BRIEF REPORT
Patterns of User Disclosure of Complementary and Alternative Medicine (CAM) Use Jae-Mahn Shim, PhD,* John Schneider, MD, MPH,wz and Farr A. Curlin, MDz
Objective: To investigate patterns of complementary and alternative medicine (CAM) use disclosure across medical and sociobehavioral factors and to provide a model that takes into account factors in explaining those patterns. Subjects: A total of 21,849 CAM use episodes from 7347 respondents in the 2007 US National Health Interview Survey which involves the latest survey on CAM use. Research Design: Respondents were a representative sample of US national population. Logistic hierarchical linear models specify how characteristics of users and their CAM use episodes influence user disclosure behaviors. Results: At the individual level, users were more likely to disclose CAM use to health care professionals when they had health problems and when they were insured. At the episode level, CAM use episodes were more likely to be disclosed when they were intended to treat a specific medical condition and recommended by a health professional. Disclosure rates were high among most susceptible users (ie, sick people intending to treat specific conditions with CAM) and among the biologically based CAM modalities (eg, herbal supplements) that are most likely to produce adverse interactions with conventional biomedical treatments. Conclusions: User disclosure was affected not only by users’ demographic and socioeconomic characteristics but also by episode-specific factors. Efforts to improve provider-user communication of CAM use should consider the varying effects of these factors. Key Words: health communication, provider-user communication, complementary and alternative medicine, disclosure (Med Care 2014;52: 704–708)
M
any patients use complementary and alternative medicine (CAM) along with conventional biomedicine. This poses challenges for health care as use of CAM can lead to delayed conventional treatment,1,2 nonutilization of
From the *Department of Sociology, University of Seoul, Seoul, Korea; wDepartment of Health Studies; and zDepartment of Medicine, University of Chicago, Chicago, IL. The authors declare no conflict of interest. Reprints: Jae-Mahn Shim, PhD, Department of Sociology, University of Seoul, 163 Siripdaero, Dongdaemun-gu, Seoul 130-743, Korea. E-mail: [email protected]. Copyright r 2014 by Lippincott Williams & Wilkins ISSN: 0025-7079/14/5208-0704
704 | www.lww-medicalcare.com
available medical resources,3–5 or unforeseen adverse interactions between conventional biomedical treatments and CAM.6–9 Communication between CAM users and conventional medical service providers is essential for delivering health services effectively and safely.10–13 We explore which patients are talking to conventional medical service providers about CAM use and in what contexts. We examine patterns in users’ disclosure of CAM in the United States. In representative national surveys, 34%– 42% of Americans used at least one of CAM modalities in the past 12 months: 34% in 1990,14 42% in 1998,15 36% in 2002,16 and 40% in 2007.17 The overall rates of individuals disclosing CAM use to conventional medical service providers were estimated at 39.8% in 199014 and 38.5% in 1997.15 These estimates provided little explanation of the determinants of disclosure. Meanwhile, studies of smaller nonrepresentative populations reported widely varying disclosure rates. A systematic review of 12 disclosure studies found that disclosure rates ranged from 12% to 90%.12 Studies of CAM users who had particular medical conditions (ie, cancer patients) reported inconsistent nondisclosure rates fluctuating from 20% to 77%.10 To examine what has produced these variations and to identify medical and sociobehavioral factors influencing disclosure behaviors, we used the latest representative US national data from the 2007 National Health Interview Survey (NHIS). Existing studies suggested potential factors influencing disclosure behaviors. First, studies of patient groups suggested to us that disclosure rates were slightly higher among CAM users with specific medical conditions compared with the general population many of whom had no medical conditions.18–22 Second, users with better access to health care (ie, the insured) were more likely to disclose their CAM use.23,24 Third, racial/ethnic minorities were less likely to disclose CAM use, whereas age, sex, income, marital status, and education were often found to influence disclosure.23–27 Fourth, disclosure rates varied among different CAM modalities, such as “provider-based” CAM and “self-care” CAM.23,28 Fifth, doctors’ attitudes and recommendations were implied to be a factor.29,30 Existing literature, however, did not demonstrate how these effects on disclosure would be related to one another in a multivariate analysis. More importantly, all prior studies investigated CAM disclosure at the individual level and not at the episode level. This is a significant limitation. Disclosure occurs at the episode level and should be examined at that level. Individuals may use multiple CAM modalities. Medical Care
Volume 52, Number 8, August 2014
Medical Care
Volume 52, Number 8, August 2014
Patterns of User Disclosure of CAM
They may disclose some episodes and not others, depending on episode-level characteristics. A multilevel approach is necessary for investigating these variations. We conducted a multivariate analysis with a multilevel design that examined users’ CAM use disclosure at the episode level while incorporating individual-level factors as well. To do this, we reorganized potential determinants of disclosure behaviors at 2 levels: the individual user level and the CAM use episode level. We then tested hypotheses at both levels. At the individual level, we hypothesized that users would be more likely to disclose CAM use to medical service providers when they have medical conditions and when they are insured. At the episode level, we hypothesized that CAM use episodes would be more likely to be disclosed when they are intended to treat a specific medical condition and recommended by a professional medical service provider. We controlled for a set of other individual-level factors (eg, demographic characteristics) and episode-level characteristics (eg, CAM modalities).
modality. CAM use for medical condition measured respondents’ answer to the following question: “did you use [a CAM modality] for a specific health problem or condition?” (1 = yes, 0 = no). CAM recommender measured who recommended CAM: friends/family/coworkers, health care provider, neither, or both. CAM modality referred to which of the 17 modalities was used for that episode. We included vitamins/minerals in the analysis. A sensitivity analysis was performed in which episodes of vitamins/minerals were excluded. The results did not change the conclusions, and are available from the authors. We used Logistic Hierarchical Linear Modeling (HLM) because the dependent variable is a dichotomous variable.36 The following analyses put CAM use disclosure on the left side of the equation with both individual-level and episode-level predictors on the right side. We applied the weights provided by the survey at both individual and episode levels. Multilevel analysis package HLM 6 was used.37
METHODS
A total of 21,849 CAM use episodes were reported by 7347 respondents over the previous 12 months. The average number of CAM use episodes per respondent was 2.97. Thirteen percent (n = 952) of respondents reported a single episode of using CAM during this period; the rest reported multiple episodes (Tables 1 and 2). Key findings are reported in terms of the odds of disclosure (Table 3). At the individual level, CAM users were more likely to disclose use as their comorbidities increased. As the CCI increased, the odds increased (OR, 1.37 for CCI = 1; OR, 1.48 for CCI = 2; OR, 1.92 for CCI = 3; OR, 2.32 for CCI = 4+). Second, the odds of disclosure for people with health insurance were >2 times those of the uninsured (OR, 2.27; 95% CI, 1.89–2.73). At the episode level, users were more likely to disclose episodes when CAM was used to treat medical conditions than when it was not (OR, 1.31; 95% CI, 1.17–1.47). There was a facilitating effect of professional recommendations on disclosure. The odds of disclosure for episodes with professional recommendations were about 3 times those of episodes without recommendations from professionals or family and friends (OR, 2.76; 95% CI, 2.41–3.16). Recommendations by family and friends did not significantly affect disclosure rates. Results support our hypotheses at 2 levels. The sick and the insured were more likely to talk about their CAM use in the user-provider communication. The episodes intended for specific medical conditions and recommended by medical professionals were more likely to be disclosed.
RESULTS We used data from the 2007 NHIS, the latest survey on health care behaviors regarding 18 CAM modalities in the United States.31,32 These 18 modalities were acupuncture, Ayurveda, biofeedback, chelation therapy, chiropractic or osteopathic manipulation, energy healing therapy, hypnosis, massage, naturopathy, traditional healers, movement therapies, herbal medicine/supplements, vitamins/minerals, homeopathy, special diets, Yoga/Tai Chi/Qigong, relaxation techniques, and prayer. We excluded prayer, because the survey did not assess disclosure behaviors regarding prayer. We analyzed 7347 respondents who reported that they had used at least 1 of the remaining 17 CAM modalities in the previous 12 months and who had no missing values for the set of variables included in the following analytical models. These respondents reported a total of 21,849 CAM use episodes. The outcome variable was CAM use disclosure at the episode level. This was a binary measure determined by respondents’ answers (for each CAM use episode) to the following question: “did you let any of conventional medical professionals know about your use of [a CAM modality]?” (yes = 1, no = 0). The survey did not address who initiated communication about CAM use. Predictor variables at the individual level were health insurance status and comorbidity with sex, age, region, race, marital status, education, and annual household income being control variables. All but age were categorical variables. Health insurance status was coded 1 if a respondent possessed an insurance plan, whether public or private. Comorbidity was measured by the Charlson Comorbidity Index (CCI) applied to NHIS data.33–35 The CCI measured the degree to which individuals were ever diagnosed by a health professional with a list of medical conditions, including heart attack, stroke, emphysema, asthma, ulcer, liver problems, diabetes, arthritis, kidney problems, and cancer. Predictor variables at the episode level were CAM use for medical condition, CAM recommender, and CAM r
2014 Lippincott Williams & Wilkins
DISCUSSION Our multivariate approach with a multilevel design contributes to CAM use disclosure research in 2 ways. First, the multilevel design allowed us to specify within-individual variations across multiple episodes of CAM use and disclosure. User disclosure was affected not only by users’ demographic and socioeconomic characteristics but also by episode-specific factors, such as treatment motivations, www.lww-medicalcare.com |
705
Medical Care
Shim et al
TABLE 1. Characteristics of Respondents (N = 7347)
Volume 52, Number 8, August 2014
TABLE 2. Characteristics of Episodes (N = 21,849)
No. Respondents (%) Variables Sex Female Male Age, mean (SD) Region Northeast Midwest South West Race Hispanic White Black Asian Other Marital status Married Widow/divorce/separated Never married Education Less than high school diploma Less than BA degree BA degree and more Income < $20,000 $20,000–$45,000 $45,000–$65,000 Z$65,000 Health insurance status Insured Uninsured Charlson Comorbidity Index 0 1 2 3 Z4
Unweighted
Weighted
4110 (55.9) 3237 (44.1) 42.3 (13.6)
3841 (52.3) 3506 (47.7) 41.8 (13.6)
1248 1805 2431 1863
(17.0) (24.6) (33.1) (25.4)
1276 (17.4) 1944 (26.5) 2382 (32.4) 1744(23.7)
960 4947 981 398 61
(13.1) (67.3) (13.4) (5.4) (0.8)
696 5544 699 345 63
(9.5) (75.5) (9.5) (4.7) (0.9)
3931 (53.5) 1583 (21.5) 1833 (24.9)
4759 (64.8) 1044 (14.2) 1544 (21.0)
549 (7.5) 4073 (55.4) 2725 (37.1)
490 (6.7) 4084 (55.6) 2773 (37.7)
2154 2678 1218 1297
2119 2606 1267 1354
(29.3) (36.5) (16.6) (17.7)
(28.8) (35.5) (17.3) (18.4)
6327 (86.1) 1020 (13.9)
6423 (87.4) 924 (12.6)
4405 1459 670 452 361
4409 1474 655 466 342
(60.0) (19.9) (9.1) (6.2) (4.9)
(60.0) (20.1) (8.9) (6.3) (4.7)
CAM modalities in use, and medical professionals’ attitudes and recommendations regarding CAM use. From the perspective of health services research, we stress the significant role that medical professionals play. Second, the multivariate analysis revealed that health status and insurance status were more significant determinants than race/ethnicity, income, or education. In addition, we found that disclosure differences between CAM modalities persisted even after individuallevel and episode-level factors were controlled for. Our findings have several implications for concerns about public health challenges and professional medical practices at the intersection of conventional biomedicine and CAM. First, disclosure rates were reportedly low among the general US population, many of whom did not have medical conditions. Users, however, were found to be much more likely to disclose CAM use when they were sick and when they used CAM for a medical condition—2 categories of users who are most susceptible to adverse interactions between conventional treatments and CAM. With regard to the finding that CAM users for medical treatment and those for general health promotion have different characteristics,38 our findings add that they show a significant difference in user disclosure behavior as well.
706 | www.lww-medicalcare.com
No. Episodes (%) Variables
Unweighted
CAM use disclosure Yes No CAM modality Alternative whole medical system Acupuncture Ayurveda Chelation therapy Naturopathy Traditional healers Homeopathy Biologically based therapy Herbal medicine/supplements Vitamins/minerals Special diets Manipulation-based therapy Chiropractic and osteopathic manipulation Massage Movement therapies Mind-body therapy Biofeedback Energy healing therapy Hypnosis Yoga/Taichi/Qigong Relaxation CAM recommender Neither friends/family nor care professionals Friends/family but not care professionals Care professionals but not friends/family Both friends/family and care professionals CAM for treating medical condition Yes No
Weighted
10201 (46.7) 10384 (47.5) 11648 (53.3) 11465 (52.5) 199 12 5 53 44 259
(0.9) (0.1) (0.02) (0.2) (0.2) (1.2)
174 16 8 48 39 250
(0.8) (0.1) (0.04) (0.2) (0.2) (1.1)
3328 (15.2) 3310 (15.1) 11370 (52.0) 11410 (52.2) 528 (2.4) 525 (2.4) 1154 (5.3) 1201 (5.5) 226 (1.0)
1208 (5.5) 1222 (5.6) 214 (1.0)
13 105 35 995 2322
17 91 31 988 2298
13576 4454 2403 1416
(0.1) (0.5) (0.2) (4.6) (10.6)
(0.1) (0.4) (0.1) (4.5) (10.5)
(62.1) 13520 (61.9) (20.4) 4546 (20.8) (11.0) 2350 (10.8) (6.5) 1433 (6.6)
4258 (19.5) 4228 (19.4) 17591 (80.5) 17621 (80.6)
CAM indicates complementary and alternative medicine.
Second, varying disclosure rates across different CAM modalities have additional implications. It is reassuring that users were more likely to disclose “biologically based therapies” (eg, herbs and vitamins/minerals); these are the therapies that are most likely to have adverse interactions with conventional pharmacotherapy39,40 and, thus, require the most user-provider communication of CAM use. Relatively high disclosure rates in “alternative whole medical systems” (eg, acupuncture) suggest that we can be optimistic about user-provider communication on CAM among ethnic minorities or immigrants of Asian origins, groups most likely engaged in these CAM modalities in the United States. Third, we draw attention to the significant gap in CAM disclosure between the insured and the uninsured. Uninsured people are less likely to see physicians and CAM practitioners.41 They discuss their uses of CAM to these practitioners less often. Thus, existing health disparities between the insured and the uninsured could be exacerbated by this disparity in CAM use disclosure and chances of adverse medical events. There are limitations to this study. The survey asked individuals whether they disclosed their use of CAM to their medical professionals. It did not include further information other than self-reports on these questions. It did not provide r
2014 Lippincott Williams & Wilkins
Medical Care
Volume 52, Number 8, August 2014
Patterns of User Disclosure of CAM
TABLE 3. Odds Ratios From Logistic Hierarchical Linear Models (HLM) of CAM Use Disclosure Regressed on Individuallevel and CAM Use Episode-level Variables Odds Ratio (95% CI)
P
Individual-level intercept 0.177 (0.122,0.258) < 0.001 Individual-level variable Sex (female = 1, male = 0) 1.722 (1.535,1.931) < 0.001 Age 1.018 (1.013,1.023) < 0.001 Region (reference = Northeast) Midwest 1.015 (0.863,1.193) 0.86 South 0.820 (0.701,0.960) 0.01 West 0.721 (0.612,0.849) < 0.001 Race (reference = white) Hispanic 0.872 (0.733,1.036) 0.12 Black 1.055 (0.889,1.254) 0.54 Asian 0.760 (0.577,1.001) 0.05 Other 1.126 (0.354,3.583) 0.84 Marital status (reference = married) Widow/divorce/separated 0.768 (0.667,0.885) < 0.001 Never married 0.713 (0.617,0.824) < 0.001 Education (reference = less than high school diploma) Less than BA degree 1.128 (0.892,1.425) 0.32 BA degree and more 1.333 (1.041,1.706) 0.02 Household income (reference = < $20,000) $20,000–$45,000 1.002 (0.872,1.152) 0.98 $45,000–$65,000 0.953 (0.796,1.141) 0.60 Z$65,000 0.930 (0.777,1.112) 0.43 Health insurance status (insured = 1, 2.268 (1.886,2.728) < 0.001 uninsured = 0) Charlson Comorbidity Index (reference = 0) 1 1.370 (1.198,1.566) < 0.001 2 1.478 (1.229,1.779) < 0.001 3 1.919 (1.483,2.484) < 0.001 Z4 2.326 (1.759,3.075) < 0.001 Episode-level variable CAM modality (reference = vitamins/minerals in biologically based therapy) Alternative whole medical system Acupuncture 0.670 (0.475,0.946) 0.02 Ayurveda 0.431 (0.171,1.086) 0.07 Chelation therapy 1.315 (0.179,9.652) 0.79 Naturopathy 0.674 (0.390,1.165) 0.16 Traditional healers 0.131 (0.038,0.455) 0.002 Homeopathy 0.307 (0.227,0.416) < 0.001 Biologically based therapy Herbal medicine/supplements 0.566 (0.505,0.635) < 0.001 Special diets 0.525 (0.405,0.680) < 0.001 Manipulation-based therapy Chiropractic and osteopathic 0.489 (0.408,0.586) < 0.001 manipulation Massage 0.194 (0.165,0.229) < 0.001 Movement therapies 0.195 (0.126,0.300) < 0.001 Mind-body therapy Biofeedback 0.385 (0.119,1.243) 0.11 Energy healing therapy 0.240 (0.158,0.364) < 0.001 Hypnosis 0.103 (0.027,0.390) 0.001 Yoga/Taichi/Qigong 0.229 (0.190,0.276) < 0.001 Relaxation 0.137 (0.118,0.159) < 0.001 CAM recommender (reference = neither friends/family nor care professionals) Friends/family but not care professionals 1.011 (0.924,1.107) 0.81 Care professionals but not friends/family 2.756 (2.408,3.155) < 0.001 Both friends/family and care 3.060 (2.537,3.692) < 0.001 professionals CAM for treating medical condition 1.310 (1.165,1.474) < 0.001 (yes = 1, no = 0) CAM indicates complementary and alternative medicine.
r
2014 Lippincott Williams & Wilkins
nuanced information about how the disclosure was made; nor did it investigate qualitative aspects of user-provider communications, such as how the communication was initiated and what was discussed. In addition, variations in disclosure between different CAM modalities merit future research on what underlies these variations. These variations can be related to different cultural valences associated with CAM modalities by different sectors of the American public or to differing regulatory and medical educational contexts across CAM modalities. The current survey data were not equipped to address these aspects. Notwithstanding these limitations, our study demonstrates that CAM use disclosure varies significantly, depending on users’ demographic, socioeconomic, and medical characteristics on the one hand and characteristics of CAM use episodes on the other. REFERENCES 1. Ayers SL, Kronenfeld JJ. Delays in seeking conventional medical care and complementary and alternative medicine utilization. Health Serv Res. 2012;47:2081–2096. 2. Tom K, Farrell TW. The complementarity and substitution between unconventional and mainstream medicine among racial and ethnic groups in the United States. Health Serv Res. 2007;42:811–826. 3. Downey L, Tyree PT, Huebner CE, et al. Pediatric vaccination and vaccine-preventable disease acquisition: associations with care by complementary and alternative medicine providers. Matern Child Health J. 2010;14:922–930. 4. Downey L, Tyree PT, Lafferty WE. Preventive screening of women who use complementary and alternative medicine providers. J Womens Health. 2009;18:1133–1143. 5. Pisani MJ, Paga´n JA, Lackan NA, et al. Substitution of formal health care services by Latinos/Hispanics in the US-Mexico border region of South Texas. Med Care. 2012;50:885–889. 6. Ang-Lee MK, Moss J, Yuan C-S. Herbal medicines and perioperative care. JAMA. 2001;286:208–216. 7. Furrer M, Naegeli B, Bertel O. Hazards of an alternative medicine device in a patient with a pacemaker. N Engl J Med. 2004;350: 1688–1690. 8. Hu Z, Yang X, Paul Chi Lui H, et al. Herb-drug interactions: a literature review. Drugs. 2005;65:1239–1282. 9. Gilmour J, Harrison C, Asadi L, et al. Natural health product-drug interactions: evolving responsibilities to take complementary and alternative medicine into account. Pediatrics. 2011;128:S155–S160. 10. Davis EL, Oh B, Butow PN, et al. Cancer patient disclosure and patientdoctor communication of complementary and alternative medicine use: a systematic review. Oncologist. 2012;17:1475–1481. 11. Langlois-Klassen D, Kipp W, Rubaale T. Who’s talking? Communication between health providers and HIV-infected adults related to herbal medicine for AIDS treatment in Western Uganda. Soc Sci Med. 2008;67:165–176. 12. Robinson A, McGrail MR. Disclosure of CAM use to medical practitioners: a review of qualitative and quantitative studies. Complement Ther Med. 2004;12:90–98. 13. Stevenson FA, Britten N, Barry CA, et al. Self-treatment and its discussion in medical consultations: how is medical pluralism managed in practice? Soc Sci Med. 2003;57:513–527. 14. Eisenberg DM, Kessler RC, Foster C, et al. Unconventional medicine in the united states: prevalence, costs, and patterns of use. N Engl J Med. 1993;328:246–252. 15. Eisenberg DM, Davis RB, Ettner SL, et al. Trends in alternative medicine use in the United States, 1990-1997. JAMA. 1998;280: 1569–1575. 16. Barnes PM, Powell-Griner E, McFann K, et al. Complementary and Alternative Medicine Use Among Adults: United States, 2002. Advance Data From Vital and Health Statistics. Hyattsville, MD: National Center for Health Statistics; 2004. 17. Barnes PM, Bloom B, Nahin RL. Complementary and Alternative Medicine Use Among Adults and Children: United States, 2007.
www.lww-medicalcare.com |
707
Shim et al
18. 19. 20. 21. 22. 23. 24. 25. 26. 27.
28.
National Health Statistics Reports. Hyattsville, MD: National Center for Health Statistics; 2008. Hilsden RJ, Scott CM, Verhoef MJ. Complementary medicine use by patients with inflammatory bowel disease. Am J Gastroenterol. 1998;93:697–701. Nam RK, Fleshner N, Rakovitch E, et al. Prevalence and patterns of the use of complementary therapies among prostate cancer patients: an epidemiological analysis. J Urol. 1999;161:1521–1524. Rao JK, Mihaliak K, Kroenke K, et al. Use of complementary therapies for arthritis among patients of rheumatologists. Ann Intern Med. 1999;131:409–416. Saxe GA, Madlensky L, Kealey S, et al. Disclosure to physicians of CAM use by breast cancer patients: findings from the Women’s Healthy Eating and Living Study. Integr Cancer Ther. 2008;7:122–129. Yates J, Mustian K, Morrow G, et al. Prevalence of complementary and alternative medicine use in cancer patients during treatment. Support Care Cancer. 2005;13:806–811. Chao M, Wade C, Kronenberg F. Disclosure of complementary and alternative medicine to conventional medical providers: variation by race/ethnicity and type of cam. J Natl Med Assoc. 2008;100:1341–1349. Mehta DH, Gardiner PM, Phillips RS, et al. Herbal and dietary supplement disclosure to health care providers by individuals with chronic conditions. J Altern Complement Med. 2008;14:1263–1269. Mehta DH, Phillips RS, Davis RB, et al. Use of complementary and alternative therapies by Asian Americans: results from the National Health Interview Survey. J Gen Intern Med. 2007;22:762–767. Thomson P, Jones J, Evans JM, et al. Factors influencing the use of complementary and alternative medicine and whether patients inform their primary care physician. Complement Ther Med. 2012;20:45–53. Graham RE, Ahn AC, Davis RB, et al. Use of complementary and alternative medical therapies among racial and ethnic minority adults: results from the 2002 National Health Interview Survey. J Natl Med Assoc. 2005;97:535–545. Gardiner P, Graham RE, Legedza ATR, et al. Factors associated with dietary supplement use among prescription medication users. Arch Intern Med. 2006;166:1968–1974.
708 | www.lww-medicalcare.com
Medical Care
Volume 52, Number 8, August 2014
29. Sleath B, Callahan L, DeVellis RF, et al. Patients’ perceptions of primary care physicians’ participatory decision-making style and communication about complementary and alternative medicine for arthritis. J Altern Complement Med. 2005;11:449–453. 30. Chung VCH, Hillier S, Lau CH, et al. Referral to and attitude towards traditional Chinese medicine amongst western medical doctors in postcolonial Hong Kong. Soc Sci Med. 2011;72:247–255. 31. Kaptchuk TJ, Eisenberg DM. Varieties of healing. 2: a taxonomy of unconventional healing practices. Ann Intern Med. 2001;135:196–204. 32. Kaptchuk TJ, Eisenberg DM. Varieties of healing. 1: medical pluralism in the United States. Ann Intern Med. 2001;135:189–195. 33. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383. 34. Nguyen LT, Davis RB, Kaptchuk TJ, et al. Use of complementary and alternative medicine and self-rated health status: results from a national survey. J Gen Intern Med. 2011;26:399–404. 35. Nerurkar A, Yeh G, Davis RB, et al. When conventional medical providers recommend unconventional medicine: results of a national study. Arch Intern Med. 2011;171:862–864. 36. Snijders TAB. Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling. London; Thousand Oaks, CA: SAGE; 2012. 37. Raudenbush SW, Bryk AS, Congdon R. HLM 6 for Windows. Lincolnwood, IL: Scientific Software International Inc; 2004. 38. Davis MA, West AN, Weeks WB, et al. Health behaviors and utilization among users of complementary and alternative medicine for treatment versus health promotion. Health Serv Res. 2011;46:1402–1416. 39. Gryzlak BM, Wallace RB, Zimmerman MB, et al. National surveillance of herbal dietary supplement exposures: the Poison Control Center Experience. Pharmacoepidemiol Drug Saf. 2007;16:947–957. 40. Gelo F, Harris R. The secret garden. Fam Med. 2009;41:391–392. 41. Shih S-F, Lew-Ting C-Y, Chang H-Y, et al. Insurance covered and noncovered complementary and alternative medicine utilisation among adults in Taiwan. Soc Sci Med. 2008;67:1183–1189.
r
2014 Lippincott Williams & Wilkins
Patterns of User Disclosure of Complementary and Alternative Medicine (CAM) Use Jae-Mahn Shim, PhD,* John Schneider, MD, MPH,wz and Farr A. Curlin, MDz
Objective: To investigate patterns of complementary and alternative medicine (CAM) use disclosure across medical and sociobehavioral factors and to provide a model that takes into account factors in explaining those patterns. Subjects: A total of 21,849 CAM use episodes from 7347 respondents in the 2007 US National Health Interview Survey which involves the latest survey on CAM use. Research Design: Respondents were a representative sample of US national population. Logistic hierarchical linear models specify how characteristics of users and their CAM use episodes influence user disclosure behaviors. Results: At the individual level, users were more likely to disclose CAM use to health care professionals when they had health problems and when they were insured. At the episode level, CAM use episodes were more likely to be disclosed when they were intended to treat a specific medical condition and recommended by a health professional. Disclosure rates were high among most susceptible users (ie, sick people intending to treat specific conditions with CAM) and among the biologically based CAM modalities (eg, herbal supplements) that are most likely to produce adverse interactions with conventional biomedical treatments. Conclusions: User disclosure was affected not only by users’ demographic and socioeconomic characteristics but also by episode-specific factors. Efforts to improve provider-user communication of CAM use should consider the varying effects of these factors. Key Words: health communication, provider-user communication, complementary and alternative medicine, disclosure (Med Care 2014;52: 704–708)
M
any patients use complementary and alternative medicine (CAM) along with conventional biomedicine. This poses challenges for health care as use of CAM can lead to delayed conventional treatment,1,2 nonutilization of
From the *Department of Sociology, University of Seoul, Seoul, Korea; wDepartment of Health Studies; and zDepartment of Medicine, University of Chicago, Chicago, IL. The authors declare no conflict of interest. Reprints: Jae-Mahn Shim, PhD, Department of Sociology, University of Seoul, 163 Siripdaero, Dongdaemun-gu, Seoul 130-743, Korea. E-mail: [email protected]. Copyright r 2014 by Lippincott Williams & Wilkins ISSN: 0025-7079/14/5208-0704
704 | www.lww-medicalcare.com
available medical resources,3–5 or unforeseen adverse interactions between conventional biomedical treatments and CAM.6–9 Communication between CAM users and conventional medical service providers is essential for delivering health services effectively and safely.10–13 We explore which patients are talking to conventional medical service providers about CAM use and in what contexts. We examine patterns in users’ disclosure of CAM in the United States. In representative national surveys, 34%– 42% of Americans used at least one of CAM modalities in the past 12 months: 34% in 1990,14 42% in 1998,15 36% in 2002,16 and 40% in 2007.17 The overall rates of individuals disclosing CAM use to conventional medical service providers were estimated at 39.8% in 199014 and 38.5% in 1997.15 These estimates provided little explanation of the determinants of disclosure. Meanwhile, studies of smaller nonrepresentative populations reported widely varying disclosure rates. A systematic review of 12 disclosure studies found that disclosure rates ranged from 12% to 90%.12 Studies of CAM users who had particular medical conditions (ie, cancer patients) reported inconsistent nondisclosure rates fluctuating from 20% to 77%.10 To examine what has produced these variations and to identify medical and sociobehavioral factors influencing disclosure behaviors, we used the latest representative US national data from the 2007 National Health Interview Survey (NHIS). Existing studies suggested potential factors influencing disclosure behaviors. First, studies of patient groups suggested to us that disclosure rates were slightly higher among CAM users with specific medical conditions compared with the general population many of whom had no medical conditions.18–22 Second, users with better access to health care (ie, the insured) were more likely to disclose their CAM use.23,24 Third, racial/ethnic minorities were less likely to disclose CAM use, whereas age, sex, income, marital status, and education were often found to influence disclosure.23–27 Fourth, disclosure rates varied among different CAM modalities, such as “provider-based” CAM and “self-care” CAM.23,28 Fifth, doctors’ attitudes and recommendations were implied to be a factor.29,30 Existing literature, however, did not demonstrate how these effects on disclosure would be related to one another in a multivariate analysis. More importantly, all prior studies investigated CAM disclosure at the individual level and not at the episode level. This is a significant limitation. Disclosure occurs at the episode level and should be examined at that level. Individuals may use multiple CAM modalities. Medical Care
Volume 52, Number 8, August 2014
Medical Care
Volume 52, Number 8, August 2014
Patterns of User Disclosure of CAM
They may disclose some episodes and not others, depending on episode-level characteristics. A multilevel approach is necessary for investigating these variations. We conducted a multivariate analysis with a multilevel design that examined users’ CAM use disclosure at the episode level while incorporating individual-level factors as well. To do this, we reorganized potential determinants of disclosure behaviors at 2 levels: the individual user level and the CAM use episode level. We then tested hypotheses at both levels. At the individual level, we hypothesized that users would be more likely to disclose CAM use to medical service providers when they have medical conditions and when they are insured. At the episode level, we hypothesized that CAM use episodes would be more likely to be disclosed when they are intended to treat a specific medical condition and recommended by a professional medical service provider. We controlled for a set of other individual-level factors (eg, demographic characteristics) and episode-level characteristics (eg, CAM modalities).
modality. CAM use for medical condition measured respondents’ answer to the following question: “did you use [a CAM modality] for a specific health problem or condition?” (1 = yes, 0 = no). CAM recommender measured who recommended CAM: friends/family/coworkers, health care provider, neither, or both. CAM modality referred to which of the 17 modalities was used for that episode. We included vitamins/minerals in the analysis. A sensitivity analysis was performed in which episodes of vitamins/minerals were excluded. The results did not change the conclusions, and are available from the authors. We used Logistic Hierarchical Linear Modeling (HLM) because the dependent variable is a dichotomous variable.36 The following analyses put CAM use disclosure on the left side of the equation with both individual-level and episode-level predictors on the right side. We applied the weights provided by the survey at both individual and episode levels. Multilevel analysis package HLM 6 was used.37
METHODS
A total of 21,849 CAM use episodes were reported by 7347 respondents over the previous 12 months. The average number of CAM use episodes per respondent was 2.97. Thirteen percent (n = 952) of respondents reported a single episode of using CAM during this period; the rest reported multiple episodes (Tables 1 and 2). Key findings are reported in terms of the odds of disclosure (Table 3). At the individual level, CAM users were more likely to disclose use as their comorbidities increased. As the CCI increased, the odds increased (OR, 1.37 for CCI = 1; OR, 1.48 for CCI = 2; OR, 1.92 for CCI = 3; OR, 2.32 for CCI = 4+). Second, the odds of disclosure for people with health insurance were >2 times those of the uninsured (OR, 2.27; 95% CI, 1.89–2.73). At the episode level, users were more likely to disclose episodes when CAM was used to treat medical conditions than when it was not (OR, 1.31; 95% CI, 1.17–1.47). There was a facilitating effect of professional recommendations on disclosure. The odds of disclosure for episodes with professional recommendations were about 3 times those of episodes without recommendations from professionals or family and friends (OR, 2.76; 95% CI, 2.41–3.16). Recommendations by family and friends did not significantly affect disclosure rates. Results support our hypotheses at 2 levels. The sick and the insured were more likely to talk about their CAM use in the user-provider communication. The episodes intended for specific medical conditions and recommended by medical professionals were more likely to be disclosed.
RESULTS We used data from the 2007 NHIS, the latest survey on health care behaviors regarding 18 CAM modalities in the United States.31,32 These 18 modalities were acupuncture, Ayurveda, biofeedback, chelation therapy, chiropractic or osteopathic manipulation, energy healing therapy, hypnosis, massage, naturopathy, traditional healers, movement therapies, herbal medicine/supplements, vitamins/minerals, homeopathy, special diets, Yoga/Tai Chi/Qigong, relaxation techniques, and prayer. We excluded prayer, because the survey did not assess disclosure behaviors regarding prayer. We analyzed 7347 respondents who reported that they had used at least 1 of the remaining 17 CAM modalities in the previous 12 months and who had no missing values for the set of variables included in the following analytical models. These respondents reported a total of 21,849 CAM use episodes. The outcome variable was CAM use disclosure at the episode level. This was a binary measure determined by respondents’ answers (for each CAM use episode) to the following question: “did you let any of conventional medical professionals know about your use of [a CAM modality]?” (yes = 1, no = 0). The survey did not address who initiated communication about CAM use. Predictor variables at the individual level were health insurance status and comorbidity with sex, age, region, race, marital status, education, and annual household income being control variables. All but age were categorical variables. Health insurance status was coded 1 if a respondent possessed an insurance plan, whether public or private. Comorbidity was measured by the Charlson Comorbidity Index (CCI) applied to NHIS data.33–35 The CCI measured the degree to which individuals were ever diagnosed by a health professional with a list of medical conditions, including heart attack, stroke, emphysema, asthma, ulcer, liver problems, diabetes, arthritis, kidney problems, and cancer. Predictor variables at the episode level were CAM use for medical condition, CAM recommender, and CAM r
2014 Lippincott Williams & Wilkins
DISCUSSION Our multivariate approach with a multilevel design contributes to CAM use disclosure research in 2 ways. First, the multilevel design allowed us to specify within-individual variations across multiple episodes of CAM use and disclosure. User disclosure was affected not only by users’ demographic and socioeconomic characteristics but also by episode-specific factors, such as treatment motivations, www.lww-medicalcare.com |
705
Medical Care
Shim et al
TABLE 1. Characteristics of Respondents (N = 7347)
Volume 52, Number 8, August 2014
TABLE 2. Characteristics of Episodes (N = 21,849)
No. Respondents (%) Variables Sex Female Male Age, mean (SD) Region Northeast Midwest South West Race Hispanic White Black Asian Other Marital status Married Widow/divorce/separated Never married Education Less than high school diploma Less than BA degree BA degree and more Income < $20,000 $20,000–$45,000 $45,000–$65,000 Z$65,000 Health insurance status Insured Uninsured Charlson Comorbidity Index 0 1 2 3 Z4
Unweighted
Weighted
4110 (55.9) 3237 (44.1) 42.3 (13.6)
3841 (52.3) 3506 (47.7) 41.8 (13.6)
1248 1805 2431 1863
(17.0) (24.6) (33.1) (25.4)
1276 (17.4) 1944 (26.5) 2382 (32.4) 1744(23.7)
960 4947 981 398 61
(13.1) (67.3) (13.4) (5.4) (0.8)
696 5544 699 345 63
(9.5) (75.5) (9.5) (4.7) (0.9)
3931 (53.5) 1583 (21.5) 1833 (24.9)
4759 (64.8) 1044 (14.2) 1544 (21.0)
549 (7.5) 4073 (55.4) 2725 (37.1)
490 (6.7) 4084 (55.6) 2773 (37.7)
2154 2678 1218 1297
2119 2606 1267 1354
(29.3) (36.5) (16.6) (17.7)
(28.8) (35.5) (17.3) (18.4)
6327 (86.1) 1020 (13.9)
6423 (87.4) 924 (12.6)
4405 1459 670 452 361
4409 1474 655 466 342
(60.0) (19.9) (9.1) (6.2) (4.9)
(60.0) (20.1) (8.9) (6.3) (4.7)
CAM modalities in use, and medical professionals’ attitudes and recommendations regarding CAM use. From the perspective of health services research, we stress the significant role that medical professionals play. Second, the multivariate analysis revealed that health status and insurance status were more significant determinants than race/ethnicity, income, or education. In addition, we found that disclosure differences between CAM modalities persisted even after individuallevel and episode-level factors were controlled for. Our findings have several implications for concerns about public health challenges and professional medical practices at the intersection of conventional biomedicine and CAM. First, disclosure rates were reportedly low among the general US population, many of whom did not have medical conditions. Users, however, were found to be much more likely to disclose CAM use when they were sick and when they used CAM for a medical condition—2 categories of users who are most susceptible to adverse interactions between conventional treatments and CAM. With regard to the finding that CAM users for medical treatment and those for general health promotion have different characteristics,38 our findings add that they show a significant difference in user disclosure behavior as well.
706 | www.lww-medicalcare.com
No. Episodes (%) Variables
Unweighted
CAM use disclosure Yes No CAM modality Alternative whole medical system Acupuncture Ayurveda Chelation therapy Naturopathy Traditional healers Homeopathy Biologically based therapy Herbal medicine/supplements Vitamins/minerals Special diets Manipulation-based therapy Chiropractic and osteopathic manipulation Massage Movement therapies Mind-body therapy Biofeedback Energy healing therapy Hypnosis Yoga/Taichi/Qigong Relaxation CAM recommender Neither friends/family nor care professionals Friends/family but not care professionals Care professionals but not friends/family Both friends/family and care professionals CAM for treating medical condition Yes No
Weighted
10201 (46.7) 10384 (47.5) 11648 (53.3) 11465 (52.5) 199 12 5 53 44 259
(0.9) (0.1) (0.02) (0.2) (0.2) (1.2)
174 16 8 48 39 250
(0.8) (0.1) (0.04) (0.2) (0.2) (1.1)
3328 (15.2) 3310 (15.1) 11370 (52.0) 11410 (52.2) 528 (2.4) 525 (2.4) 1154 (5.3) 1201 (5.5) 226 (1.0)
1208 (5.5) 1222 (5.6) 214 (1.0)
13 105 35 995 2322
17 91 31 988 2298
13576 4454 2403 1416
(0.1) (0.5) (0.2) (4.6) (10.6)
(0.1) (0.4) (0.1) (4.5) (10.5)
(62.1) 13520 (61.9) (20.4) 4546 (20.8) (11.0) 2350 (10.8) (6.5) 1433 (6.6)
4258 (19.5) 4228 (19.4) 17591 (80.5) 17621 (80.6)
CAM indicates complementary and alternative medicine.
Second, varying disclosure rates across different CAM modalities have additional implications. It is reassuring that users were more likely to disclose “biologically based therapies” (eg, herbs and vitamins/minerals); these are the therapies that are most likely to have adverse interactions with conventional pharmacotherapy39,40 and, thus, require the most user-provider communication of CAM use. Relatively high disclosure rates in “alternative whole medical systems” (eg, acupuncture) suggest that we can be optimistic about user-provider communication on CAM among ethnic minorities or immigrants of Asian origins, groups most likely engaged in these CAM modalities in the United States. Third, we draw attention to the significant gap in CAM disclosure between the insured and the uninsured. Uninsured people are less likely to see physicians and CAM practitioners.41 They discuss their uses of CAM to these practitioners less often. Thus, existing health disparities between the insured and the uninsured could be exacerbated by this disparity in CAM use disclosure and chances of adverse medical events. There are limitations to this study. The survey asked individuals whether they disclosed their use of CAM to their medical professionals. It did not include further information other than self-reports on these questions. It did not provide r
2014 Lippincott Williams & Wilkins
Medical Care
Volume 52, Number 8, August 2014
Patterns of User Disclosure of CAM
TABLE 3. Odds Ratios From Logistic Hierarchical Linear Models (HLM) of CAM Use Disclosure Regressed on Individuallevel and CAM Use Episode-level Variables Odds Ratio (95% CI)
P
Individual-level intercept 0.177 (0.122,0.258) < 0.001 Individual-level variable Sex (female = 1, male = 0) 1.722 (1.535,1.931) < 0.001 Age 1.018 (1.013,1.023) < 0.001 Region (reference = Northeast) Midwest 1.015 (0.863,1.193) 0.86 South 0.820 (0.701,0.960) 0.01 West 0.721 (0.612,0.849) < 0.001 Race (reference = white) Hispanic 0.872 (0.733,1.036) 0.12 Black 1.055 (0.889,1.254) 0.54 Asian 0.760 (0.577,1.001) 0.05 Other 1.126 (0.354,3.583) 0.84 Marital status (reference = married) Widow/divorce/separated 0.768 (0.667,0.885) < 0.001 Never married 0.713 (0.617,0.824) < 0.001 Education (reference = less than high school diploma) Less than BA degree 1.128 (0.892,1.425) 0.32 BA degree and more 1.333 (1.041,1.706) 0.02 Household income (reference = < $20,000) $20,000–$45,000 1.002 (0.872,1.152) 0.98 $45,000–$65,000 0.953 (0.796,1.141) 0.60 Z$65,000 0.930 (0.777,1.112) 0.43 Health insurance status (insured = 1, 2.268 (1.886,2.728) < 0.001 uninsured = 0) Charlson Comorbidity Index (reference = 0) 1 1.370 (1.198,1.566) < 0.001 2 1.478 (1.229,1.779) < 0.001 3 1.919 (1.483,2.484) < 0.001 Z4 2.326 (1.759,3.075) < 0.001 Episode-level variable CAM modality (reference = vitamins/minerals in biologically based therapy) Alternative whole medical system Acupuncture 0.670 (0.475,0.946) 0.02 Ayurveda 0.431 (0.171,1.086) 0.07 Chelation therapy 1.315 (0.179,9.652) 0.79 Naturopathy 0.674 (0.390,1.165) 0.16 Traditional healers 0.131 (0.038,0.455) 0.002 Homeopathy 0.307 (0.227,0.416) < 0.001 Biologically based therapy Herbal medicine/supplements 0.566 (0.505,0.635) < 0.001 Special diets 0.525 (0.405,0.680) < 0.001 Manipulation-based therapy Chiropractic and osteopathic 0.489 (0.408,0.586) < 0.001 manipulation Massage 0.194 (0.165,0.229) < 0.001 Movement therapies 0.195 (0.126,0.300) < 0.001 Mind-body therapy Biofeedback 0.385 (0.119,1.243) 0.11 Energy healing therapy 0.240 (0.158,0.364) < 0.001 Hypnosis 0.103 (0.027,0.390) 0.001 Yoga/Taichi/Qigong 0.229 (0.190,0.276) < 0.001 Relaxation 0.137 (0.118,0.159) < 0.001 CAM recommender (reference = neither friends/family nor care professionals) Friends/family but not care professionals 1.011 (0.924,1.107) 0.81 Care professionals but not friends/family 2.756 (2.408,3.155) < 0.001 Both friends/family and care 3.060 (2.537,3.692) < 0.001 professionals CAM for treating medical condition 1.310 (1.165,1.474) < 0.001 (yes = 1, no = 0) CAM indicates complementary and alternative medicine.
r
2014 Lippincott Williams & Wilkins
nuanced information about how the disclosure was made; nor did it investigate qualitative aspects of user-provider communications, such as how the communication was initiated and what was discussed. In addition, variations in disclosure between different CAM modalities merit future research on what underlies these variations. These variations can be related to different cultural valences associated with CAM modalities by different sectors of the American public or to differing regulatory and medical educational contexts across CAM modalities. The current survey data were not equipped to address these aspects. Notwithstanding these limitations, our study demonstrates that CAM use disclosure varies significantly, depending on users’ demographic, socioeconomic, and medical characteristics on the one hand and characteristics of CAM use episodes on the other. REFERENCES 1. Ayers SL, Kronenfeld JJ. Delays in seeking conventional medical care and complementary and alternative medicine utilization. Health Serv Res. 2012;47:2081–2096. 2. Tom K, Farrell TW. The complementarity and substitution between unconventional and mainstream medicine among racial and ethnic groups in the United States. Health Serv Res. 2007;42:811–826. 3. Downey L, Tyree PT, Huebner CE, et al. Pediatric vaccination and vaccine-preventable disease acquisition: associations with care by complementary and alternative medicine providers. Matern Child Health J. 2010;14:922–930. 4. Downey L, Tyree PT, Lafferty WE. Preventive screening of women who use complementary and alternative medicine providers. J Womens Health. 2009;18:1133–1143. 5. Pisani MJ, Paga´n JA, Lackan NA, et al. Substitution of formal health care services by Latinos/Hispanics in the US-Mexico border region of South Texas. Med Care. 2012;50:885–889. 6. Ang-Lee MK, Moss J, Yuan C-S. Herbal medicines and perioperative care. JAMA. 2001;286:208–216. 7. Furrer M, Naegeli B, Bertel O. Hazards of an alternative medicine device in a patient with a pacemaker. N Engl J Med. 2004;350: 1688–1690. 8. Hu Z, Yang X, Paul Chi Lui H, et al. Herb-drug interactions: a literature review. Drugs. 2005;65:1239–1282. 9. Gilmour J, Harrison C, Asadi L, et al. Natural health product-drug interactions: evolving responsibilities to take complementary and alternative medicine into account. Pediatrics. 2011;128:S155–S160. 10. Davis EL, Oh B, Butow PN, et al. Cancer patient disclosure and patientdoctor communication of complementary and alternative medicine use: a systematic review. Oncologist. 2012;17:1475–1481. 11. Langlois-Klassen D, Kipp W, Rubaale T. Who’s talking? Communication between health providers and HIV-infected adults related to herbal medicine for AIDS treatment in Western Uganda. Soc Sci Med. 2008;67:165–176. 12. Robinson A, McGrail MR. Disclosure of CAM use to medical practitioners: a review of qualitative and quantitative studies. Complement Ther Med. 2004;12:90–98. 13. Stevenson FA, Britten N, Barry CA, et al. Self-treatment and its discussion in medical consultations: how is medical pluralism managed in practice? Soc Sci Med. 2003;57:513–527. 14. Eisenberg DM, Kessler RC, Foster C, et al. Unconventional medicine in the united states: prevalence, costs, and patterns of use. N Engl J Med. 1993;328:246–252. 15. Eisenberg DM, Davis RB, Ettner SL, et al. Trends in alternative medicine use in the United States, 1990-1997. JAMA. 1998;280: 1569–1575. 16. Barnes PM, Powell-Griner E, McFann K, et al. Complementary and Alternative Medicine Use Among Adults: United States, 2002. Advance Data From Vital and Health Statistics. Hyattsville, MD: National Center for Health Statistics; 2004. 17. Barnes PM, Bloom B, Nahin RL. Complementary and Alternative Medicine Use Among Adults and Children: United States, 2007.
www.lww-medicalcare.com |
707
Shim et al
18. 19. 20. 21. 22. 23. 24. 25. 26. 27.
28.
National Health Statistics Reports. Hyattsville, MD: National Center for Health Statistics; 2008. Hilsden RJ, Scott CM, Verhoef MJ. Complementary medicine use by patients with inflammatory bowel disease. Am J Gastroenterol. 1998;93:697–701. Nam RK, Fleshner N, Rakovitch E, et al. Prevalence and patterns of the use of complementary therapies among prostate cancer patients: an epidemiological analysis. J Urol. 1999;161:1521–1524. Rao JK, Mihaliak K, Kroenke K, et al. Use of complementary therapies for arthritis among patients of rheumatologists. Ann Intern Med. 1999;131:409–416. Saxe GA, Madlensky L, Kealey S, et al. Disclosure to physicians of CAM use by breast cancer patients: findings from the Women’s Healthy Eating and Living Study. Integr Cancer Ther. 2008;7:122–129. Yates J, Mustian K, Morrow G, et al. Prevalence of complementary and alternative medicine use in cancer patients during treatment. Support Care Cancer. 2005;13:806–811. Chao M, Wade C, Kronenberg F. Disclosure of complementary and alternative medicine to conventional medical providers: variation by race/ethnicity and type of cam. J Natl Med Assoc. 2008;100:1341–1349. Mehta DH, Gardiner PM, Phillips RS, et al. Herbal and dietary supplement disclosure to health care providers by individuals with chronic conditions. J Altern Complement Med. 2008;14:1263–1269. Mehta DH, Phillips RS, Davis RB, et al. Use of complementary and alternative therapies by Asian Americans: results from the National Health Interview Survey. J Gen Intern Med. 2007;22:762–767. Thomson P, Jones J, Evans JM, et al. Factors influencing the use of complementary and alternative medicine and whether patients inform their primary care physician. Complement Ther Med. 2012;20:45–53. Graham RE, Ahn AC, Davis RB, et al. Use of complementary and alternative medical therapies among racial and ethnic minority adults: results from the 2002 National Health Interview Survey. J Natl Med Assoc. 2005;97:535–545. Gardiner P, Graham RE, Legedza ATR, et al. Factors associated with dietary supplement use among prescription medication users. Arch Intern Med. 2006;166:1968–1974.
708 | www.lww-medicalcare.com
Medical Care
Volume 52, Number 8, August 2014
29. Sleath B, Callahan L, DeVellis RF, et al. Patients’ perceptions of primary care physicians’ participatory decision-making style and communication about complementary and alternative medicine for arthritis. J Altern Complement Med. 2005;11:449–453. 30. Chung VCH, Hillier S, Lau CH, et al. Referral to and attitude towards traditional Chinese medicine amongst western medical doctors in postcolonial Hong Kong. Soc Sci Med. 2011;72:247–255. 31. Kaptchuk TJ, Eisenberg DM. Varieties of healing. 2: a taxonomy of unconventional healing practices. Ann Intern Med. 2001;135:196–204. 32. Kaptchuk TJ, Eisenberg DM. Varieties of healing. 1: medical pluralism in the United States. Ann Intern Med. 2001;135:189–195. 33. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383. 34. Nguyen LT, Davis RB, Kaptchuk TJ, et al. Use of complementary and alternative medicine and self-rated health status: results from a national survey. J Gen Intern Med. 2011;26:399–404. 35. Nerurkar A, Yeh G, Davis RB, et al. When conventional medical providers recommend unconventional medicine: results of a national study. Arch Intern Med. 2011;171:862–864. 36. Snijders TAB. Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling. London; Thousand Oaks, CA: SAGE; 2012. 37. Raudenbush SW, Bryk AS, Congdon R. HLM 6 for Windows. Lincolnwood, IL: Scientific Software International Inc; 2004. 38. Davis MA, West AN, Weeks WB, et al. Health behaviors and utilization among users of complementary and alternative medicine for treatment versus health promotion. Health Serv Res. 2011;46:1402–1416. 39. Gryzlak BM, Wallace RB, Zimmerman MB, et al. National surveillance of herbal dietary supplement exposures: the Poison Control Center Experience. Pharmacoepidemiol Drug Saf. 2007;16:947–957. 40. Gelo F, Harris R. The secret garden. Fam Med. 2009;41:391–392. 41. Shih S-F, Lew-Ting C-Y, Chang H-Y, et al. Insurance covered and noncovered complementary and alternative medicine utilisation among adults in Taiwan. Soc Sci Med. 2008;67:1183–1189.
r
2014 Lippincott Williams & Wilkins
