cancer risk assessment. The search was conducted by entering the search terms calculate cancer risk, cancer risk calculator, estimate cancer risk, assess cancer risk, and cancer risk assessment into the Google, MSN, and Yahoo! search engines. These search engines accounted for 82.6% of all Internet searches that originated in the United States in 2006 [34]. To locate the Internet-based cancer risk assessment tools, a total of 1500 websites were examined (ie, the first 100 search results for each of the five search terms, for each of the three search engines).

    Out of the 1500 websites examined, we identified 51 websites that gave specific cancer risk estimates. Forty-four of the identified websites were unique interactive websites that provided individualized cancer risk estimates. These websites required visitors to enter information about their status on several cancer risk factors. Seven of the identified websites were non-interactive; they stratified risk information by two or three variables (eg, lung cancer risk by smoking status and gender). These non-interactive websites were included because they provided more specific risk information than general population data. Of these 51 websites, four were excluded because they required information seekers to download a software program (n = 2), provide a mailing address (n = 1), or provide payment (n = 1) before obtaining results. A total of 47 websites were evaluated (Table 2).

    Table 2. Websites hosting cancer risk assessment tools (WebCite® links are listed below the original URL)
    View this table

    Standardized coding criteria were developed (Table 3) and each website was coded independently by two of the authors (EW and HS). The first 15 websites were used to calibrate the coding procedure, and inconsistencies in the remaining 32 websites were resolved through discussion. Inter-rater reliability was high (κ = 86.4). If information about the statistical model used to calculate the risk could not be found on the website, EW attempted to obtain it by contacting the website’s developer. Two attempts to reach the developers were made over a four-week period.

    Table 3. Website coding criteria
    View this table


    General website characteristics, risk communication formats, and the presence of information about the quality of the risk estimate were examined by calculating simple frequencies on all coding categories. The number and percentage of websites that used formats that had the most empirical support for facilitating comprehension and reducing bias was recorded (see Table 1 for a list of formats). Affiliation-based differences in the use of formats were examined using crosstabulations. Formal statistical analysis and significance testing was not possible due to small cell sizes.


    General Website Characteristics

    The general characteristics of the websites are described in Table 4. There are two areas of particular interest. First, the three most common organ sites for which websites provided assessments (breast, lung, and colorectal cancer) coincide with the three leading causes of cancer mortality in the United States [35]. The second item of interest is the widespread use of technical language. Although laypeople were the intended audience for nearly all of the sites, an overwhelming majority did not define medical and technical terms such as “biopsy,” “DCIS,” “mastectomy,” or “radon progeny” (Table 4).

    Table 4. General website characteristics (N = 47)
    View this table

    Formats for Communicating Individualized Cancer Risk Estimates

    In general, few websites used risk communication formats that facilitate comprehension of probabilistic information (Table 5). Few websites provided the risk estimate as numbers and words, or described how the information seeker’s cancer risk compared to the risk of experiencing other hazards. Only three websites framed the risk positively (eg, “998 chances in 1000 that you will not develop cancer”), and slightly less than half informed participants of the duration of the risk estimate (eg, 5-year risk). Approximately one-third of the websites provided a visual display. However, some risk communication formats were used widely. Seventeen of the 21 websites that provided any numeric information did so using the percentage format, six used the “N in 1000” format, and three used both. Twenty-four websites compared the information seeker’s cancer risk to other people’s risk, and fourteen of these also provided the absolute risk. An overwhelming majority of the websites provided information seekers with safety messages like “stop smoking.” Slightly more than half of the websites made at least one statement acknowledging that the estimate contained some degree of uncertainty.

    Table 5. Risk communication formats used by Internet-based cancer risk assessment tools to communicate individualized risk estimates (N = 47)
    View this table

    The use of risk communication formats that facilitate comprehension and reduce bias varied by website affiliation (Table 6). For example, websites affiliated with the health care industry were the least likely to communicate risk as percentages or as frequencies (1 of 12 websites) and to provide both absolute and comparative risk information (1 of 12 websites). Websites affiliated with cancer centers were the least likely to provide information about the duration of the risk estimate (1 of 8 websites) and to make at least one statement acknowledging that the estimate contained uncertainty (1 of 8 websites). Health portals were the least likely to provide a safety message or risk reduction strategies (2 of 5 websites). As mentioned in the Analyses section, formal statistical analyses and significance testing could not be performed due to small cell sizes (ie, out of 72 possible cells, only eight contained five or more websites and fifteen included no websites; see Table 6).

    Table 6. Website affiliation-based variations in the use of risk communication formats that facilitate comprehension and reduce bias (N = 47)
    View this table

    Information Necessary to Evaluate Website Quality

    The number of websites that provided information to help visitors evaluate the quality of the risk estimate was limited. Only 25 of the 47 (53.2%) websites provided either information about the statistical model used to calculate the risk or peer-reviewed citations, where such information could be found. Only 13 websites (27.7%) provided both a description of the model and peer-reviewed citations. One of the authors (EW) attempted to contact all 22 of the websites that did not provide any information about the model. Of the 15 websites that provided a valid email address or phone number, only 8 responded to either the first or second inquiry, and only 2 were able to provide the information.


    Internet-based cancer risk assessment tools can provide cancer risk and prevention information to millions of people worldwide. Effective risk communication has the potential to reduce cancer morbidity and mortality by motivating people to engage in healthy behaviors. However, poorly communicated risk information could mislead people or frighten them unduly, resulting in maladaptive health behaviors [37], such as over- or under-utilization of health care services.

    In October 2007, 47 Internet-based cancer risk assessment tools provided individualized cancer risk information to the public. Almost half (20 out of 47) of these websites were developed by cancer centers (n = 8) and the health care industry (n = 12). The three most common organ sites for which assessment tools were available (breast, lung, colorectal) corresponded to the most common causes of cancer mortality among men and women in the United States [35]. This suggests that the assessment tools are responding to a public health need. However, three factors suggest that some of these websites might not be as useful to the public as they seem at first glance. First, many of the websites did not identify the statistical model used to calculate the risk estimate. Thus, visitors to these sites would have no way to verify that the risk assessment model had been scientifically vetted. The second factor that raises questions about the utility of some of these tools is the extensive use of undefined scientific terminology. Terms like “biopsy” and “radon progeny” may be confusing for individuals with limited literacy, thereby limiting their usefulness.

    The third factor that raises concerns about these tools is the fact that they varied widely in their use of risk communication formats that facilitate comprehension of probabilistic information. For some risk communication formats, most websites followed experts’ recommendations (eg, 39 of 47 sites provided safety messages). For other formats, few websites followed recommendations (eg, 3 of 47 sites used positive and negative framing). It is unclear whether it is necessary to include all of the risk communication formats identified in Table 1 in a single communication effort. Indeed, it is possible that doing so would overwhelm people with too much information and bias their risk perceptions [38]. However, risk perceptions can also be biased if people receive too little information. For example, seven websites provided information seekers with a simple count of risk factors or a “risk score.” One such website then informed people, “You should do something to reduce your risk,” but it did not provide specific safety messages, such as “stop smoking.” Telling people to “do something” without providing specific recommendations does not facilitate comprehension [33] and might be counterproductive [32,37].

    Two websites underscore the potential pitfalls of communicating risk using only numbers or only words. One of these websites provided two numerical risk estimates: one represented the risk if the information seeker continued the health-damaging behavior, and the other represented the risk if the individual stopped the behavior. However, the absolute risk was low in both cases (2% if the behavior continued and 1% if the behavior stopped), which might discourage people from changing their behavior [24,33,39]. The second website provided risk information as words only. It described cancer risk as “moderate,” “increased,” “high,” and “highest.” There was no “low risk” category. Thus, an 18-year-old with no risk factors for the type of cancer addressed on this website would be informed that he has a “moderate” risk of cancer. The combination of using only words to describe a risk and using the word “moderate” to describe the lowest-risk category might exaggerate risk perceptions among people at lowest risk [40].

    It is also important to note that online risk assessment tools often have different objectives. For example, some risk assessment tools are designed to educate people about healthy lifestyles, whereas others are designed to persuade people to use their services, purchase their products, or engage in a particular health behavior. While it may be acceptable to attempt to persuade people to purchase necessary goods or services, or to engage in appropriate health behaviors, it is incumbent on developers of cancer risk assessment tools to consider the point at which persuasion may infringe upon an individual’s ability to make an informed decision. It is beyond the scope of this study to assess the underlying motivations of the various risk assessment tool developers, but website developers, clinicians, researchers, and public health officials should consider the ethical implications of the tools they design.

    Despite these concerns, cancer risk assessment tools have the potential to play an important role in cancer prevention and control. However, in order to be acceptable and effective, they should use language that is appropriate for low-literacy populations, communicate risk estimates using formats that increase comprehension and reduce biased interpretations of risk, and provide information about the model used to calculate the estimate.

    As mentioned previously, the communication formats chosen will depend upon the goals of the communication. For example, if a developer’s goal is to educate women about how the risk of breast cancer increases with age, the ideal website might use a line graph to portray how her risk changes over time [27]. It might describe in words how the risk changes and add the probability estimates at 5-year intervals. It might also include lines that depict the “average woman’s” risk over the same time period and the risk of heart disease over the same time period. The website would include accurate information about how often women should undergo mammography screening, how to obtain a mammogram, and how to maintain a healthy weight. A statement would inform visitors that the risk estimate contains some uncertainty, and it would identify the statistical model that was used to generate the risk estimate. Finally, before the website is released to the public, it would be tested for readability and comprehensibility among people with low literacy and to ensure that it elicited relatively unbiased risk perceptions. If it was found to be confusing, the website developer might consider removing some of the information, such as how the woman’s risk of breast cancer compared to her risk of heart disease. It might also test whether providing users with control over the amount of information displayed—such as being able to add or remove the 5-year probability estimates—results in better outcomes.

    Strengths, Limitations, and Future Directions

    This content analysis fills a gap in existing research on Internet-based cancer risk assessment tools. This is the first study to describe how these tools communicate individualized risk estimates to the public and to examine the relationship between risk communication format and developer affiliation. Furthermore, this study replicated the finding that information needed to evaluate the quality of the risk estimate is often missing from these tools [5].

    Our extensive search methodology increases the likelihood that most of the Internet-based cancer risk assessment tools that were available in mid-October 2007 were included in this analysis. However, since then, some tools might have added or removed features, additional websites might have been activated, and existing websites might have been deactivated. The small number of websites precluded formal statistical analysis of affiliation-based differences in communication formats. Furthermore, we only evaluated the presence or absence of specific risk formats. We did not evaluate whether the websites implemented the risk communication formats appropriately. To the best of our knowledge, empirical research has not identified the relative importance of each of the communication formats examined in this study. Consequently, we did not evaluate the overall quality of each website, nor can we say that using a greater number of risk communication formats results in better comprehension than using a lower number of formats.

    Future research should examine the demographic and cancer risk profiles of people who use Internet-based cancer risk assessment tools to identify how these tools influence cancer-related cognitions, emotions, and behaviors. This will be increasingly important as we enter the second generation of the Internet [41]. Web 2.0 is eroding the conventional boundaries between information providers and information seekers. The move from static personal websites to blogs, from online encyclopedias to Wikipedia, and from one-player video games to complex interactive virtual worlds exemplifies this transformation. The Centers for Disease Control and the American Cancer Society already have virtual outposts in the online world Second Life [42]. The interactive capabilities of Web 2.0 to help people understand cancer risk and cancer risk reduction strategies may be an important means to promote a healthy lifestyle.


    Internet-based cancer risk assessment tools have the potential to reach a wide audience and motivate people to engage in cancer preventive behaviors. However, because the tools do not always communicate risk in ways that facilitate comprehension, patients may misperceive or be confused about their cancer risk. This confusion may result in unintended negative consequences, such as failing to seek appropriate medical care. Researchers, organizations and clinicians who wish to provide risk assessment services to the public or their patients should refer to several excellent reviews [20,22,25] for specific recommendations on communicating risk in ways that facilitate comprehension, rather than foment confusion.


    Drs. Erika Waters and Helen Sullivan conducted this research as Cancer Prevention Fellows, Office of Preventive Oncology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland. The National Cancer Institute reviewed and approved this manuscript for publication.

    Conflicts of Interest

    None declared.


    1. Finney Rutten LJ, Moser RP, Beckjord EB, Hesse BW, Croyle RT. Cancer Communication: Health Information National Trends Survey. NIH Pub. No. 07-6214. Washington, DC: National Cancer Institute; 2007.
    2. Ajzen I. From intentions to action: a theory of planned behavior. In: Kuhl J, Beckman J, editors. Action Control: From Cognitions to Behaviors. New York, NY: Springer; 1985:11-39.
    3. Rosenstock IM. The health belief model. In: Glanz K, Lewis F, Rimer B, editors. Health Behavior and Health Education: Theory, Research and Practice. San Francisco, CA: Jossey-Bass; 1990:39-62.
    4. Wills CE, Holmes-Rovner M. Patient comprehension of information for shared treatment decision making: state of the art and future directions. Patient Educ Couns 2003 Jul;50(3):285-290. [Medline] [CrossRef]
    5. Gurmankin Levy A, Sonnad SS, Kurichi JE, Sherman M, Armstrong K. Making sense of cancer risk calculators on the web. J Gen Intern Med 2008 Mar;23(3):229-235 [FREE Full text] [Medline] [CrossRef]
    6. Ekman A, Hall P, Litton JE. Can we trust cancer information on the Internet?—A comparison of interactive cancer risk sites. Cancer Causes Control 2005 Aug;16(6):765-772. [Medline] [CrossRef]
    7. Lipkus IM, Samsa G, Rimer BK. General performance on a numeracy scale among highly educated samples. Med Decis Making 2001;21(1):37-44. [Medline] [CrossRef]
    8. Schwartz LM, Woloshin S, Welch HG. Can patients interpret health information? An assessment of the medical data interpretation test. Med Decis Making 2005;25(3):290-300. [Medline] [CrossRef]
    9. Reyna V, Brainerd CJ. The importance of mathematics in health and human judgment: numeracy, risk communication, and medical decision making. Learn Individ Differ 2007;17(2):147-159. [CrossRef]
    10. Yamagishi K. When a 12.86% mortality is more dangerous than 24.14%: implications for risk communication. Appl Cogn Psychol 1997;11(6):495-506. [CrossRef]
    11. Zikmund-Fisher BJ, Smith DM, Ubel PA, Fagerlin A. Validation of the Subjective Numeracy Scale: effects of low numeracy on comprehension of risk communications and utility elicitations. Med Decis Making 2007;27(5):663-671. [Medline] [CrossRef]
    12. Woloshin S, Schwartz LM, Welch HG. Patients and medical statistics. Interest, confidence, and ability. J Gen Intern Med 2005 Nov;20(11):996-1000 [FREE Full text] [Medline]
    13. Edwards A, Elwyn G, Gwyn R. General practice registrar responses to the use of different risk communication tools in simulated consultations: a focus group study. BMJ 1999 Sep 18;319(7212):749-752 [FREE Full text] [Medline]
    14. Rothman AJ, Martino SC, Bedell BT, Detweiler JB, Salovey P. The systematic influence of gain- and loss-framed messages on interest in and use of different types of health behavior. Pers Soc Psychol Bull 1999;25(11):1355-1369. [CrossRef]
    15. Waters EA, Weinstein ND, Colditz GA, Emmons K. Formats for improving risk communication in medical tradeoff decisions. J Health Commun 2006 Mar;11(2):167-182. [Medline] [CrossRef]
    16. Cuite CL, Weinstein ND, Emmons K, Colditz G. A test of numeric formats for communicating risk probabilities. Med Decis Making 2008;28(3):377-384. [Medline] [CrossRef]
    17. Gigerenzer G, Hoffrage U. How to improve Bayesian reasoning without instruction: frequency formats. Psychol Rev 1995;102(4):684-704. [CrossRef]
    18. McNeil BJ, Pauker SG, Sox HC, Tversky A. On the elicitation of preferences for alternative therapies. N Engl J Med 1982 May 27;306(21):1259-1262. [Medline]
    19. Waters EA, Weinstein ND, Colditz GA, Emmons KM. Reducing aversion to side effects in preventive medical treatment decisions. J Exp Psychol Appl 2007 Mar;13(1):11-21. [Medline] [CrossRef]
    20. Fagerlin A, Ubel PA, Smith DM, Zikmund-Fisher BJ. Making numbers matter: present and future research in risk communication. Am J Health Behav 2007;31(Suppl 1):S47-S56. [Medline]
    21. Feldman-Stewart D, Brundage MD, Zotov V. Further insight into the perception of quantitative information: judgments of gist in treatment decisions. Med Decis Making 2007;27(1):34-43. [Medline] [CrossRef]
    22. Lipkus IM. Numeric, verbal, and visual formats of conveying health risks: suggested best practices and future recommendations. Med Decis Making 2007;27(5):696-713. [Medline] [CrossRef]
    23. Forrow L, Taylor WC, Arnold RM. Absolutely relative: how research results are summarized can affect treatment decisions. Am J Med 1992 Feb;92(2):121-124. [Medline] [CrossRef]
    24. Fagerlin A, Zikmund-Fisher BJ, Ubel PA. "If I'm better than average, then I'm ok?": Comparative information influences beliefs about risk and benefits. Patient Educ Couns 2007 Dec;69(1-3):140-144 [FREE Full text] [Medline] [CrossRef]
    25. Schwartz LM, Woloshin S, Welch HG. Risk communication in clinical practice: putting cancer in context. J Natl Cancer Inst Monogr 1999;25(1):124-133. [Medline]
    26. Ancker JS, Senathirajah Y, Kukafka R, Starren JB. Design features of graphs in health risk communication: a systematic review. J Am Med Inform Assoc 2006;13(6):608-618 [FREE Full text] [Medline] [CrossRef]
    27. Lipkus IM, Hollands JG. The visual communication of risk. J Natl Cancer Inst Monogr 1999;25(1):149-163.
    28. Klein WM. Comparative risk estimates relative to the average peer predict behavioral intentions and concern about absolute risk. Risk, Decision and Policy 2002;7(2):193-202. [CrossRef]
    29. Klein WM. Objective standards are not enough: affective, self-evaluative, and behavioral responses to social comparison information. J Pers Soc Psychol 1997 Apr;72(4):763-774. [Medline] [CrossRef]
    30. Lipkus IM, Klein WM. Effects of communicating social comparison information on risk perceptions for colorectal cancer. J Health Commun 2006 Jun;11(4):391-407. [Medline] [CrossRef]
    31. Weinstein ND, Sandman PM. The precaution adoption process model. In: Glanz K, Rimer B, Lewis F, editors. Health Behavior and Health Education: Theory, Research, and Practice. 3rd edition. San Francisco, CA: John Wiley and Sons; 2002:121-143.
    32. Leventhal H, Diefenbach M, Leventhal EA. Illness cognition: using common sense to understand treatment adherence and affect cognition interactions. Cognit Ther Res 1992;16(2):143-163. [CrossRef]
    33. Weinstein ND. What does it mean to understand a risk? Evaluating risk comprehension. J Natl Cancer Inst Monogr 1999;25(1):15-20. [Medline]
    34. Sullivan D. Nielsen NetRatings Search Engine Ratings. 2006.   URL: [accessed 2009 Jul 14]
    35. ; American Cancer Society. Cancer Facts & Figures. American Cancer Society. 2007.   URL: [accessed 2009 Jul 14]
    36. WebMD - Better Information. Better Health. WebMD. 2009.   URL: [accessed 2009 Jul 17] [WebCite Cache]
    37. Rogers RW, Prentice-Dunn S. Protection motivation theory. In: Gochman DS, editor. Handbook of Health Behavior Research I: Personal and Social Determinants. New York, NY: Plenum Press; 1997.
    38. Hay J, Coups E, Ford J. Predictors of perceived risk for colon cancer in a national probability sample in the United States. J Health Commun 2006;11(Suppl 1):71-92. [Medline] [CrossRef]
    39. Slovic P, Monahan J, MacGregor DG. Violence risk assessment and risk communication: the effects of using actual cases, providing instruction, and employing probability versus frequency formats. Law Hum Behav 2000 Jun;24(3):271-296. [Medline] [CrossRef]
    40. Berry DC, Michas IC, Bersellini E. Communicating information about medication side effects: effects on satisfaction, perceived risk to health, and intention to comply. Psychol Health 2002;17(3):247-267.
    41. O’Reilly T. What is Web 2.0? Design Patterns and Business Models for the Next Generation of Software. O’Reilly Media Inc. 2005.   URL: [accessed 2009 Jan 28] [WebCite Cache]
    42. Linden Research, Inc. Second Life. 2009.   URL: [accessed 2009 Jul 17] [WebCite Cache]


    HINTS: Health Information National Trends Survey

    Edited by G Eysenbach; submitted 24.02.09; peer-reviewed by J Hay, I Lipkus; comments to author 16.04.09; revised version received 29.04.09; accepted 17.06.09; published 31.07.09

    All of the authors conducted this work as a part of their duties as Federal employees, and so copyright does not apply. Originally published in the Journal of Medical Internet Research (, 31.07.2009.  

    Articles published in JMIR are - unless stated otherwise - open-access articles distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on, as well as this copyright and license information must be included.