Effect of Computer Support on Younger Women with Breast Cancer]]>Logo of jgimedspringer.comThis journalToc AlertsSubmit OnlineOpen Choice ]]>
J Gen Intern Med. 2001 July; 16(7): 435–445.
PMCID: PMC1495237

Effect of Computer Support on Younger Women with Breast Cancer



Assess impact of a computer-based patient support system on quality of life in younger women with breast cancer, with particular emphasis on assisting the underserved.


Randomized controlled trial conducted between 1995 and 1998.


Five sites: two teaching hospitals (Madison, Wis, and Chicago, Ill), two nonteaching hospitals (Chicago), and a cancer resource center (Indianapolis, Ill). The latter three sites treat many underserved patients.


Newly diagnosed breast cancer patients (N = 246) under age 60.


Experimental group received Comprehensive Health Enhancement Support System (CHESS), a home-based computer system providing information, decision-making, and emotional support.


Pretest and two post-test surveys (at two- and five-month follow-up) measured aspects of participation in care, social/information support, and quality of life. At two-month follow-up, the CHESS group was significantly more competent at seeking information, more comfortable participating in care, and had greater confidence in doctor(s). At five-month follow-up, the CHESS group had significantly better social support and also greater information competence. In addition, experimental assignment interacted with several indicators of medical underservice (race, education, and lack of insurance), such that CHESS benefits were greater for the disadvantaged than the advantaged group.


Computer-based patient support systems such as CHESS may benefit patients by providing information and social support, and increasing their participation in health care. These benefits may be largest for currently underserved populations.

Keywords: breast cancer, quality of life, patient participation, computer, patient education, disadvantaged, Digital Divide

Breast cancer, the most common cancer among women, accounts for 29% of all new cancer diagnoses in the United States, and is the leading cause of cancer death among women between 20 and 59 years old. The physical effects of breast cancer and its treatment (depression, fatigue, hair loss, weight gain, and changes in physical appearance) can have a devastating impact on quality of life,14 affecting women's self image and ability to carry out roles within and outside the home. Ability to deal with complex information, obtain social support, and participate (as desired) in one's care become crucial in minimizing that devastation.5,6

The impact of breast cancer is even more serious among African-American women, whose breast cancer-related mortality rates are higher than for Caucasians.1,7 Moreover, medical underservice (often associated with minority status and with less health insurance, income and formal education) leads to fewer opportunities to participate in care or obtain needed information and social support.810

An intervention that improved information seeking skills, social support, and patient participation in health care might improve psychosocial aspects of quality of life.1113 Computer-based systems provide a venue for such an intervention, because multiple services (e.g., rapid access to information, support groups, and decision support systems) can be integrated to support each other. Furthermore, in-home systems are available to patients when and where needed. Because the medically underserved often lack other resources, such services should have greater impact for them than for traditionally better-served groups.

This paper reports on a randomized controlled evaluation of one such system, Comprehensive Health Enhancement Support System (CHESS), developed at the University of Wisconsin-Madison. In a previous randomized trial, HIV-infected patients used CHESS extensively, and access to CHESS reduced mood distress and improved cognitive functioning, social support, and active involvement in life.14 Unpublished anecdotal comments by clinicians suggested that CHESS users also participated more effectively in their care. Qualitative analyses found that those who improved most had done so by using the CHESS information and decision components in thoughtful and connected ways for moderate amounts of time, rather than by spending large amounts of time in discussions with other patients.15 Underserved minority HIV patients used CHESS overall as much as middle class Caucasians, but participated less in discussions, while making heavier use of information and decision services.16,17 Although CHESS effects did not show significant interactions with race for HIV patients, these two findings suggest that such interactions might occur.

Pilot tests of CHESS with breast cancer patients indicated widespread acceptance and use of CHESS, including by low-income African-American women.18,19 Thus, we proposed that:

  1. CHESS would affect breast cancer patients' perceptions of six outcomes: 1) social support, 2) unmet need for information, 3) competence dealing with that information, 4) participation in health care, 5) confidence in doctor(s), and 6) psychosocial quality of life.
  2. Effects of CHESS would be greater for the medically underserved, primarily indexed by race.


Subjects and Recruitment

Between April 1995 and May 1997, this randomized controlled trial accrued younger (age 60 years or less) breast cancer patients treated at two large teaching hospitals in Madison, Wis, and Chicago, Ill, two other hospitals in Chicago, and a community-based cancer resource center in Indianapolis, Ind. The last three treated mostly low-income minorities and were sought out to meet study goals of including 100 minority patients (mostly African American with lower income, less education, and no health insurance) among the total sample of 300.

Nurses and doctors at the participating organizations introduced the study to breast cancer patients at a clinic visit following diagnosis. Eligible patients were within six months of diagnosis, not homeless, not active illegal drug users, able to give informed consent, and able to understand and answer sample questions from the pretest survey. All qualifying patients were encouraged to speak with a study representative who described the study. Consent was documented by obtaining signed, University of Wisconsin IRB-approved consent forms.

Recruitment rates were 93% in Madison, Wis, 84% in Indianapolis, Ind, and 61% in Chicago, Ill, where CHESS was less integrated into normal practice at the two nonteaching hospitals and staff may have been less able to take time to offer CHESS to patients. Among patients contacted, the most frequent reasons for refusal to participate included 1) no time, 2) didn't want to make another decision, 3) no interest in using a computer, 4) didn't want more information, and 5) unwilling to be randomized.

We accrued 295 women with newly diagnosed breast cancer. After dropouts, 246 women remained, including 55 African Americans and 9 other minorities (see Figure 1 for the study flowchart). We called dropouts and found that being too busy and feeling too ill were the most frequent reasons for dropping out after randomization. Other reasons included wanting to move beyond breast cancer, disappointment at not being assigned to receive CHESS, and just changing their mind.

Study profile.


At the time of the study, CHESS programs resided on a home computer, with modem connections to a central server for communication. (See chess.chsra.wisc.edu for a demonstration of the system, since migrated to the Internet.) CHESS content was developed by clinician and patient panels to respond to needs identified by interviews and prioritized with surveys of 400 breast cancer patients and families.20 Designed to be attractive and easy to use, CHESS integrates eleven services to provide breast cancer information, decision-making tools, and support services.

  • Information Services: Questions & Answers (short answers to hundreds of frequently asked breast cancer questions); Instant Library (full articles drawn from the scientific and popular press); Consumer Guide (descriptions of health services, identifying good providers and being an effective consumer); and Referral Directory (descriptions of and ways to contact local and national breast cancer services).
  • Support Services: Discussion Groups (limited-access, facilitated bulletin boards for sharing information and support); Ask an Expert (patients receive a confidential response from experts); and Personal Stories (how others facing breast cancer coped).
  • Decision Services: Assessment (measures a person's emotional status and offers coping advice); Health Charts (patients record and track health changes); Decision Aid (learn about options, clarify values and understand consequences of choices); and Action Plan (plan successful changes by identifying goals, resources and ways to overcome obstacles).

Study Administration

After completing the consent form and pretest, patients were randomly assigned to the control or experimental group. To ensure consistency across sites and provide subjects some benefit for participation, the control group received a copy of Dr. Susan Love's Breast Book.21 Those in the experimental group had CHESS at home for six months (a period long enough for most patients to finish treatment). CHESS staff delivered a computer, installed it, gave brief training (less than one hour), and left a user's manual.

Subjects completed a pretest at the clinic and two- and five-month post-tests and were compensated $25 for each. Post-test surveys were mailed to subjects and returned in pre-addressed stamped envelopes. If a subject did not respond within two weeks, a follow-up phone call was made and a second copy of the survey mailed. Surveys were usually entered into the database within a week of arrival at the central office. If missing data were found, site coordinators phoned to obtain remaining answers.


Patient Characteristics

Sociodemographic characteristics (age, race, income, education, and insurance status) were assessed at the pretest survey; disease-related information (diagnosis date, type and date of surgery, and schedule for radiation and chemotherapy) was collected on post-test surveys and validated by medical record reviews conducted by trained research staff. All surveys collected Karnofsky Performance Status22,23 to assess impairment of normal activities. Staging data (tumor size, nodal status, and metastasis) were abstracted from medical record reviews, and the TNM classification scheme applied. For analysis, we classified patients with breast cancer at stage 0, I, and II as early and those at stage III and IV as late stage.

Patient Outcomes

For convenience here and in tables, patient outcomes are presented in three groupings, with explanation of conceptual independence or nonindependence for multiple comparison purposes.24

Social Support and Information Measures

A six-item social support scale (Cronbach's α = 0.89)14 assessed a woman's perception of available emotional and instrumental support (see Appendix for sample items). A three-item unmet information need scale (α = 0.83) assessed the extent to which a woman lacked information about her health or health care. A five-item information competence scale (α = 0.80)25 assessed a woman's level of competence seeking information.

Participation-related Measures

A four-item scale (α= 0.77)14 assessed how much a woman perceived herself involved in her health care. Two items (α = 0.68)25 assessed comfort asking clinicians questions and confidence making major decisions about her health. These two scales are closely related outcomes and will divide α in analyses. Level of confidence in her doctor(s) came from a single item in the Functional Assessment of Cancer Therapy-Breast (FACT-B) instrument.2629

Quality of Life Measures

The FACT-B instrument satisfied assessment guidelines30 by tapping several psychosocial dimensions of QOL with test-retest reliability and sensitivity to change.2629 Six items assessed social/family well-being (α = 0.79). Six other items assessed emotional well-being (α = 0.80). And seven items assessed functional well-being in everyday life (α = 0.84). The FACT-B also includes a measure of nine breast-cancer-specific concerns. While these assess potentially independent problems (and thus have lower scale reliability; α = 0.66), a scale of them reflects the degree to which a patient was worried about treatments, side effects, and recurrence. These four QOL measures are closely related and will divide α (.05) in analyses.

All scale scores were linearly transformed to a 0–100 scale with a higher score indicating more positive outcomes, e.g., more support, greater participation, less unmet information need, and fewer concerns.


Tests of intervention effects used an intention-to-treat approach that compared the original randomized groups regardless of individuals' amount of use of the experimental intervention. To test whether CHESS enhanced patient outcomes over time, separate multiple linear regression models31,32 predicted the two- and five-month post-test scores with experimental assignment (CHESS or control) as the main predictor, controlling for the pretest score. In addition, because health outcomes may be affected by time since diagnosis, stage of disease, and recent (within one month) chemotherapy, we also controlled for these and the individual's Karnofsky score. Finally, because sociodemographic variables may influence a patient's quality of life scores and access to services, we controlled for patient's level of education, race, and whether the patient had private health insurance.

To test whether medically underserved patients benefited more for each outcome variable, a subsequent model tested the interaction of experimental assignment (CHESS or control) with race. Although study design and sampling treated race as the primary indicator of medical underservice, some African-American women in this sample were not socioeconomically disadvantaged, while some Caucasian women were. Therefore, we also tested two additional interactions with insurance (private insurance vs no insurance) and education. Because race, education, and insurance status were moderately correlated (average R =.33), their interactions were added separately to the main effects model.

Statistical significance is reported for P ≤ .05, two tailed. For the conceptually-related measures identified above, the Holm procedure will be employed to divide α among multiple comparisons.33,34 Our sample size of 246 subjects (121 CHESS, 125 control) at an α = 0.05, provided a 0.80 power to detect a 0.36 SD difference between the CHESS and control groups (between small and medium effect sizes, as defined by Cohen35). For the interaction of experimental assignment with race, we have 0.80 power to detect an interaction in which the women of color benefit more than Caucasians by 0.82 standard deviations (a large effect).


Table 1)shows that the CHESS and control groups did not differ significantly on any sociodemographic and clinical characteristics. Women in the CHESS group averaged 155 uses of the system during the 26 weeks, or almost 6 times per week, with an average of 13 minutes per use. Caucasian women and women of color used CHESS equally frequently (5.9 and 6.2 times a week, respectively), but the average duration of use per episode was much higher for Caucasian women (14.6 vs 7.7 minutes). This difference came from Caucasians' greater use of Discussion Group (the bulletin board exchanges with fellow patients), which accounted for 68% of Caucasian women's use, but only 38% for women of color. Use of CHESS information services was equal for the groups, although women of color used the self-tracking and decision services considerably more than Caucasian women (over four hours during the study vs just under one hour). Because decision services demand more of users and are most directly linked to behavior changes,15 the use of this service by women of color could provide a mechanism for stronger CHESS effects.

Table 1
Characteristics of Patients Who Completed the Clinical Trial (N = 246)

Main Effects

After two months of implementation, the CHESS group scored 4.8 points (95% confidence interval [CI], 1.5 to 8.1) higher on competence dealing with information than controls (Table 2)Similarly, the CHESS group scored significantly higher than controls on two of three measures of interactions with the health care system: 6.4 points (95% CI, 2.1 to 10.7) on level of comfort with their participation (for which the multiple comparison significance level is 0.025) and 5.7 points higher (95% CI, 1.0 to 11.3) on confidence in their doctor. However, there were no significant main effects of CHESS among the four quality of life measures.

Table 2
Group Differences on Support, Participation, and Quality-of-life Variables at Two-month Follow-up — Results from Regression Models (N = 246)*

Five months after the study began (Table 3)), CHESS subjects scored 4.9 points (95% CI, 1.4 to 8.4) higher than controls on social support and, similar to results at two months, 3.5 points higher (95% CI, 0 to 6.9) on information competence. While differences on the three participation measures were all still positive, none was significant. Again, there were no significant main effects for quality of life indicators.

Table 3
Group Differences on Support, Participation, and Quality of Life Variables at Five-month Follow-up-Results from Regression Models (N = 246)*

Interactions with Indicators of Medical Underservice

Although the limited number of women with color only provided sufficient power to reliably detect large interactions, several were found, and a number of others presented similar patterns. Thus,Figures 2 and and33 plot all interactions with P < .05 to allow examination of a consistent pattern, even though some are not statistically significant after controlling for multiple comparisons. After two months, significant interactions with race and education were present for two variables for which there had not been significant main effects: information need and behavioral involvement in one's health care. For overcoming the perception of unmet information needs, CHESS helped minority women more than Caucasians (Figure 2). This interaction was repeated when education was substituted for race, with education more strongly related to information need for controls than for the CHESS group. Similarly, for women participating in their own health care, CHESS again helped minority women more than Caucasians, and education was related to participation more for controls than for the CHESS group (Figure 2).

Interactions for participation and information need at two-month follow-up.
Interactions for quality of life measures at two months and for information competence and functional well-being at five months. Due to multiple comparisons, significance required P < .0125 for the two functional well-being interactions and P ...

For interactions involving quality of life measures (Figure 3), access to private health insurance appeared more important than race or education, although controls for multiple comparisons across four QOL variables meant that these interactions did not reach significance. Being assigned to CHESS was more related to functional well-being for women without private insurance than those with private insurance (P = .02, where significance required .0125). Women who were assigned to CHESS reported fewer breast cancer concerns than did controls, particularly for those without insurance (P = .025, where significance required .0167). Two additional interactions for quality of life variables (with race for breast cancer concerns, and with insurance for emotional well-being) had similar shapes and P values less than .10 (.06 and .065, respectively), and perhaps widen the pattern to potentially involve three of the four QOL variables.

At five months, there was only one significant interaction (Figure 3), but it was again similar to the above. Education was related more strongly to information competence for controls than for the CHESS group. And women who had CHESS reported better functional well-being, particularly for those without insurance (P = .036, where significance required .0125).


Based on needs assessment studies,13 CHESS was designed to help breast cancer patients and their families deal with information, be more active and confident participants in their health care, increase social support, and better maintain their socio-emotional quality of life against the impact of a breast cancer diagnosis and its treatment. And because minority women may have limited access to health care information and other resources, benefits for them are potentially even greater.

Our results indicate that CHESS improved social support, information competence, participation in health care, and quality of life in this group of patients. After two months of implementation, there was either a significant main effect or interaction with an indicator of medical underservice for information needs and competence, and all three measures of health care participation. No significant main effects were found among four quality of life measures, but there was a consistent pattern among the interactions matching significant ones found elsewhere. CHESS' mixture of disease information, self-monitoring and decision tools, along with shared experiences and support from other patients, may help patients tolerate treatment better, have more positive feelings about treatment, and therefore may improve outcomes as well.

The benefits of CHESS were stronger in the short run. Five months after implementation, benefits of CHESS were considerably less consistent. There were still significant increases in information competence and social support measures, but no difference for health care participation (possibly because there is less contact with the health system at this time) or quality of life. Thus, the benefits of a consumer support tool like CHESS may dissipate over time. However, inspection of group means in Tables 2 and and33 suggests instead that control subjects eventually improved on their own between two and five months, as time since diagnosis and treatment passed, while the CHESS group changed little beyond their two-month scores. Given the strains on breast cancer patients around the times of diagnosis and treatment (corresponding to the first post-test), CHESS appeared to help patients when they most need it.

There were several interactions of indicators of medical underservice with the CHESS intervention: greater benefits for women of color, the uninsured, and those with less education. Note that race was not always the best locator of these interactions; low education and lack of health insurance were equally or more germane in identifying women who benefited most. Because the underserved have fewer resources, any incremental resource may have the potential to make a bigger difference for them.

Although we predicted such interactions, what is always at issue is whether the disadvantaged will have the ability and motivation to take advantage of that potential.17 Because CHESS (and other consumer health support services) relies heavily on written information, literacy and ease of reading are potential barriers for the underserved. In addition, the disadvantaged may be less experienced and comfortable with computers. Our results encourage optimism that computer technologies can assist the disadvantaged, because motivation provided by a life-threatening disease may overcome barriers to use. Disadvantaged women gain relatively more benefit because alternative resources are equally available to advantaged women in both experimental and control groups. Moreover, minority women made less use of the Discussion Group, but more use of self-monitoring, planning and decision services, which we argued earlier may be more beneficial.

The clinical or personal importance of the changes reported here is a more complex issue, because no clear benchmarks exist for the information or health care participation variables. For the quality of life variables, however, two studies have demonstrated that 7.1–10.7 unit (translating to the 0–100 scale used here) changes in the functional well-being scale were associated with meaningful improvements in hemoglobin for patients treated for a variety of cancers.36,37 None of the main effects for this measure or the other quality of life measures developed in the same way by the same researchers2629 were significant. If the same standard were applied to the information and health care participation measures for which significant main effects were observed, one could say that those main effects ranged in size from one-half to the lower bound of the clinically important range. More importantly, although the estimates are less reliable for the smaller sample size, the size of effects for the medically underserved (women of color, lower education, no private insurance) are consistently in or larger than this range, including two effects on functional well-being and one other measure of quality of life. Thus, it would appear that the effects observed here for the medically underserved probably are personally and clinically, as well as statistically, significant.

Our results are probably generalizable to the U.S. population of younger women with breast cancer, although there are some remaining limitations. The ethnic diversity of the patients in the study (74% Caucasian, 22.4% African American, and 3.6% other minorities) is similar to that of the U.S. (72.2% Caucasian, 12.5% African American, and 15.3% other minorities).38 The study sites included both public and private hospitals, some of which were teaching hospitals. Patients came from urban, small-city, and rural areas and included all socioeconomic strata, although all were from the Midwest. We did not include older women with breast cancer in this study. However, a recent CHESS evaluation in Medicare women found similar levels of acceptance and use.39 In addition, patients treated in small, rural general surgeries were excluded, because including many clinics with few patients each would have been prohibitively expensive. Nonetheless, reactions to CHESS by patients at small clinics could conceivably be different from those at the specialized clinics studied here.

Our decision to compare CHESS to a control condition employing Dr. Susan Love's Breast Book creates a potential limitation to generalizability. Although we did this so that control-group patients from five different clinics would have access to a comparable base of information, this could produce a conservative bias in experimental group comparisons. If so, actual implementations of CHESS might produce slightly larger effects than those seen here. Conversely, one could argue that some benefits (particularly social and information support) could result simply from being loaned a computer, rather than the particular content it contained. We are skeptical of that explanation, given the heavy use of CHESS, but a current study is comparing CHESS to mere Internet access.

A final limitation is that the study reports aggregate effects for a support system, having many components designed to work together. While this approach was deliberate, and derived from the variation of patient needs,18,20 we cannot say whether specific portions of CHESS were responsible for these effects. Assigning each of a dozen or so CHESS components to different groups could address this. However, we feel this would be a disservice to the patients involved by depriving each of the majority of services, and the ability to integrate material from different services. However, given interactions with indicators of medical underservice, lower use of Discussion Group by women of color, and the match this presents with prior work on HIV-CHESS,15 one might wonder whether CHESS could function effectively without the Discussion Groups (which accounted for 55% of all use). The ongoing study mentioned above includes a reduced-CHESS condition without the Discussion Group and Ask an Expert components.

Future research should compare CHESS against alternative support services of comparable or less cost. Many professional programs (such as group psychotherapy for women with breast cancer) exist for people with life threatening illnesses. How would they compare to CHESS and how would a combination of CHESS and such programs affect women with breast cancer?

Although this study was conducted while CHESS was a stand-alone program, it has now migrated to the Internet, potentially changing its accessibility and cost. More people are now gaining access and using the unregulated and possibly overwhelming world-wide web.40 In fact, nearly 75% of all breast cancer patients now offered CHESS already have Internet access. However, although this research suggests that the underserved may benefit most, they currently are least likely to have such access.41,42

Costs of a wide scale dissemination of any interactive health communication (IHC) system depend on the implementation strategy employed. One could provide the system in the same way as most web sites: offering it to anyone with Internet access. Alternatively patients could be actively recruited to and supported in their use of the IHC.

In that case, costs would be divided into those associated with “marketing” the IHC to patients and costs of operating the IHC once patients accept it. Assuming three-month access, we estimate that a system like CHESS might be delivered for approximately $150 per patient, including purchasing computers to loan. Delivery would be cheaper for patients who have Internet access, but relying on this strategy would exacerbate digital divide problems. An aggressive recruiting effort would increase costs. However, given the benefits observed here, and their concentration among those most in need of help, we believe an aggressive strategy is justified.


This research was supported by the National Institute of Child Health and Human Development and the National Cancer Institute under grant #5R01HD32922. We wish to acknowledge the very important contribution to this project made by Haile Behre, the lead programmer for the CHESS projects.


Patient Outcome Measures (With Example Items)

Social and Information Support

  1. A six-item scale of perceived emotional and instrumental support (Cronbach's α = 0.89).14
    “There are people I can count on for emotional support.”
    “There are people I can rely on when I need help doing something.”
  2. A five-item scale of competence at seeking information (Cronbach's α = 0.80).25
    “At this time, I can figure out how and where to get the information I need.”
    “I feel that I am in control over how and what I learn about my health.”
  3. A three-item unmet information need scale (Cronbach's α = 0.83).
    “There is information I need but don't know how to get.”

Participation in Own Health Care

  1. A four-item scale of perceiving oneself involved in one's health care (Cronbach's α = 0.77)14
    “I went to the right provider at the right time.”
    “I thought about what was going to happen ahead of time.”
  2. A two-item level of comfort participating in one's own health care scale (Cronbach's α = 0.68).25
    “I feel comfortable asking my doctor questions about my care.”
    “I feel confident about my ability to make major decisions about my health.”
  3. A single item assessed a woman's level of confidence in her doctor(s).26
    “I have confidence in my doctor(s).”

Psychosocial Quality of Life26

  1. A six-item scale of social/family well-being (Cronbach's α = 0.79).
    “Family communication about my illness was poor.”
    “I felt close to my partner.”
  2. A six-item scale of emotional well-being (Cronbach's α = 0.80).
    “I worried about dying.”
    “I felt nervous.”
  3. A seven-item scale of functional well-being in everyday life (Cronbach's α = 0.84).
    “I was enjoying my usual leisure pursuits.”
    “My work was fulfilling.”
  4. A nine-item scale of the extent a woman was concerned about side effects and risks (Cronbach's α = 0.66).
    “I felt self-conscious about the way I dressed.”
    “I felt sexually attractive.”
    “I worried about the effect of stress on my illness.”

A complete set of items is available from the authors.


1. Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1999. CA Cancer J Clin. 1999;49:8–31. [PubMed]
2. Meyerowitz BE. Psychosocial correlates of breast cancer and its treatments. Psychol Bull. 1980;87:108–31. [PubMed]
3. Vinokur AD, Threatt BA, Caplan RD, et al. Physical and psychosocial functioning and adjustment to breast cancer: long-term follow-up of a screening population. Cancer. 1989;63:394–405. [PubMed]
4. Ganz PA, Rowland JH, Desmond K, Meyerowitz BE, Wyatt GE. Life after breast cancer: understanding women's health-related quality of life and sexual functioning. J Clin Oncol. 1998;16:501–14. [PubMed]
5. Nayfield SG, Ganz PA, Moinpour CM, Cella DF, Hailey BJ. Report from a National Cancer Institute (USA) workshop on quality of life assessment in cancer clinical trials. Qual Life Res. 1992;1:203–10. [PubMed]
6. Aaronson NK, Meyerowitz BE, Bard M, et al. Quality of life research in oncology: past achievements and future priorities. Cancer. 1991;67:839–43. [PubMed]
7. Chevarley F, White E. Recent trends in breast cancer mortality among white and black U.S. women. Am J Public Health. 1997;87:775–81. [PubMed]
8. Health Services Research Group. Development of the Index of Medical Underservice. Health Services Research. 1975:168–80. Summer. [PMC free article] [PubMed]
9. Marino C, Gerlach KK. An analysis of breast cancer coverage in selected women's magazines, 1987–1995. Am J Health Promot. 1999;13:163–70. [PubMed]
10. Scroggins TG, Jr, Bartley TK. Enhancing cancer control: assessing cancer knowledge; attitudes and beliefs in disadvantaged communities. J La State Med Soc. 1999;151:202–4. [PubMed]
11. Nelles WB, McCaffrey RJ, Blanchard CG, Ruckdeschel JC. Social support and breast cancer: a review. J Psychosoc Oncol. 1991;9:21–34.
12. Street RL, Voigt B. Patient participation in deciding breast cancer treatment and subsequent quality of life. Med Decis Making. 1997;17:298–306. [PubMed]
13. Bilodeau BA, Degner LF. Information needs, sources of information, and decisional roles in women with breast cancer. Oncol Nurs Forum. 1996;23:691–6. [PubMed]
14. Gustafson DH, Hawkins RP, Boberg E, et al. Impact of a patient-centered, computer-based health information/support system. Am J Prev Med. 1999;16:1–9. [PubMed]
15. Smaglik P, Hawkins R, Pingree S, Gustafson D. The quality of interactive computer use among HIV infected individuals. J Health Communication. 1998;3:53–68. [PubMed]
16. Boberg E, Gustafson D, Hawkins R, et al. Development, acceptance and use patterns of a computer based education and social support system for people living with AIDS/HIV infection. Comp Hum Behav. 1995;11:289–311.
17. Pingree S, Hawkins RP, Gustafson D, et al. Will the disadvantaged ride the information highway? Hopeful answers from a computer-based health crisis system. J Broadcasting Electronic Media. 1996;40:331–53.
18. Gustafson D, Wise M, McTavish F, et al. Development and pilot evaluation of a computer-based support system for women with breast cancer. J Psychosoc Oncol. 1993;11:69–93.
19. McTavish FM, Gustafson DH, Owens BH, Ozbolt JG, et al. Transforming Information, Changing Health Care; Proceedings of the Eighteenth Annual Symposium on Computer Applications in Medical Care. Philadelphia: Hanley & Belfus; 1994. CHESS: an interactive computer system for women with breast cancer piloted with an under-served population.
20. Gustafson DH, Taylor JO, Thompson S, Chesney P. Assessing the needs of breast cancer patients and their families. Qual Manage Healthcare. 1993;2(1):6–17. [PubMed]
21. Love S, Lindsey K. Dr. Susan Love's Breast Book. Reading, Mass: Addison-Wesley Publishing Company; 1995.
22. Karnofsky DA, Burchenal JH. The clinical evaluation of chemotherapeutic agents in cancer. In: Macleod CM, editor. Evaluation of Chemotherapeutic Agents. New York: Columbia University Press; 1949. pp. 199–205.
23. Schag CC, Heinrich RL, Ganz PA. Karnofsky Performance Status revisited: reliability, validity, and guidelines. J Clin Oncol. 1984;2:187–93. [PubMed]
24. Darlington RB. Regression and Linear Models. New York: McGraw Hill; 1990. pp. 263–6.
25. Ende-Murphy K. The relationship of self-directed learning, self-efficacy, and health value in young women with cancer using a computer health education program. Madison, Wis: University of Wisconsin; 1996. Unpublished Ph.D. Thesis.
26. Brady MJ, Cella DF, Mo F, et al. Reliability and validity of the Functional Assessment of Cancer Therapy-Breast quality-of-life instrument. J Clin Oncol. 1997:974–86. [PubMed]
27. Cella DF, Tulsky DS, Gray G, et al. The Functional Assessment of Cancer Therapy scale: development and validation of the general measure. J Clin Oncol. 1993;11:570–9. [PubMed]
28. Fairclough DL, Cella DF. Functional Assessment of Cancer Therapy (FACT-G): non-response to individual questions. Qual Life Res. 1996;5:321–9. [PubMed]
29. Cella DF, Bonomi AE, Lloyd SR, et al. Reliability and validity of the Functional Assessment of Cancer Therapy-Lung (FACT-L) quality of life instrument. Lung Cancer. 1995;12:199–220. [PubMed]
30. Moinpour CM, Feigl P, Metch B, Hayden KA, Meyskens FL, Jr, Crowley J. Quality of life end points in cancer trials: review and recommendations. J Natl Cancer Inst. 1989;81:485–95. [PubMed]
31. Hardy MA. Regression with Dummy Variables. Newbury Park, Calif: Sage Publications Inc; 1993. Sage University Paper Series: Quantitative Applications in the Social Sciences, Series No. 07–093.
32. Cohen J, Cohen P. Applied multiple regression/correlation analysis for the behavioral sciences. 2nd Ed. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.; 1983.
33. Seaman MA, Levin JR, Serlin RC. New developments in pairwise multiple comparisons: some powerful and practicable procedures. Psychol Bull. 1991;110:577–86.
34. Holm S. A simple sequentially rejective multiple test procedure. Scand J Stat. 1979;6:65–70.
35. Cohen J. Statistical power analysis for the behavioral sciences. 2nd Ed. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.; 1988.
36. Demetri GD, Kris M, Wade J, Degos L, Cella D. Quality of life benefit in chemotherapy patients treated with Epoetin Alfa is independent of disease response or tumor type: results from a prospective community oncology study. J Clin Oncol. 1998;16:3412–25. [PubMed]
37. Bonomi P, Kim K, Fairclough D, et al. Comparison of survival and quality of life in advanced non-small-cell lung cancer patients treated with pacilitaxel-cisplatin vs. etoposide-cisplatin: results of an Eastern Cooperative Oncology Group trial. J Clin Oncol. 2000;18:623–31. [PubMed]
38. U.S. Bureau of the Census. United States Population Estimates by Age, Sex, Race and Hispanic Origin, 1990–1997. Washington, DC: Population Division, U.S. Bureau of the Census; 1998. p. 20233.
39. Gustafson D, McTavish F, Hawkins R, et al. Computer support for elderly women with breast cancer: results of a population-based intervention. (Letter) JAMA. 1998;280(15) [PubMed]
40. CommerceNet/Nielsen. Electronic commerce on the rise according to CommerceNet /Nielsen Media Research Survey. 1998 Jan. 7 http://www.commerce.net/news/press/121197.html.
41. Eng T, Maxfield A, Patrick K, Deering M, Ratzan S, Gustafson D. Access to health information and support: a public highway or a private road? JAMA. 1998;280:1371–5. [PubMed]
42. Robinson T, Eng T, Patrick K, Gustafson D. An evidence-based approach to interactive health communication: a challenge to medicine in the information age. JAMA. 1998;280:1264–9. [PubMed]

Articles from Journal of General Internal Medicine are provided here courtesy of Society of General Internal Medicine