Authors

  1. SHAW, BRET R. PhD
  2. HAWKINS, ROBERT PhD
  3. ARORA, NEERAJ PhD
  4. McTAVISH, FIONA MS
  5. PINGREE, SUZANNE PhD
  6. GUSTAFSON, DAVID H. PhD

Abstract

This study examined what characteristics predict participation in online support groups for women with breast cancer when users are provided free training, computer hardware, and Internet service removing lack of access as a barrier to use. The only significant difference between active and inactive participants was that active users were more likely at pretest to consider themselves active participants in their healthcare. Among active participants, being white and having a higher energy level predicted higher volumes of writing. There were also trends toward the following characteristics predictive of a higher volume of words written, including having a more positive relationship with their doctors, fewer breast cancer concerns, higher perceived health competence, and greater social/family well-being. Implications for improving psychosocial interventions for women with breast cancer are discussed, and future research objectives are suggested.

 

Article Content

The number of people participating in computer- mediated social support (CMSS) groups currently counts in the millions 1 and has been rising steadily in the past few years.2 Experts have predicted that computer-based health communication services will continue to expand, and consumers will increasingly turn to them for health information, communication, and support.3 In the past decade, several studies on the ways women with breast cancer use CMSS groups have been published. Numerous descriptive accounts related to whether people will use such groups, the extent to which they will do so, and the general nature of the support exchanged within these groups have been published.4-6 Other research offers a more qualitative, phenomenological account of the discourse within CMSS groups for women with breast cancer.7,8 Recent research illustrates some of the unique ways that women with breast cancer from different demographic backgrounds participate in CMSS groups9 and demonstrates that participation in CMSS groups can affect various mental health outcomes.10,11 Despite the growing body of research, there is still a lack of understanding about these groups.12,13 One current gap in nursing informatics and health communication research is an understanding of what types of people are most likely to use computer-mediated support groups in the first place. The purpose of this current study is to provide insights about the characteristics of individuals who participate in CMSS groups for women with breast cancer.

 

Such questions are relevant to a number of practical concerns. First, it is important to know whether certain population segments are more likely to be the most active participants in CMSS groups. If training and access are removed as barriers to use, what characteristics predict level of participation in such groups? Research identifying the most appropriate psychosocial interventions for various population segments can help healthcare providers efficiently allocate resources to serve different populations most effectively. With increased awareness about the value of social support in helping women to cope with breast cancer, it has been suggested that additional sources of support be identified.14 In this pursuit, one of the purposes of the current research endeavor is to discover how various antecedent characteristics affect level of participation in the groups with the hope that these efforts will inform future interventions designed to meet the psychosocial needs of women who are most likely to use and benefit from participation in CMSS groups.

 

REVIEW OF LITERATURE

Although there is very little research examining what characteristics predict participation in CMSS groups, various factors influence how women cope with a breast cancer diagnosis. For example, demographics affect what channels of communication people use to cope with their cancer.15 Two small studies have found that age and educational level have no effect on participants' propensity to use online support groups.4,16 Nevertheless, evidence that demographic variables affect participation in a computer-mediated support group is quite limited, and in other contexts, research has consistently found that demographics do affect how people respond to a cancer diagnosis.17 For instance, some research indicates that younger and less educated women tend to be more anxious and depressed following a breast cancer diagnosis and that more educated women are likely to employ less avoidant coping strategies.18 Other research has found that women of color are generally less likely to seek out help with health issues than whites.19 Thus, although there are indications that demographic variables can influence reactions and coping strategies, there are no clear insights as to how these background variables may affect utilization of a CMSS group.

 

Not surprisingly, clinical status as measured by stage of cancer affects how women cope with their breast cancer. For instance, women with later-stage cancer are more likely to be socially withdrawn and more prone to self-criticism.18 Payne et al20 have also described how psychological reactions to breast cancer may vary depending on the stage of disease. Self-reports of mental and physical health states affect how women cope with breast cancer; symptom distress21 has been found to be negatively associated with psychosocial adjustment among women with breast cancer.

 

Despite the generally accepted notion in health psychology that attitudes affect coping styles,22 there has been a surprising lack of research examining how preexisting attitudes affect breast cancer patients' propensity to participate in CMSS groups, which is one of the factors examined in this current study.

 

There is also evidence that social support positively influences how women cope with breast cancer and, in some cases, appears to improve health outcomes. However, research about how preexisting levels of social support may influence an individual's participation in a CMSS group is lacking. Logic suggests that people who are lonelier may be more likely to participate in CMSS. However, Shaw and Yun23 found that women who lived with a spouse or partner were more likely to participate in the Comprehensive Health Enhancement Support System (CHESS) support groups 3 months after receiving the system than women who lived alone.

 

In this study, research questions are employed rather than hypotheses because so little is actually known about the characteristics that influence participation in CMSS groups,24 particularly when people are provided free training, computer hardware, and network access. The specific research questions are listed below:

 

RQ1: What differences in preexisting characteristics are there between active and inactive participants in CMSS groups for women with breast cancer who receive free access and training?

 

RQ2: Are demographic variables associated with volume of participation in CMSS groups?

 

RQ3: Is clinical status associated with volume of participation in CMSS groups?

 

RQ4: Are self-reported physical and mental health indicators associated with volume of participation in CMSS groups?

 

RQ5: Are healthcare-related variables associated with volume of participation in CMSS groups?

 

RQ6: Are social support factors associated with volume of participation in CMSS groups?

 

 

METHODS

Sample

Breast cancer patients were recruited from a large teaching hospital in Madison, WI, a large teaching hospital and two community-based cancer resource centers in Chicago, IL, and a community-based cancer resource center in Indianapolis, IN. In a 5-month experiment, 297 women with breast cancer were randomly assigned to an experimental group that had access to the CHESS "Living with Breast Cancer" Program or to a control group who did not. CHESS was developed at the Center for Health Systems Research and Analysis at the University of Wisconsin-Madison.4,8-10,16,23,25-31 It is designed to be an easy-to-use, interactive computer program that provides brief answers to hundreds of questions about breast cancer, as well as detailed articles and descriptions of services users may need. Users can anonymously ask questions of experts and communicate with and read personal accounts of other women with breast cancer. Decision support tools help patients monitor their health status and risk behaviors, share in important decisions, and plan how to implement their decisions.

 

CHESS was designed with multiple services because it was presumed that different people need different things to cope with disease, and that they learn in different ways.16 Although women do use and value other CHESS services, data consistently show that support groups account for a majority of CHESS use.9 This particular study, which examined characteristics of women who had access to the CMSS group, only looked at the 147 women randomized into the experimental group who had access to and training in how to use CHESS. After three women dropped out of this group with an attrition rate of 2%, the experimental group included 144 women who are included in the analysis below. The participants ranged in age from 30 to 60 years with a mean age of 44.5. Participants had a diverse educational background; 23.8% had a high school education or less, 30.1% had attended some college, 19.6% were college graduates, and 26.6% attended graduate school. The racial characteristics of the sample were 73.8% white and 26.2% were women of color (33 African Americans, two Asians, one Latin American, and two Native Americans).

 

Healthcare providers from participating sites-mostly oncology nurses-introduced the study to patients at a clinic visit following a breast cancer diagnosis. Eligible patients were within 6 months of diagnosis, not homeless, not active illegal drug users, able to understand and sign the informed consent forms approved by the Institutional Review Board, and able to understand and answer sample questions from the pretest survey. Qualifying individuals were encouraged to speak further with a study representative, who described the research project in greater detail. Recruitment rates were 93% in Madison, 84% in Indianapolis, and 61% in Chicago.

 

Before the study began, participants filled out surveys about their health status, attitudes, and preexisting levels of social support. For women in the experimental group, computers were delivered to their homes and they received formal training about how to use the system. In the rare instances that people insisted on using their own computers, CHESS software was installed on their machines before training. Users had access to ongoing technical support (by toll-free number) if they needed it. CHESS was operated as a software program on a computer network rather than on the Internet at the time of this study.

 

Measures

DEMOGRAPHIC VARIABLES

Demographic variables included age, annual household income, education, and race. Due to the preponderance of white women in this sample, the group was dichotomized into white women and women of color.

 

CLINICAL VARIABLES

Stage of cancer was obtained by medical chart review. Stage of cancer is classified by measuring the size, grade, lymph node involvement, and whether the cancer has spread to any other sites.32 Data were categorized into two groups: early stage (stages 0, 1, 2, and DCIS) and late stage (stages 3 and 4). This dichotomy for early and late stage cancer is a generally accepted criterion in the medical community.18

 

MENTAL AND PHYSICAL HEALTH INDICATORS

Emotional Well-being. Six five-point items asked how often participants had felt each of the following in the past 7 days: sad; proud of how I've coped with my illness; losing hope in the fight against my illness; nervous; worried about dying; and, worried that my condition would get worse.33 The scale had an internal consistency [alpha] of 0.80.

 

Negative Mood. Eight five-point items taken from the Positive Affect Negative Affect Schedule (PANAS)34 asked how often participants had felt each of the following in the past 7 days: tense; angry; worried; frustrated; sad; hopeless; anxious; and helpless. The scale had an internal consistency [alpha] of 0.89.

 

Physical Well-being. Seven seven-point items asked how often participants had felt each of the following over the past 7 days: lack of energy; nauseous; trouble meeting needs of family; pain; bothered by side effects of treatment; sick; and, forced to spend time in bed. The scale had an internal consistency [alpha] of 0.84.33

 

HEALTHCARE-RELATED VARIABLES

Perceived Health Competence. Eleven five-point items35 asked how often participants had felt each of the following over the past 7 days: It was difficult for me to find effective solutions for health problems that come my way (reverse coded); I was actively involved in maintaining my health; I had little influence over my health (reverse coded); I was unable to change things I don't like about my health (reverse coded); I handled myself well with respect to my health; I succeeded in the projects I undertook to improve my health; I was generally able to accomplish my goals with respect to my health; I found the efforts to change things I didn't like about my health were ineffective (reverse coded); I was able to do things for my health as well as most; typically, my plans for my health don't work out well (reverse coded); and no matter how hard I try, my health doesn't turn out the way I would like (reverse coded). The scale had an internal consistency [alpha] of 0.82.

 

Breast Cancer-related Concerns. Nine five-point items asked the extent to which respondents felt the following ways in the past 7 days: short of breath; self-conscious about the way I dress; bothered by swollen or tender arms; sexually attractive; hair loss bothered me; worried about the risk of cancer in other family members; worried about the effect of stress on my illness; my change in weight bothered me; and able to feel like a woman. The scale, which had been validated in other studies, had an internal consistency [alpha] of 0.66.33 The relatively lower internal consistency for this scale may be explained because this scale taps into a range of physical, emotional, and body image issues which may not occur all at the same time.

 

Desire for Health Information. Five five-point items developed in our previous research31 asked how often in the past 7 days respondents had felt each of the following: Having information about my illness, treatment, and prognosis gives me a sense of control; I prefer to have all the detailed information (including possible risks) regarding diagnostic procedures and treatment options; it is my responsibility to learn about my health; I only want to have basic health information as opposed to detailed information (reverse coded), and, I seek out health information on my own. The scale had an internal consistency [alpha] of 0.71.

 

Social Information Seeking. Two five-point items asked whether respondents agreed or disagreed with the following statements: I am able to learn from talking to others with breast cancer, and I am able to feel a sense of control by talking to others with breast cancer. The scale had an internal consistency [alpha] of 0.82.

 

Participation in Healthcare. Four five-point items developed in our previous research31 asked how often in the past 2 months respondents had felt each of the following: I went to the right provider at the right time; I understood what was going on; I thought about what was going to happen ahead of time; and I knew the right questions to ask. The scale had an internal consistency [alpha] of 0.75.

 

Relationship With Doctor. Two five-point items asked how often in the past 7 days respondents had felt each of the following: I had confidence in my doctor(s); and, my doctor was available to answer my questions. In the current study, the scale had an internal consistency [alpha] of 0.81. This scale comes from the two-item FACT-B Relationship with Doctor scale, which is a subscale of the Functional Assessment of Cancer Therapy FACT-B scale for patients with breast cancer.36

 

SOCIAL SUPPORT VARIABLES

The Wisconsin Social Support Scale. Six five-point items asked how often in the past 7 days respondents had felt each of the following: there were people I could count on for emotional support; there were people I could rely on when I needed help doing something; my friends supported me even when they disagreed with me; I felt like no one cared for me (reverse coded); I was pretty much all alone (reverse coded); and there was no one I could turn to for help when I needed it (reverse coded). The scale had an internal consistency [alpha] of 0.89.31

 

Social/Family Well-being. Respondents were asked how often in the past 7 days they had felt each of the following: distant from my friends (reverse coded); got emotional support from my family; got support from my friends and neighbors; my family accepted my illness; family communication about my illness was poor (reverse coded); and, felt close to my partner (or the person who is my main support). In the current study, the measure had an internal consistency [alpha] of 0.79.36

 

CATEGORIZATION OF ACTIVE VERSUS INACTIVE PARTICIPANTS

Women who wrote three or fewer messages were categorized as inactive participants, and those who wrote four or more were categorized as active. This distinction was based on two criteria. First, women were required to write one or two messages introducing themselves to the rest of the group during training. Second and third messages tended to be short, containing simple background information such as their diagnosis, and basic personal information such as whether they were married, had children, or where they lived. More substantive interaction with the group generally did not begin until the third or fourth message. Second, the median number of messages across the entire sample was three over the 5-month testing period, so this represented a natural cut in the data.

 

VOLUME/LEVEL OF PARTICIPATION

The volume/level of participation was characterized in terms of the number of words written to the CMSS group. Words were selected rather than individual messages because there was significant variance in message length and so number of words was considered a better measure for volume of writing and level of participation.

 

RESULTS

General Use of Patterns for the Entire Sample

During the 5 months of the experiment, the 144 study participants wrote a total of 472 398 words and an average of 3280 words (SD 6117). The minimum number of words written was zero and the maximum was 29 571 (to provide a frame of reference, the word count for this article including tables and references is 6661 words). Fifty-four percent of the sample wrote three or fewer messages and were categorized as nonactive members of the CMSS group.

 

Predictors of Participation

This section provides two analyses of how antecedent variables affect participation in CMSS groups. The first section employs the entire sample (N = 144) and compares characteristics of active participants and inactive participants, utilizing t tests and Chi-square tests. The second section uses correlation and regression analyses to explore differences in the same set of antecedent variables in the number of words written by the active participants (n = 66) in the CMSS group.

 

Test Comparisons

Table 1 displays the t-test comparisons between active and nonactive participants. Although the active group had a slightly higher education, a greater desire for health information and propensity toward social information seeking, higher health competence, lower levels of preexisting social support and social/family well being, more negative mood, lower emotional well-being and higher functional well being, the only statistically significant difference was that active users were more likely at pretest to consider themselves active participants in their healthcare (t = -2.99, P < .05).

  
Table 1 - Click to enlarge in new windowTable 1

Chi-square Tests

Because race and stage of cancer were created as dichotomous variables, Chi-square tests were used to determine if the number of women of color as active participants was proportional to their ratio of active participants from the entire sample. The difference between observed and expected frequencies on the Chi-square test was not significant. Additionally, comparing early and late stage breast cancer, the difference between observed and expected frequencies on the Chi-square test was not significant.

 

The following tests examined use behaviors among active members and specifically what antecedent variables predicted level of participation among this group. The mean number of words written for the 66 active group members was 7158, and the standard deviation was 7359.

 

Correlation Analyses

Analyses were conducted to examine the association of key demographic, health indicators, healthcare-related attitudes, and social support variables at pretest with volume of words written by active participants over the course of the entire study. Table 2 presents the zero-order correlations between pretest scores and volume of participation among active members.

  
Table 2 - Click to enlarge in new windowTable 2 Zero-Order Correlations of Pretest Scores and Volume of Words Written

Within demographics, a significant correlation was found for race, r = 0.332, P < .01, with white women writing more than women of color. No significant correlations were found between mental or physical health indicators and volume of writing. Among healthcare-related variables, significant correlations were found for relationship with doctor, r = 0.324, P < .01, breast cancer-related concerns, r = 0.321, P < .01, and perceived health competence, r = 0.301, P < .05. Significant correlations were also found between social support, r = 0.279, P < .05, and social/family well-being, r = 0.308, P < .05. All significant correlation coefficients were positive, indicating that, in general, women who were doing better were writing more.

 

Multiple Regression Analyses

Hierarchical regression analyses were used to determine whether the relationships between pretest scores and volume of writing held when controlling for demographics. The first step regressed the effects of demographics, whereas the second step involved individually entering the other antecedent characteristics (ie, health, attitudinal or social support-related variables) that were tested in the correlational analyses resulting in 14 regression models. Betas for predictors of volume of writing, controlling for demographics, are displayed in Table 3. Of the demographics, race was significant; being white predicted writing more words ([beta] = .347). The only significant self-reported health indicator was that having a higher energy level predicted a higher volume of writing ([beta] = .275). Three attitudinal variables showed a trend toward predicting a higher level of words written: a more positive relationship with doctor predicted writing more words ([beta] = .273); fewer concerns about breast cancer predicted writing more words ([beta] = .220); and higher levels of perceived health competence predicted writing more words within the CMSS group ([beta] = .253). Among the social support variables, there was a trend toward social/family well-being being predictive of writing more words ([beta] = .234).

  
Table 3 - Click to enlarge in new windowTable 3 Summary of 14 Regression Equations Predicting Volume of Participation (Words Written) from Pretest Measures Controlling for Demographics

DISCUSSION

The lack of differences in predictive variables between people who are likely to participate in a CMSS group and those who are not gives us some reasons to be optimistic. We had speculated that free access and training would be great equalizers in reducing differences one might expect based on age, race, education, or income in terms of propensity to use a CMSS group. This is precisely what was found in comparing pretest characteristics of active and inactive participants.

 

Hence, stereotypes about who is most likely to use CMSS groups should be questioned as there do not appear to be many distinguishing antecedent characteristics that influence the choice to participate in these groups when free training and access are provided. When cost and access may be prohibitive for particular patients, nurses may help identify free computer access, which is available in many inpatient and outpatient settings.13 Furthermore, where benefits can be demonstrated, nurses may advocate within their organizations to lend computers and temporarily provide Internet access to women diagnosed with breast cancer for a specified period to help them cope, as is done with CHESS.

 

These findings also suggest that providing free training and access is a viable policy option. The lack of differences between active and inactive members in their pretest characteristics has especially important implications for providing support to the underserved who are much less likely to have access to CMSS due to economic or other reasons.

 

Although the underserved will use and benefit from a system such as CHESS,4,30 this does not matter if the underserved do not have access. As earlier researchers have pointed out, many people do not have the access to computers, and this will remain a problem in the future.37 More research is needed on access-related issues, examining the potential benefits of universal access to Internet-based health information and support.38

 

As age was not associated with frequency of use, stereotypes about older people being less likely to utilize such services than younger people should be questioned. To the extent that older people use online health service less often than younger people, the debate should shift to issues related to access, cost, and training, rather than actual willingness to use such systems.

 

That the most active participants in this CMSS group were more likely to report a positive relationship with their doctors is an interesting trend for several reasons. As suggested, some might speculate that people join CMSS groups to find information not available from their healthcare providers, but this study found that doctor/patient relationship was not a significant predictor of whether people given access to CHESS participated in the CMSS group; in fact, active CMSS participants were likely to be more satisfied with their healthcare services. It seems that CMSS groups complement the information provided by healthcare professionals, who clearly have important information to share with patients, particularly about the medical treatment of a disease. However, patients may also gain a great deal by exchanging information with each other about how to cope with breast cancer on a day-to-day basis.

 

The only significant difference between inactive and active participants was higher pretest scores among active members in their self-reported participation in healthcare. This scale tapped into self-efficacy issues, including indicators such as whether they understood what was going on, felt they went to the right provider at the right time, and knew the right questions to ask. People with higher perceptions of self-efficacy are those who are more likely to utilize services that may actually help them become more competent in coping with cancer. To the extent that participation in CMSS groups contributes to positive outcomes, future research and interventions should focus on increasing perceived self-efficacy, and thus, participation in such services.

 

Race affected volume of participation among those who were active members; white women wrote more prolifically than women of color. It is likely that various ethnic groups have different needs resulting from a variety of societal, economic, and cultural influences, and future research should keep this in mind. However, as the most active group of users was almost exclusively white, it is possible that women of color wrote less than the white women because they were not comfortable with the group culture. Future studies should examine whether homogeneity affects the level of participation within CMSS groups and, in fact, there are current studies at CHESS examining this very issue. It is also worth noting that there were so few minority women among active CMSS group participants that any inferences derived from comparing the two groups must be viewed merely as suggestive. A larger sample is needed for more generalizable inferences.

 

Somewhat surprisingly, there was an association between fewer breast cancer-related worries and greater levels of involvement in the CMSS groups. One possible explanation is that women who are more worried about their breast cancer write less because they feel emotionally overwhelmed by their diagnosis, treatments, side effects, and fears related to their own mortality.

 

The association of a higher energy level with greater levels of interaction in the self-selected group of active participants, but not as a predictor of whether those in the randomized sample would choose to participate, is also worth noting. People who chose to participate and had higher energy levels in general understandably had more energy for communication. More energetic women may be better able to adopt adaptive coping strategies to deal with the emotional and physical challenges of breast cancer. Healthcare professionals who note a lack of energy in patients may want to encourage them to seek out ways to increase energy levels. Such options might include modifications in adjuvant therapy regimens, diet, exercise, or lifestyle changes (eg, reducing work- or family-related responsibilities that may push patients to the limits of their available energy resources).

 

Among active participants in the CHESS support group, higher levels of social/family well-being predicted more words written within the CMSS groups. In some ways, this finding appears counterintuitive; one might hypothesize that someone with less everyday social support would be more likely to seek CMSS. However, other research presents findings consistent with this. Using a different data set, Shaw and Yun23 found that women with breast cancer who lived with a spouse or a partner were more likely to use social support services 3 months after introduction to the system, compared to women who lived alone.

 

These individuals may perceive they have more to lose from breast cancer, and therefore are more inclined to communicate about it. This finding agrees with self-in-relation theory,39 which posits that women's strengths lie in their relationships with others as opposed to a more autonomous orientation. Crosby40 also argues that women's identities are more likely to be formed and sustained in relation to the interpersonal relationships that compose their social network. According to these theoretical perspectives, illness is isolating, which is all the more significant to those who have sufficient social networks from which they can feel isolated.

 

Past research supporting the theory that an awareness and sensitivity to other people is one of the significant hallmarks of the psychology of women may also explain the correlation between a higher level of social support and more participation in the CMSS group.41,42 Perhaps because of this, part of what the women disclosed relative to their breast cancer experiences is that they did not understand how or why their husbands or partners were acting in a fashion that they perceived to be emotionally insensitive during their difficult time. In his many years of working with support groups for women with breast cancer, Spiegel43 found that male caregivers tended to express affection by doing things to show that they loved their wives and partners, when the women often just wanted to talk and be heard.

 

One final possible explanation for heavier use of CMSS services by women with more social support is that people who are comfortable receiving support are more competent at getting it.44 It is possible that women who are in an intimate relationship are also more likely to build and maintain other intimate relationships via the social support services in CHESS.

 

CONCLUSION

As CMSS groups have been found to affect outcome variables of interest and are being used by many people, research questions about what kinds of people are most likely to use such a system are important ones. Based on the results from this study, there are some reasons to be optimistic about the relative lack of differences in predictive variables between people who are likely to participate in a CMSS group and those who are not. Because the CHESS program provided computers, network access, and training, these may have been great equalizers in reducing differences one might expect based on age, education, or income. Findings that women with more social support were more active participants in the CMSS groups also counter stereotypes that such groups are used by people who are more socially isolated. Future research should examine whether similar antecedent characteristics that influence level of participation in CMSS groups for women with breast cancer apply to other health concerns as well. Finally, although there is some promising research indicating that breast cancer patients can benefit from participation in CMSS groups, much more research needs to be conducted to understand who is most likely to benefit from such participation as well as acknowledge that CMSS groups could have positive or negative effects on group members.45 Research in these areas can help nurses refer patients to resources that are most likely to improve their quality of care.

 

Acknowledgments

The authors thank Patty Brennan, PhD, RN, for her editorial assistance. This work was done while Dr Arora was working at the University of Wisconsin. Any opinion expressed in this article reflects Dr Arora's personal views and not the official position of the National Cancer Institute.

 

REFERENCES

 

1. Horrigan J. Online communities: networks that nurture long-distance relationships and local ties. Pew Internet & American Life Project. October 31, 2001. Available at: http://www.pewinternet.org/reports/toc.asp?Report=47. Accessed March 24, 2004. [Context Link]

 

2. Fox S, Fallows D. Internet Health Resources. Pew Internet & American Life Project. 2003. Available at: http://www.pewinternet.org/reports/toc.asp?Report=95. Accessed March 24, 2004. [Context Link]

 

3. Science Panel on Interactive Communication and Health In: Eng TR, Gustafson DH, eds. Wired for Health and Well-Being: The Emergence of Interactive Health Communication. Washington, DC: US Department of Health and Human Services, US Government Printing Office; 1999. [Context Link]

 

4. McTavish FM, Gustafson DH, Owens BH, et alet al. CHESS: an interactive computer system for women with breast cancer piloted with an underserved population. J Ambul Care Manage. 1995;18:35-41. [Context Link]

 

5. Weinberg N, Schmale J, Uken J, Adamek M. Therapeutic factors: their presence in a computer-mediated support group. Soc Work Groups. 1995;18:57-69. [Context Link]

 

6. Weinberg N, Schmale J, Uken J, Wessel K. Online help: cancer patients participate in a computer-mediated support group. Health Soc Work. 1996;21(1):24-29. [Context Link]

 

7. Sharf BF. Communicating breast cancer on-line: support and empowerment on the Internet. Women Health. 1997;26(1):65-84. [Context Link]

 

8. Shaw B, McTavish F, Hawkins RP, Gustafson DH, Pingree S. Experiences of women with breast cancer: exchanging social support over the CHESS computer network. J Health Commun. 2000;5(2):135-149. [Context Link]

 

9. McTavish FM, Pingree S, Hawkins RP, Gustafson DH. Cultural differences in use of an electronic discussion group. J Health Psychol. 2003;8:105-117. [Context Link]

 

10. Shaw BR, Hawkins RP, McTavish F, Pingree S, Gustafson DH. Effects of insightful disclosure within computer mediated support groups on women with breast cancer. Health Commun. In press. [Context Link]

 

11. Lieberman MA, Golant M, Giese-Davis J, et alet al. Electronic support groups for breast carcinoma. Cancer. 2003;97:920-925. [Context Link]

 

12. Klemm P, Bunnell D, Cullen M, Soneji R, Gibbons P, Holecek A. Online cancer support groups: a review of the research literature. Comput Inform Nurs. 2003;21:136-142. [Context Link]

 

13. LaCoursiere SP. A theory of online social support. ANS. 2001;24:60-77. [Context Link]

 

14. Cope DG. Functions of a breast cancer support group as perceived by the participants; an ethnographic study. Cancer Nurs. 1995;18:472-478. [Context Link]

 

15. Johnson DJ. Cancer-Related Information Seeking. Cresskill, NJ: Hampton Press; 1997. [Context Link]

 

16. Gustafson DH, Wise M, McTavish FM, et alet al. Development and pilot evaluation of a computer-based support system for women with breast cancer. J Psychosoc Oncol. 1993;11:69-93. [Context Link]

 

17. National Institutes of Health. Cancer Prevention Awareness Survey: Technical Report. Rockville, Md: Public Health Service; 1983. [Context Link]

 

18. Epping-Jordan JE, Compas BE, Osowiecki DM, et alet al. Psychological adjustment in breast cancer: processes of emotional distress. Health Psychol. 1999;18:315-326. [Context Link]

 

19. Gourash N. Help-seeking: a review of the literature. Am J Community Psychol. 1978;6:413-423. [Context Link]

 

20. Payne DK, Sullivan MD, Massie MJ. Women's psychological reactions to breast cancer. Semin Oncol. 1996;23:89-97. [Context Link]

 

21. Budin WC. Psychosocial adjustment to breast cancer in unmarried women. Res Nurs Health. 1998;21:155-166. [Context Link]

 

22. Salovey P, Rothman AJ, Rodin J. Health behavior. In: Gilbert DT, Fiske ST, Gardner L, eds. The Handbook of Social Psychology. Boston, Mass: McGraw-Hill; 1998:633-683. [Context Link]

 

23. Shaw B, Yun GW. Using databases from interactive health communication databases for formative research on program development and inductive theory building: a case study of the CHESS program. Paper presented to: The Science Communication Interest Group of the Annual Conference of the Association for Education in Journalism and Mass Communication; August 2000; Phoenix, AZ. [Context Link]

 

24. Wimmer RD. New Media Research: An Introduction. Belmont, Calif: Wadsworth Publishing; 1987. [Context Link]

 

25. Gustafson DH, Hawkins RP, Boberg EW, et alet al. The use and impact of a computer-based support system for people living with AIDS and HIV infection. JAMIA. 1994:604-608. [Context Link]

 

26. Gustafson DH, McTavish F, Hawkins RD, et alet al. Computer support for elderly women with breast cancer. JAMA. 1998;280(15):1305. [Context Link]

 

27. Gustafson DH, McTavish FM, Boberg EW, et alet al. Empowering patients using computer-based health support systems. Qual Healthc. 1999;8:49-56. [Context Link]

 

28. Hawkins RP, Pingree S, Gustafson DH, et alet al. Aiding those facing health crises: the experience of the CHESS project. In: Street RL, Gold WR, Manning T, eds. Health Promotion and Interactive Technology: Theoretical Applications and Future Directions. Mahwah, NJ: Lawrence Erlbaum Associates; 1997:79-102. [Context Link]

 

29. Smaglik P, Hawkins RP, Pingree S, et alet al. The quality of interactive computer use among HIV-infected individuals. J Health Commun. 1998;3:39-54. [Context Link]

 

30. Pingree S, Hawkins RP, Gustafson DH, et alet al. Will the disadvantaged ride the information highway? Hopeful answers from a computer-based health crisis system. J Broadcasting Electric Media. 1996;40:331-351. [Context Link]

 

31. Gustafson DH, Hawkins R, Pingree S, et alet al. Effect of computer support on younger women with breast cancer. J Gen Intern Med. 2001;16:435-445. [Context Link]

 

32. Love SM, Lindsey K. Dr. Susan Love's Breast Book. Reading, Mass: Perseus Books; 1995. [Context Link]

 

33. Cella DF, Jacobsen PG, Orav EJ. A brief POMS measure of distress for cancer patients. J Chronic Dis. 1987;40:939-942. [Context Link]

 

34. Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect: the PANAS scales. J Pers Soc Psychol. 1988;54:1063-1070. [Context Link]

 

35. Smith MS, Wallston KA, Smith CA. The development and validation of the perceived health competence scale. Health Educ Res. 1995;10:51-64. [Context Link]

 

36. Cella DF, Tulsky DS, Gray G. The functional assessment of cancer therapy: development and validation of the general measure. J Clin Oncol. 1993;11:570-579. [Context Link]

 

37. Rimer BK. Use of multiple media and breast cancer screening: an introduction. J Health Commun. 2000;5:113-116. [Context Link]

 

38. Eng TR, Maxfield A, Patrick K, et alet al. Access to health information and support: a public highway or a private road? JAMA. 1998;280:1371-1375. [Context Link]

 

39. Candib LM. Self-in-relation theory: implications for women's health. In: Dan AJ, ed. Reframing Women's Health: Multidisciplinary Research and Practice. Thousand Oaks, Calif: Sage Publications; 1994:67-78. [Context Link]

 

40. Crosby FJ. Juggling: The Unexpected Advantages of Balancing Career and Home for Women and Their Families. New York, NY: Free Press; 1991. [Context Link]

 

41. Gilligan C. Remapping the social domain: new images of the self in relationship. In: Heller TC, Sosna M, Wellbery DE, eds. Reconstructing Individualism: Autonomy, Individuality, and the Self in Western Thought. Stanford, Calif: Stanford University Press; 1986:237-252. [Context Link]

 

42. Markus HR, Kitayama S. Culture and the self: implications for cognition, emotion and motivation. Psychol Rev. 1991;98:224-253. [Context Link]

 

43. Spiegel D. Living Beyond the Limits. New York, NY: Time Books/Random House; 1993. [Context Link]

 

44. Hobfoll SE, Vaux A. Social support: social resources and social context. In: Goldberger L, Breznitz S, eds. Handbook of Social Stress. New York, NY: Free Press; 1993:685-705. [Context Link]

 

45. Klemm P, Hardie T. Depression in Internet and face-to-face cancer support groups: a pilot study. Oncol Nurs Forum. 2002:45-51. [Context Link]