Of 165 BRCA+ women who knew their mutation status prior to enrolling in the Breast Imaging Screening Study, 164 answered the pre-visit telephone CAM questionnaire and were included in this analysis. The cohort was highly-educated (92.7% attending college); predominately white (97.6%); most were married (72%); over half reported having children (54.3%); and 17.7% had a personal history of cancer prior to study entry.
The overall CAM use rate in our BRCA+ population was 78% (128/164), with an average of 2.3 CAM therapies used per person. Thirty-four percent of the cohort reported having used three or more CAM therapies within the past year. As shown in Figure , only 15 (9.1%) BRCA+ women reported prayer as the sole CAM therapy used. If spiritual healing/prayer are excluded, 68.9% of participants had used CAM in the previous year. Spiritual healing/prayer and lifestyle diet were the most commonly reported modalities (48.8% and 48.2%, respectively). Figure shows CAM use grouped by NCCAM domains: biologically-based practices (herbal medicine, lifestyle diet); mind-body medicine (spiritual healing or prayer, meditation, yoga or tai chi, relaxation techniques, biofeedback, imagery, hypnosis); energy healing, manipulative and body-based practices (massage therapy); and alternative medical systems (homeopathic treatment, acupuncture). Mind-body therapies and biologically-based practices were the most commonly used domains (61.6% and 51.8%, respectively). However, if prayer is excluded from the computations, biologically-based practices become the most prevalent modality.
Patterns of CAM Use in BRCA+ Women.
There were significant positive associations between CAM use and several demographic, health-related, and psychosocial variables (Tables and ). Higher education (p = 0.015), a previous cancer diagnosis (p = 0.006), and older age (p = 0.007) were all associated with increased CAM use. Although the overall scores were not high (median = 1.33), higher ovarian cancer worry scores were statistically significantly associated with increased CAM use (p = 0.011) (Table ). These findings did not change when the ovarian cancer worry analysis was restricted to the 95 women with intact ovaries (data not shown); therefore, all women were included in the multivariate regression analysis. Among the demographic indicators, race (p = 1.00), number of children (p = 0.52), and marital status (p = 0.91) were not associated with CAM use, and neither were the health-related behaviors of breast or ovarian cancer screening, nor the psychosocial variables depression, perceived cancer risk, and breast cancer worry.
Associations Between CAM Use and Demographic, Psychosocial and Health-related Categorical Predictor Variables in BRCA+ Women
CAM Use and Demographic, Psychosocial and Health-related Continuous Predictor Variables in BRCA+ Women
Factors with p ≤ 0.20 (age, education, previous cancer diagnosis, ovarian cancer worry, breast cancer worry, and BSE) were included in the multivariate logistic regression model. Odds ratios and 95% confidence intervals are presented in Table . Age, higher education, previous cancer diagnosis, and increased ovarian cancer worry remained statistically significantly associated with CAM use (p ≤ 0.05). Previous cancer diagnosis showed the strongest association with increased CAM use; women with any previous cancer were 17 times more likely to use CAM than unaffected women. Multivariate analysis showed a weak, yet statistically significant, inverse association between frequency of BSE and CAM use; women who did BSE less than once a month were more likely to use CAM.
Multivariate Associations Between CAM Use and Predictor Variables in BRCA+ Women
Since a previous cancer diagnosis was highly-associated with CAM use in our population, a finding previously-reported in the general population literature, we repeated our analyses after excluding the women with a previous cancer diagnosis. The overall patterns of the CAM therapies used remained unchanged (data not shown), with the exception of lifestyle diet being more commonly used than spiritual healing/prayer (47.4% and 44.4%, respectively) compared with 48.2% and 48.8%, respectively in the previous analysis. When spiritual healing/prayer was excluded, overall CAM use was 65.2% compared to 68.9% when all BRCA+ women were included in the analysis.
Age, education, previous cancer diagnosis, ovarian cancer worry, breast cancer worry, and BSE all had p ≤ 0.20 in the univariate analyses again and were included in the multivariate logistic regression model. Ovarian cancer risk perception was also included in the multivariate analyses with p = 0.15. As was the case with the previous analysis, age, higher education, ovarian cancer worry, and BSE remained statistically significantly associated with CAM use (p ≤ 0.05), with similar odds ratios. Ovarian cancer risk perception was no longer statistically significant.