Effectiveness of Internet vs. Face to Face Support Groups

Institution: University of California, San Francisco
Investigator(s): Morton Lieberman, Ph.D. - Mitch Golant, Ph.D. -
Award Cycle: 2001 (Cycle VII) Grant #: 7BB-1100 Award: $251,763
Award Type: CRC Full Research Award
Research Priorities
Community Impact of Breast Cancer>Sociocultural, Behavioral, and Psychological Issues: the human side

This is a collaboration with: 7BB-1101 -

Initial Award Abstract (2001)
The use of the Internet for delivering health services has seen explosive growth in the past several years. Based on current research we know that 56,000,000 adult Americans go online for help with health problems. Recent surveys indicate that many are from minority communities and the socio-economically deprived. Such growth has not been matched by quality empirical research to understand the implications and potential uses of such internet use.

The study is a collaborative effort investigative teams in two universities, UCSF and Stanford, and a community organization The Wellness Community (TWC) who have worked successfully together for 5 years. We propose a randomized study comparing online (ESG) Wellness Community groups and face-to-face (FTF) support groups to a TWC usual care educational activities group. This will enable us to evaluate the true effects of systematic psychological support delivered in two modalities. Patients will be recruited regardless of their computer skills and after randomization will be offered training and provided E Boxes (Web TV). SeniorNet will provide computer training.

Since TWC is in a position to implement this service quickly and effectively, this study has far-reaching implications. A cost analysis found that TWC FTF is 5 times the cost of their ESG groups. TWC has 20 facilities nationwide serving cancer patients. If found effective, online support groups would be made available to many more women than are served currently, particularly underserved rural populations. As the first random trial of Internet intervention, we believe that the findings may enable increased use of support services for a variety of health problems. The manual for leading online support groups, although based on breast cancer, could be a significant starting point in other health-related Internet support groups.

The CBCRP IDEA I Pilot found: a) women with breast cancer can be successfully recruited and retained for ESG’s; b) two-thirds of the women reside in rural areas or small towns; c) participants showed significant decreases in depression, expressed more zest for life, saw more new possibilities for their future, deepened their spiritual lives and substantially decreased their negative reactions to their cancer pain; and d) critical group processes were isolated that help identify characteristics associated with successful groups.

Three broad questions are asked in this research: 1) Do online TWC support groups improve quality of life? 2) What are the psychological characteristics of women who benefit? and 3) What are the processes of online groups that facilitate beneficial changes?

To answer these broad questions a series of hypotheses will be tested, the primary one being that participation in both TWC support settings will improve their attitudes and beliefs about their illness, quality of life, decreased marital strain and levels of depression compared to the usual-care group. We will recruit 240 primary (stages 1-3) breast cancer patients to participate in this randomized trial for 20 weeks. Measures will include: Efficacy: Depression, Self-Efficacy, Quality of Life, Marital Strain, Personal Growth. Mediators: Personality traits. Moderators: Cancer-Related Coping, Emotional Control, Repression, and Illness Level. Process measures: Leader Behavior, Group Norms, Cohesiveness, Helpful Experiences, Group Content, Analysis of Group Interaction using two computer linguistic dictionaries.


Final Report (2006)
The overall aims of this randomized study were to: I. Evaluate efficacy of face-to-face (F2F) and Online Internet support groups (OSG). We hypothesize that participation in both TWC’s F2F and OSG support will improve attitudes and beliefs about their illness, and quality of life, decrease marital strain and levels of depression. II. To test mediators of group support in both the FTF and OSG.

We were not able to recruit a sufficient number of women to accomplish both of these aims. Details of our recruit efforts and statistics on the sample we studied are provided in the scientific report. An analysis of the OSG vs. F2F groups using the FACTB scale to measure quality of life, the CESD scale to measure depression, PTGI scale to measure emotional growth, and MINIMAC to meassure coping and pain. No significant differences were found in changes over time for both groups or the groups (F2F vs. OSG) on the FACTB, CESD AND MINIMAC measures. The sample size did not permit an analysis for Aim 2. During the course of our work, we soon realized that recruitment was a serious problem and made a number of changes in how we recruited. We also set out, using some of the resources of the BCRP grant (not the field staff) to embark on a series of studies that would broadly expand our knowledge about OSG, particularly peer-facilitated BC groups. These studies are summarized below.

First in importance was the finding that peer facilitated large internet groups for women with breast cancer result in substantial positive improvement in quality of life and lowered depression. Our analysis of peer led cancer OSG suggested that one possible mechanism for this success is the ability of women to search among such groupsfor one that they believe matches their needs through the mechanism of “lurking” before posting a message. Much of our effort during the BCRP grant has been directed towards isolating characteristics of the participants and the group conditions or processes that lead to positive outcomes of the group members. We found that the expression of specific negative emotions in the group is associated with positive change and that professionally facilitated OSGs are more likely to enhance the frequency of such expressions. We found that certain therapeutic factors (helpful group experiences) are linked to positive change. We also found that the norms of the groups (the regulatory group culture) enhanced these HGEs. Leader behavior made a difference and certain types of behavior are more successful in creating a group where participants make positive change. This leader behavior works by enhancing the group’s cohesiveness. When we compared F2F and OSG we found that by and large they are more similar than different in both processes and content of discussion. In a study of those who drop out of OSG, we found that they experience at the outset of the group more fear and anxiety about the group. Finally, in looking at various non-gender related cancers, we found that men rarely gave expression to the critical negative emotion compared to women with the same cancers.

These studies have provided invaluable information on how OSG can be improved to maximize their effectiveness.


Progress Report 2 (2004)
This year we have been focused on recruitment and the development of new studies. To date we have enrolled 90 women in the randomization trial.

With respect to the effectiveness of self-directed internet newsgroups, we expanded our research to Internet Bulletin Boards. Bulletin Boards provide a forum for people to communicate. A common theme is that all boards are open discussions with a large community of BC patients. We recruited 114 women with breast cancer. A six-month follow-up generated relies from 80% of the original sample. A repeated measures multivariate analysis of variance was performed using time since diagnosis as a control for each measure. The analysis showed statistically significant positive changes over the six months of study. These results were not effected by the length of time the women had cancer. We found lower depression (using the CES-D scale) quality of life (using the FACTB) and more personal growth (using the PTGI). The findings led us to extend this study, and we are now collecting long term data- from one to one and one-half years since the women joined the newsgroup

The same sample of women was used to test a critical mechanism of BC support groups: the expression of negative emotions. The repression/suppression of negative emotions has long been considered detrimental for breast cancer (BC) patients, leading to poor coping, progression of symptoms, and general lower quality of life. Using a variety of text analysis programs we found that the group of women who showed that greater expression of anger predicted higher quality of life and lower depression, while the expression of fear and anxiety predicted lower quality of life and higher depression. The expression of sadness had no predictive utility. Our results challenged the commonly held belief that the expression of all negative emotions is beneficial for BC patients. Instead expressing some negative emotions are beneficial, while expressing others are not.

Our research group has also been active in extending the internet studies for cancer patients. The Wellness Community Virtual Community, a direct outgrowth of previously funded BCRP pilot studies has grown considerably in the past year in both the number of participants and the range of services provided cancer patients, about 25% of them are BC patients. Caregiver services have been added, and we now have two active caregiver support groups as well as 5 patient groups. Many of the visitors to our site, about 4,000 monthly, chose not to join the support group but avail themselves of the educational and stress reduction activities. The methods we have developed used in the BCRP supported research, outcomes, process and text analysis have been applied to the VC groups.

The collaboration we described last year with NYU’s study, online support groups for immigrant Hispanic women is currently collecting Data on their first online group. Finally, a new study, providing online TWC led support groups for adolescent cancer patients and their caregivers has been successfully launched.


Symposium Abstract (2003)
Investigators have linked repression or suppression of negative affect with the incidence and progression of cancer. Implied is that the expression of negative emotions is beneficial. The empirical evidence that the frequency of negative emotions in women with BC is linked to their well being has been studied in a variety of ways. The direct exploitation of this finding in psychotherapy lacks important research, since the studies are based on input-output information. Needed are studies that go beyond the “black box” and generate data on the behaviors of patients in the therapeutic setting itself.

The current studies central question, the comparisons between professional and lay led internet support groups. Eleven groups were examined, 7 led by professionals and 4 by lay leaders. The 11 groups were composed of 4 TWC, 3 Bosom Buddies, 1 CHESS and 3 bulletin boards. The first three groups had closed group membership. The professionally moderated groups included The Wellness Community chat rooms and an asynchronous newsgroup, Bosom Buddies. The lay led groups included CHESS and 3 open participation (“bulletin board”) lay-led asynchronous support groups (Yahoo, IVillage, and WebMD).

The expression of negative emotions in participants’ interaction was scored using the PCAD, a computer text program. Extensive empirical research has established the validity and reliability of scales measuring a variety of emotional states including Anxiety, Hostility-Anger, and Depression.

Professionally led and lay led internet groups were compared using three multivariate analysis of variance. All 3 were highly significant. (Anxiety F=7.1) (DF, 6,166) P=.00; Hostility F=20.5) (DF, 5,167) P=.00; Depression F=19.7) (DF, 7,165) P=.00. Members in the professionally led groups expressed on Anxiety, significantly more death, mutilation, guilt, shame and diffuse anxiety. On the Hostility scales they expressed more overt hostility as wells as higher covert, inwardly directed and ambivalent hostility. On the Depression scales members of professionally led groups expressed more hopelessness, self accusation, psychomotor retardation, somatic, separation and hostile depression.

Groups conducted by professional leaders express significantly more negative emotions than those led by lay leaders. Our analysis of positive emotions indicates that this difference is not simply the product of amount of emotional expression. A review of the interventions by leaders points to a glaring difference in how professional and lay leaders respond to members when they bring up such negative feelings. The professional leaders encourage the further expression of such emotions, believing that such expressions are therapeutic. The lay leaders clearly do not signal that such feelings are illegitimate, but they are more likely to offer support and suggest fighting the cause of these feelings.


Symposium Abstract (2005)
Morton A. Lieberman Ph.D. and Benjamin A. Goldstein B.A.

For women with Breast Cancer (BC), the combination of the illness and its treatment often leads to depression, anxiety, and changes in physical appearance, causing a decline in emotional quality of life. Many turn to support groups for help with the psycho social sequella of the illness. The advent of on-line communications and the rapid spread of Internet access have given rise to both self-directed as well as professionally led support groups for a variety of health problems. Of the 165 million Internet users in the United States, approximately 15 million have visited Internet health support groups. The most common format for seeking Internet support is on public bulletin boards (also known as newsgroups, message boards, and Listserv). Bulletin boards (BBs) allow the user to Public BBs found on the Internet are characterized by having 100s of members, as being self-directed, and are peer led. They are open discussions containing multiple “threads” with support provided by others having the same medical condition. Most importantly, these boards are readily available at no cost to the users.

Women were recruited when they joined one of five breast cancer BB’s, they completed questionnaires online and then again 6 months later (N = 114) Measures included depression (CES-D), growth (PTGI) and psycho-social well-being (FACT-B). The sample’s mean age was 45 years. Participants joined the board usually within 6 months from their diagnosis. Twenty-three percent of the women turned to professional psycho-social help prior to their participation in the BB. Forty-eight percent had attained a college or graduate degree, while 6% had attained only a high school diploma. Finally, 15% were identified as stage 0 (DCIS), 21% of as stage I, 33% and 25% as stage II and III respectively and 6% as stage IV, providing a distribution of breast cancer diagnoses.

A repeated measure MANCOVA was performed using time since diagnosis as a covariate for each measure. Overall follow-up was significant ( P=.001, eta=.37) while time since diagnosis failed to produce an effect as a covariate. Depression, (P=.005, eta=.17.) psycho-social well-being (P=.001, eta=.33) and growth (P=.05, eta=.08) were significant at follow-up with moderate to strong effect sizes. To further test the hypothesis, using the standard cutoff for possible clinical depression (CES-D > 16), a chi square was computed to see if participants improved from depressed to not depressed. Thirty participants met criterion for clinical depression at baseline and of those, 15 were not depressed after 6 months, X2 = .023.

Improvement was statistically significant on all three measures. This serves as a first validation of Internet bulletin boards as a source of support and help for breast cancer patients. These boards are of particular interest because they are free, accessible, and support comes from peers and not from professional facilitators.

Electronic support groups for breast carcinoma: a clinical trial of effectiveness
Periodical:Cancer
Index Medicus: Cancer
Authors: Lieberman MA, Golant M, Giese-Davis J, Winzlenberg A, Benjamin H, Humphreys K, et al
Yr: 2003 Vol: 97 Nbr: 4 Abs: Pg:920-5