RESEARCH Changes in diet, physical activity, and supplement use among adults diagnosed with cancer RUTH E. PATTERSON, PhD, RD; MARIAN L. NEUHOUSER, PhD, RD; MONIQUE M. HEDDERSON, MPH; STEPHEN M. SCHWARTZ, PhD; LEANNA J. STANDISH, PhD; DEBORAH J. BOWEN, PhD ABSTRACT Objective To investigate the prevalence and predictors of changes in diet, physical activity, and dietary supplement use among cancer patients. Design/subjects Telephone interviews of a populationbased sample of 126 breast, 114 prostate, and 116 colorectal cancer patients from the state of Washington. Analysis Logistic regression was used to estimate the odds ratio as a measure of the association of participant characteristics with lifestyle changes in the 12 months before the interview. Results Overall, 66.3% of patients reported making lifestyle changes: 40.4% made one or more dietary changes, 20.8% added new physical activity, and 48.0% started taking new dietary supplements. Compared with men, women were 2.2 times more likely to take new dietary supplements (P ⬍ .01). Compared with patients aged 35 to 59, those aged 60 to 69 and 70 or older were statistically significantly less likely to make dietary changes (odds ratio ⫽ 0.39 and 0.54, respectively) or to take new supplements (odds ratio ⫽ 0.42 and 0.69, respectively). Compared with patients who received only one medical treatment, those receiving three or more treatments were more likely to make dietary changes (odds ratio ⫽ 2.6) or to start new physical activity (odds ratio ⫽ 3.0). Patients diagnosed 12 to 24 months before the interview were as likely to report making lifestyle changes as those diagnosed within one year of the interview. Having a stronger desire for personal control or internal locus of control predicted use of new dietary supplements (P for trend ⬍ .05 for both). Applications/conclusions Cancer survivors are likely to be making lifestyle changes and represent a group that could benefit from counseling on diet and physical activity. J Am Diet Assoc. 2003;103:323-328. A pproximately 1,285,000 Americans will be diagnosed with cancer in 2002 (1). Although cancer is generally thought of as a life-threatening disease, survival rates have improved dramatically for many malignancies (2). For example, five-year survival in the United States is about 80% for prostate and breast cancers and 60% for colorectal cancer (1,3), and the National Cancer Institute estimates that approximately 8.9 million Americans with a history of cancer were alive in 1997 (1). Life-threatening health events such as a diagnosis of cancer may prompt psychological distress that motivates individuals to reduce risk (4,5). The experience of cancer diagnosis and treatment is difficult and may serve as a critical cue for an individual to undertake future health promotion activities (6). In a survey of 1,667 cancer survivors, Demark-Wahnefried and colleagues reported that there was a strong interest in health promotion programs aimed at healthier diets, exercise, and smoking cessation (7). The major role of a dietetics professional in relation to cancer patients is often nutritional status assessment and support during the early phases of treatment and recovery. However, nutrition interventions may be beneficial to cancer survivors during the health maintenance phase of R. Patterson is a research associate professor and S. Schwartz is an associate professor, Department of Epidemiology and Nutritional Sciences Program, and D. Bowen is a professor, Department of Health Services, University of Washington, Seattle, WA. R. Patterson is an associate member, M. Neuhouser is a senior staff scientist, and D. Bowen is a member, Cancer Prevention Research Program; and S. Schwartz is a member, Program in Epidemiology, Fred Hutchinson Cancer Research Center, Seattle, WA. M. Hedderson is a research assistant, Division of Research, Kaiser-Permanente, Oakland, CA. L. Standish is Director of Research, Bastyr University, Bothell, WA. Address correspondence to: Ruth E. Patterson, PhD, RD, Cancer Prevention Research Program, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, MP 702, Seattle, WA 98109. E-mail: rpatters@fhcrc.org Copyright © 2003 by the American Dietetic Association. 0002-8223/03/10303-0002$35.00/0 doi: 10.1053/jada.2003.50045 Journal of THE AMERICAN DIETETIC ASSOCIATION / 323 RESEARCH cancer, in which the goals are improved quality of life as well as preventing cancer recurrence, second primary tumors, and other preventable diseases (8). Data regarding changes in lifestyle in response to a diagnosis of cancer are scarce. This report examines self-reported changes in diet, physical activity, and dietary supplement use among cancer patients diagnosed up to 24 months in the past. We investigate whether patients felt that these changes improved their health and well-being and identify sociodemographic, clinical, and selected psychosocial characteristics that predict these lifestyle changes. METHODS Overview and Study Participants Participants were identified from the population-based Cancer Surveillance System, which is part of the National Cancer Institute’s Surveillance Epidemiology and End Results program (9). The registry catchment area includes 13 counties in western Washington state. We obtained permission from each patient’s physician before sending a letter of invitation to participate in the survey, which was followed up by a telephone call approximately one week later. We made at least 15 attempts to contact patients by telephone. Trained interviewers used a computer-assisted telephone interview system to administer a structured questionnaire. Interviews were conducted between February and July of 1999. The protocol for the study was approved by the Institutional Review Board of the Fred Hutchinson Cancer Research Center. Data regarding changes in lifestyle in response to a diagnosis of cancer are scarce Eligible participants were English-speaking patients, aged 20 to 79, diagnosed with first primary invasive breast, prostate, or colorectal cancer between February 1997 and December 1998. We chose to examine these cancers because they are the major ones affecting Americans that also have high survival rates (1). We randomly selected a sample of 511 adult patients to obtain approximately equal numbers of breast, prostate, and colon cancers such that half were diagnosed within the previous 11 months and the other half were diagnosed 12 to 24 months before sampling. Two patients were ineligible because they did not speak English, and five patients were deceased. We completed interviews with 356 (70.6%) of 504 eligible participants. Reasons for nonparticipation were physician refusal (5.0%), inability to locate (7.9%), patient refusal (15.3%), and illness (1.2%). Changes in Diet, Physical Activity, and Supplement Use Each patient was asked to report on lifestyle changes made in the past 12 months. For all changes made, each respondent was asked whether he or she strongly agreed, agreed, disagreed or strongly disagreed with the following statement: “This ⬍lifestyle change/dietary supplement⬎ has improved my physical and mental well-being.” 324 / March 2003 Volume 103 Number 3 Diet Each patient was asked if he or she had made any major changes in dietary intake that were aimed at coping with cancer or reducing the risk of cancer spreading or returning. If yes, the patient was asked to specify the dietary changes. Interviewers coded the response(s) from a list of options: (a) became a vegetarian; (b) ate more fruits and vegetables; (c) ate less red meat; (d) ate less fat; (e) ate more healthy in general; (f) fasted; (g) juiced; (h) used special supplements such as Ensure, Boost, or protein powders; (i) followed a weight-loss diet; or (j) other. The exact text of “other” responses was entered into computers by interviewers. More than one response was allowed. This method of assessing recent change in consumption has proven useful in other studies of trends in diet (10,11). Physical activity Each patient was asked if he or she had begun physical activities in the past 12 months that were aimed at coping with cancer or reducing the risk of cancer spreading or returning. If yes, the patient was asked to specify the activities. Interviewers entered the response(s) from a list of options: (a) aerobic exercise, (b) other exercise such as weight training, (c) breathing exercises, (d) qi gong, (e) tai chi, (f) yoga, or (g) other. Again, the exact text of “other” responses was entered into the computer, and more than one response was allowed. Dietary supplements Each patient was asked whether in the past 12 months he or she had taken any vitamin, mineral, herbal, or other supplements. If yes, we recorded the type of supplement, such as one-a-day type multivitamins (with or without minerals); single supplements of vitamins, minerals, or herbals; and other mixtures. For other mixtures, we recorded the ingredients (ie, the exact mixture of vitamins, minerals, and/or herbals) and flagged the entry as being a single pill. For each supplement taken, each patient was asked whether he or she had taken the product before the cancer diagnosis. Psychosocial factors The diagnosis of cancer and the health effects of the associated treatment regimens can leave patients with a sense of hopelessness and loss of control (4,5,12), and lifestyle is an area in which cancer patients can regain a sense of personal control. Therefore we assessed three psychosocial factors that we hypothesized might predict changes in health behavior: perceived health status, desire for personal control, and internal locus of control. For perceived health status, each patient was asked whether they would classify their health as excellent, very good, good, fair, or poor. To assess each patient’s desire for control over cancer, we used questions from a validated scale developed to measure desire for control among dental patients (13). Each patient was asked how much control he or she wanted in decisions related to cancer treatment and in decisions related to cancer-related symptoms. Response options were on a four-point scale from strongly disagree to strongly agree and were summed to create a summary score. The six-item internal health locus of control scale from the Multidimensional Health Locus of Control Instrument was administered (14). This model was developed to assess people’s beliefs that their own behavior determines their health. People with a stronger internal locus of control regard their health as largely within their own control and are more likely to engage in health-maintaining behavior. To facilitate patient comprehension of the tool during a telephone interview, we shortened the RESEARCH response categories from the standard six-point scale to a fourpoint scale with Likert-type responses from strongly disagree (rated as one) to strongly agree (rated as four). A single score was calculated by summing responses to the six questions. The Cronbach coefficient alpha for this scale was 0.73, indicating that it had high internal consistency. Other measures Sociodemographic data were collected on age, sex, race, and education. Patients were asked to list the medical treatments received: surgery, radiotherapy, chemotherapy, and/or hormone therapy. We created a variable for these data to indicate intensity of medical treatment by grouping participants by number of treatments received (zero, one, two, or more). Seventeen participants reported no medical treatment and were excluded from this analysis. Clinical information on type of cancer and stage of disease were abstracted from the Cancer Surveillance System registry. Statistical Analysis We provide data on prevalence of lifestyle changes after cancer diagnosis, including changes in diet, physical activity, and use of dietary supplements. We give data only for the most common changes, defined as those made by at least 5% of patients. For each lifestyle change, we present the percent of participants who strongly agreed or agreed that the change improved their health and well-being. Lifestyle is an area in which cancer patients can regain a sense of personal control Logistic regression was used to estimate odds ratios (OR) as the measure of association of lifestyle changes with sex, age (in years, 35 to 54, 55 to 69, 70 or older), education (in years, ⱕ12, 13 to 15, 16 or more), type of cancer, stage at diagnosis, number of medical treatments, and time since diagnosis (in months, ⬍12, 12 to 24). Psychosocial predictors of lifestyle changes were divided into approximate tertiles to provide stable estimates of associations. All ORs are adjusted for confounding factors (age, sex, education, stage, and medical treatments). For ordered independent variables (age, education, medical treatments, and psychosocial factors), the association with lifestyle changes was assessed with a test for trend using the logistic regression analogue to the Mantel extension test (ie, a test based on the significance of a single trend variable coded as the category of exposure). It is important to note that because of the choice of cancer sites included in this study, sex and type of cancer are confounded. All breast cancer cases were female, all prostate cancer cases were male, and colorectal cancer cases were 44.8% female. Therefore we are unable to fully separate the effects of sex vs type of cancer, and in models adjusted for sex, we are also partially adjusting for the type of cancer. For the models of associations of type of cancer with use of alternative medicine use, we are unable to adjust for sex because of this confounding. RESULTS The final sample was composed of 356 adult cancer patients, of whom 126 had breast cancer, 114 had prostate cancer, and 116 Table 1 Most common lifestyle changes among breast (n⫽126), prostate (n⫽114), and colorectal (n⫽116) cancer patients diagnosed up to 2 years previously Lifestyle change Dietary changes Eat more fruits and vegetables Eat less red meat Eat less fat Eat more healthily New physical activity Aerobic exercise Qi gong, tai chi, yoga, others Participants making change in past 12 months Participants reporting that change improved health and well-being 4 144 (40.4) n (%) 128 (88.9) 65 (45.1) 38 (26.4) 41 (28.5) 18 (12.5) 62 (95.4) 31 (81.6) 36 (87.8) 18 (100.0) 74 (20.8) 61 (82.4) 73 (98.6) 61 (100.0) 16 (21.6) 20 (95.2) New dietary supplements Multivitamins Vitamin E compound(s)a Calcium Vitamin C compound(s)a Garlic Echinacea 171 (48.0) 50 (29.2) 48 (28.1) 29 (17.0) 25 (14.6) 16 (9.4) 14 (8.2) 158 (92.3) 46 (92.0) 43 (89.6) 27 (93.1) 25 (100.0) 16 (100.0) 14 (100.0) Number of changes made None One change Two changes ⱖThree changes 120 (33.7) 117 (32.9) 85 (23.9) 34 (9.6) Not Not Not Not 3 applicable applicable applicable applicable a Any mixture or single supplement containing vitamin E (or vitamin C), not including multivitamins. had colorectal cancer. Half of the patients were female, 91.3% were white, the mean age was 62.5 years (SD ⫽ 10.7), and 37.1% had a college education. Fifty-one percent of patients had been diagnosed within the past 12 months, and the remainder were diagnosed 12 to 24 months in the past. Table 1 gives the most common changes in lifestyle, defined as changes that at least five percent of patients reported making. Overall, lifestyle changes were very common after a cancer diagnosis. Two thirds of patients reported making at least one behavior change in the past 12 months: about 40% made dietary changes, 20% began a new physical activity, and 50% started taking new dietary supplements. Approximately 25% of patients made two changes, and almost 10% made all three lifestyle changes. Among those making dietary changes, the most common was eating more fruits and vegetables; for physical activity, the most common change was adding aerobic exercise; and for supplements, the most common change was starting to take multivitamins and/or vitamin E. The vast majority of patients reported that these lifestyle changes improved their health and well-being. Table 2 gives the ORs of making these lifestyle changes by sociodemographic and other characteristics. Female patients were more than twice as likely to take new dietary supplements as males (P ⬍ .01). Compared with younger patients (35-59), older patients were about half as likely to make dietary changes (P for trend ⬍ .001) and to take new dietary supplements (P for trend ⬍ .05). More educated patients were statistically significantly more likely to make dietary changes and to take new supplements (P for trend ⬍ .05), with a similar trend for physical activity that did not reach statistical significance. ComJournal of THE AMERICAN DIETETIC ASSOCIATION / 325 RESEARCH Table 2 Association of lifestyle changes with sociodemographic and clinical characteristics among breast (n⫽126), prostate (n⫽114), and colorectal (n⫽116) cancer patients diagnosed up to 2 years previously Lifestyle changes made in past 12 months Diet changes Began new physical activity Took new dietary supplements % Adjusted ORb % Adjusted ORb % Adjusted ORb 356 40.4 – 20.8 – 48.0 – 178 178 31.4 49.4 1.0 1.19 15.2 26.4 1.0 1.53 36.0 60.1 1.0 2.19** Age (y) 35-59 60-69 70 or more P for trend 132 119 101 59.9 31.9 24.7 1.0 0.39*** 0.52*** ⬍.001 27.3 14.3 17.8 1.0 0.57 0.80 65.9 37.8 37.6 1.0 0.42** 0.69* ⬍.05 Education (y) ⱕ12 13-15 16 or more P for trend 130 93 132 26.9 49.5 47.7 1.0 3.09 2.12 ⬍.05 16.1 18.3 27.3 1.0 1.27 1.93 36.9 49.5 58.3 1.0 1.53 2.21** ⬍.01 Type of cancer Colorectal Prostate Breast 116 114 126 32.8 27.7 59.5 1.0 0.83 1.72 15.5 15.8 30.2 1.0 0.70 1.0 46.5 31.6 64.3 1.0 0.50* 1.63 Stage at diagnosis I (local) II/III (distant/regional) 192 140 37.5 45.0 1.0 1.41 20.3 21.4 1.0 1.02 44.8 55.0 1.0 1.65* 132 107 100 25.0 43.0 58.0 1.0 1.89 2.62 ⬍.01 14.4 16.8 34.0 1.0 1.33 3.01 ⬍.01 37.1 48.6 62.0 1.0 1.13 1.26 183 173 37.2 43.9 1.0 1.51 18.6 23.1 1.0 1.54 46.5 49.7 1.0 1.06 Characteristic Na All Sex Male Female Medical treatmentc One treatment Two treatments Three treatments or more P for trend Diagnosis date ⬍12 mo 12-24 mo a Sample sizes vary slightly because of missing data. OR⫽Odds ratio; adjusted for age, sex, education, stage, and number of medical treatments. Treatments included surgery, radiation, chemotherapy, and hormone therapy. *P⬍.05. **P⬍.001. ***P⬍.001. b c pared with colorectal cancer patients, men with prostate cancer were half as likely to take new dietary supplements (P ⬍ .05). Compared with those with local disease, patients with distant/regional disease were about 1.7 times more likely to take new supplements (P ⬍ .05). Patients receiving three or more medical treatments were two to three times more likely to make dietary changes or to start new physical activity compared with patients receiving only one medical treatment (P for trend ⱕ .05). Patients diagnosed 12 to 24 months in the past were equally as likely to report making lifestyle changes as those diagnosed less than 12 months ago. Table 3 gives psychosocial factors that could predict making lifestyle changes after a cancer diagnosis. Perceived health status did not significantly predict lifestyle changes. However, stronger desire for personal control and having a high internal locus of control were associated with almost a two times greater likelihood of taking new dietary supplements (P for trend ⬍ .05). 326 / March 2003 Volume 103 Number 3 DISCUSSION This article indicates that about two thirds of cancer patients reported making changes in diet, physical activity, and/or supplement use in the past 12 months. Most of these changes were healthful (ie, eating more fruits and vegetables and less fat, adding new aerobic activity) and are consistent with recommendations made by the American Cancer Society for cancer prevention (15). Other changes were more controversial (taking dietary supplements such as vitamin C, garlic, or echinacea), especially during the treatment phase, in which antioxidants may interfere with chemotherapy or radiotherapy (16,17). It is important to note that patients overwhelmingly thought that these lifestyle changes improved their health and well-being. We know of no research on changes in health-related behavior after cancer diagnosis, although a number of studies have reported on the prevalence of health behavior in cancer survivors. In a survey of 1,667 cancer survivors, investigators found RESEARCH Table 3 Lifestyle changes by psychosocial characteristics among breast (n⫽126), prostate (n⫽114), and colorectal (n⫽116) cancer patients diagnosed up to 2 years previously Lifestyle changes made in past 12 months Diet changes Began new physical activity Took new dietary supplements Psychosocial characteristic Na % Adjusted ORb % Adjusted ORb % Adjusted ORb Perceived health status Excellent/very good Good Fair/poor 175 118 60 41.7 39.0 36.7 1.0 1.19 1.31 21.7 22.0 13.3 1.0 1.31 0.71 53.7 39.8 46.7 1.0 0.76 1.06 169 45 133 32.5 44.4 51.9 1.0 1.52 1.40 17.7 11.1 27.8 1.0 0.47 1.31 36.7 51.1 61.6 1.0 1.52 1.74 ⬍.05 71 127 134 45.1 39.4 38.8 1.0 0.68 0.82 21.1 17.3 23.1 1.0 0.80 1.44 42.3 44.1 56.0 1.0 1.07 1.94 Desire for personal control Low Medium High P for trend Internal locus of control Low Medium High P for trend ⬍.05 a Sample sizes vary slightly because of missing data. b OR⫽Odds ratio; adjusted for age, sex, education, stage, and number of medical treatments. that 69% reported adhering to a low-fat diet, 45% ate five or more daily servings of fruits and vegetables, and 58% reported routine exercise (7). A study of dietary supplement use by women at risk for breast cancer reoccurrence (n ⫽ 435) reported that 80.9% of respondents reported using supplements (18). However, this was among a sample of women enrolled in a diet intervention to reduce breast cancer recurrence, and therefore was a self-selected volunteer sample of patients that were likely to be very interested in health issues. In the same sample of women, 60% of women reported weight gain, 26% reported weight loss, and 14% reported no change within the four years after a diagnosis of breast cancer; suggesting changes in diet, physical activity, or both (19). Older patients were less likely to adopt new lifestyle practices in response to a diagnosis of cancer, whereas more educated patients, females, and patients undergoing multiple treatments were more likely to make changes. These results are similar to findings from studies on prevalence of alternative medicine use and taking supplements among cancer patients (12,20-26). It is interesting to note that a high desire for control was only associated with increased use of dietary supplements and did not predict making changes in diet or physical activity, which may be motivated more by other psychosocial factors such as belief in a diet-disease connection (10) or concerns about weight changes (19). It is possible that taking supplements may be a coping behavior for dealing with the emotional aftermath of cancer rather than a health improvement strategy. Research into patients’ beliefs about the functions of supplements is needed. This study has several limitations. First, although our response rate was good (70.8%), we cannot ensure that our respondents have the same behavior as the nonrespondents. It is possible that patients with less favorable behavior change might be less likely to participate in a cancer study. Therefore our estimates of health behavior change may be somewhat inflated. Our prevalence estimates cannot be completely generalized to all cancer patients because our sample was composed only of breast, colon, and prostate cancer patients selected up to two years postdiagnosis. Because survival rates for these cancers are high, motivations for making lifestyle changes among these patients may be different than for patients having cancers with a poor prognosis, such as lung or pancreatic cancers. In addition, this study population reflects western Washington population demographics, and therefore was mostly white. Additional studies would be needed to assess lifestyle changes in other race and ethnicity groups. Finally, lifestyle changes are entirely based on self-report and were collected as patient memories of the change rather than by gathering data at two or more time points. It is possible that the experiences of cancer diagnosis, treatment, and coping may have affected these memories. These data suggest that cancer survivors are highly motivated to attempt lifestyle changes in diet, physical activity, and dietary supplement use as much as two years postdiagnosis. Little is known about how lifestyle behaviors affect cancer recurrence and prognosis. Nonetheless, family, friends, media, and health food stores may readily give advice, much of which is offered with a degree of certainty that is out of proportion with the evidence available (6). Therefore studies are needed that test hypotheses about changing health behavior to improve postdiagnosis survival rates. APPLICATIONS The results of these analyses indicate that cancer patients often make lifestyle changes months to years after diagnosis and represent a group of adults that could benefit from counseling on diet and physical activity. We suggest that dietetics professionals who interact with these patients provide them counseling and/or resources to make these lifestyle changes. Finally, dietetics professionals working with cancer survivors need to be wary of dismissing or discounting dietary supple■ Journal of THE AMERICAN DIETETIC ASSOCIATION / 327 RESEARCH ments or other diet or physical activity regimens because patients overwhelmingly thought that these activities improved their health and well-being. References 1. American Cancer Society. Cancer facts and figures 2002. Available at: http://www.cancer.org. Accessed March 15, 2002. 2. Welch HG, Schwartz LM, Woloshin S. Are increasing 5-year survival rates evidence of success against cancer? JAMA. 2000;283:2975-1978. 3. Gatta G, Capocaccia R, Coleman MP, Gloeckler Ries LA, Hakulinen T, Micheli A, Sant M, Verdecchia A, Berrino F. Toward a comparison of survival in American and European cancer patients. Cancer. 2000;89:893-900. 4. McBride CM, Clipp E, Peterson BL, Lipkus IM, Demark-Wahnefried W. Psychological impact of diagnosis and risk reduction among cancer survivors. Psychooncology. 2000;9:418-427. 5. Eyre HJ. Nutritional advice for cancer survivors. CA Cancer J Clin. 2001; 51:151-152. 6. Rowland JH, Massie MJ. Breast cancer. In: Holland JC, ed. Psychooncology. New York, NY: Oxford University Press; 1998. 7. Demark-Wahnefried W, Peterson B, McBride C, Lipkus I, Clipp E. Current health behaviors and readiness to pursue life-style changes among men and women diagnosed with early stage prostate and breast carcinomas. Cancer. 2000;88:674-684. 8. Brown J, Byers T, Thompson K, Eldridge B, Doyle C, Williams AM. Nutrition during and after cancer treatment: A guide for informed choices by cancer survivors. CA Cancer J Clin. 2001;51:153-187. 9. Hankey BF, Ries LA, Edwards BK. The Surveillance, Epidemiology, and End Results program: A national resource. Cancer Epidemiol Biomarkers Prev. 1999;8:1117-1121. 10. Patterson RE, Kristal AR, White E. Do beliefs, knowledge and norms about diet and cancer predict dietary change? Am J Public Health. 1996;86:13941400. 11. Patterson RE, Satia JA, Kristal AR, Neuhouser NH, Drewnowski A. Is there a nutrition backlash? J Am Diet Assoc. 2001;101:43-46. 12. Burstein HJ, Gelber S, Guadagnoli E, Weeks JC. Use of alternative medicine by women with early stage breast cancer. N Engl J Med. 1999;340: 1733-1739. 13. Logan HL, Baron RS, Keeley K, Law A, Stein S. Desired control and felt control as mediators of stress in a dental setting. Health Psychol. 1991;10: 352-359. 14. Wallston KA, Wallston BS. Development of the Multidimensional Health Locus of Control (MHLC) Scales. Health Education Monographs. 1978;6:160179. 328 / March 2003 Volume 103 Number 3 15. The American Cancer Society 1996 Advisory Committee on Diet, Nutrition, and Cancer Prevention. Guidelines on diet, nutrition, and cancer prevention: Reducing the risk of cancer with healthy food choices and physical activity. CA Cancer J Clin. 1996;24:325-341. 16. Lamson DW, Brignall ML. Antioxidants in cancer therapy; their actions and interactions with oncologic therapies. Alt Med Rev. 1999;4:304-329. 17. Miller LG. Herbal medicinals: Selected clinical considerations focusing on known or potential drug-herb interactions. Arch Intern Med. 1998;158:22002211. 18. Newman V, Rock CL, Faerber S, Flatt S, Wright AA, Pierce JP. Dietary supplement use by women at risk for breast cancer recurrence. J Am Diet Assoc. 1998;98:285-292. 19. Rock CL, Flatt SW, Newman V, Caan BJ, Haan MN, Stefanick ML. Factors associated with weight gain in women after diagnosis of breast cancer. J Am Diet Assoc. 1999;99:1212-1218, 1221. 20. Boon H, Stewart M, Kennard MA, Gray R, Sawka C, Brown JB, McWilliam C, Gavin A, Baron RA, Aaron D, Haines-Kamka T. Use of complementary/ alternative medicine by breast cancer survivors in Ontario: Prevalence and perceptions. J Clin Oncol. 2000;18:2515-2521. 21. Cassileth BR, Chapman CC. Alternative and complementary cancer therapies. Cancer. 1996;77:1026-1034. 22. Eisenberg DM, Kessler RC, Foster C, Norlock FE, Calkins DR, Delbanco TL. Unconventional medicine in the United States. N Engl J Med. 1993;328: 246-252. 23. Eisenberg DM, Davis RM, Ettner SL, Appel S, Wilkey S, Van Rompay M, Kessler RC. Trends in alternative medicine use in the United States, 19901997. JAMA. 1998;280:1569-1575. 24. Ernst E, Cassileth BR. The prevalence of complementary/alternative medicine in cancer. Cancer. 1998;83:777-782. 25. Richardson MA, Sander T, Palmer JL, Greisinger A, Singletary SE. Complementary/alternative medicine use in a comprehensive cancer center and the implications for oncology. J Clin Oncol. 2000;18:2505-2514. 26. Sollner W, Maislinger S, DeVries A, Steixner E. Rumpold G. Lukas P. Use of complementary and alternative medicine by cancer patients is not associated with perceived distress or poor compliance with standard treatment but with active coping behavior. Cancer. 2000;89:873-880. This study was supported by contracts N01-PC-67009 and N01-CN-05230, both from the National Cancer Institute, as well as funds from the Fred Hutchinson Cancer Research Center.