Berman BM, Bausell RB, Lee W. Use and Referral Patterns for 22 Complementary and Alternative Medical Therapies by Members of the American College of RheumatologyResults of a National Survey. Arch Intern Med. 2002;162(7):766-770. doi:10.1001/archinte.162.7.766
This study was designed to determine rheumatologists' self-reported knowledge, perceptions of legitimacy, referral patterns, and use in practice of 22 complementary and alternative medicine (CAM) therapies.
A survey was mailed to a random sample of 2000 physician members of the American College of Rheumatology asking respondents which (if any) CAM therapies they (1) knew enough about to discuss with patients, (2) considered part of "legitimate medical practice," and (3) "personally administered" to patients, or "referred patients to someone else" to administer. The response rate was 47%.
On average, the respondents reported knowing enough to discuss 10 of the therapies with patients, considered 9 to be part of legitimate medical practice, and had referred patients to someone else for 8 of the 22 therapies. Correlates of use and/or referral included sex, age, belief in the legitimacy of the therapies, and self-reported knowledge.
These results provide potentially important preliminary data regarding rheumatologists' responses to dramatic increases in the use of CAM therapies among their patients.
IT HAS BEEN ESTIMATED that there are now more office visits made for complementary and alternative medicine (CAM) therapies in the United States than for the services of conventional primary care physicians.1 While a number of studies that have also used nationally representative samples2- 4 have found considerably lower estimates of CAM use, there is little doubt that patients' use of therapies other than those prescribed by their convention physicians is a clinical issue that must be reckoned with. This is especially true for rheumatologists and those primary care physicians who treat rheumatological patients, since an ever-growing number of consumer-based surveys have indicated that visits to alternative therapists are especially prevalent among patients with chronic pain–related conditions, such as arthritis5- 14 and fibromyalgia.15- 18 In general, while some of these surveys (1) were relatively small, (2) did not use probability sampling methods, (3) used different time intervals on which to base prevalence estimates, and (4) used varying definitions of what constitutes a CAM therapy, it appears that estimates of the use of CAM therapies among rheumatological patients are generally higher than among patients with a number of other specific conditions. Although this literature is difficult to summarize, the cumulative weight of the evidence appears to indicate that the prevalence of the use of CAM therapies among patients with arthritis and fibromylagia is extremely high and certainly encompasses the majority of rheumatological patients.
There have been many fewer surveys of rheumatologists with respect to their clinical response to this movement, although there have been numerous physician surveys targeting attitudes toward, and use of, CAM therapies among primary care physicians in general. Two meta-analyses of the survey literature regarding CAM use19,20 and a number of individual physician surveys21- 44 have shown that there is considerable professional interest in complementary therapies as well as a general willingness to refer patients to providers of at least some of the CAM modalities.
Of the 2 published surveys targeting rheumatologists of which we are aware, one was published in a popular health magazine45 in 1999 and was difficult to evaluate because its methodology was not specified, while the other involved an exclusively Dutch sample of rheumatologists.8 Both articles identified a number of CAM therapies that the majority of the respondents viewed positively (capsaicin, relaxation, biofeedback, meditation, and journal writing in the Arthritis Today sample; spa treatment, acupuncture, and manipulation in the Dutch sample), but neither survey was designed to systematically assess the clinical responses to this growing phenomenon.
Therefore, the primary purpose of the present study was to assess the extent to which present-day rheumatologists incorporate CAM therapies into their professional practices, either through direct patient care or through referral to other providers. A secondary purpose was to assess the extent to which these behaviors could be explained by the demographic, professional, and affective characteristics of the clinicians who participated in the survey.
A survey instrument soliciting self-reported knowledge of, attitudes toward, clinical use of, and referral to providers of 22 separate CAM therapies was mailed to a random sample of 2000 of the 4879 physician membership of the American College of Rheumatology residing in the United States. After 2 additional mailings to nonrespondents, 924 usable questionnaires were received (28 were not delivered), representing an effective response rate of 47%.
As depicted in Table 1, the majority of the responding physicians were male (87%) and older than 50 years (78%). In general, the respondents were heavily engaged in direct clinical practice (75% saw patients at least 24 hours per week) and were qualified, with 94% reporting possessing board certifications in both rheumatology and internal medicine.
The questionnaire that was used in the present study was based on previous survey research conducted by the University of Maryland Complementary Medicine Program.43,44 The instrument solicited information regarding basic demographic and practice variables as well as self-reported attitudes and clinical behaviors regarding the above-mentioned list of 22 CAM therapies. The items related to the 22 CAM therapies were contained in an alphabetical list of the behaviors, with instructions for the respondents to indicate, for each separate therapy, whether or not they (1) knew enough about the behavior to discuss it with patients; (2) considered it to be a part of legitimate medical practice that was designed as a global attitudinal indicator (and hence possibly related to respondents' perceptions of efficacy); (3) had personally administered the therapy to patients; and (4) had referred patients to someone else to administer the therapy. Each of the resulting 88 responses was scored dichotomously (yes or no), with total composite scores (theoretically ranging from 0-22) computed for the 4 affective/behavioral dimensions, ie, self-reported knowledge, opinions regarding legitimacy, clinical use, and referral) by summing across the 22 therapies.
Descriptive statistics were used to present physicians' knowledge, attitudes, and clinical behavior relevant to the 22 individual therapies, while multiple linear regression was used to explore the correlates of clinical utilization of these CAM therapies considered as a whole.
As indicated in Table 2, the respondents were much more likely to refer patients to other practitioners for the CAM therapies than to administer them themselves. Trigger point therapy and nutraceuticals were the 2 therapeutic exceptions, with 51% and 34% of the respondents, respectively, reporting direct clinical involvement with these therapies. The other 4 therapies for which the most active direct clinical use was reported were (1) exercise intervention (41%), (2) dietary prescription (33%), (3) counseling/psychotherapy (24%), and (4) electromagnetic applications such as transcutaneous or pericutaneous electrical nerve stimulation (10%). Totaled across all 22 therapies, the average physician reported having clinically administered approximately 2.5% of the behaviors, although fewer than 5% of the sample reported having had direct clinical experience with 11 of the behaviors (ie, acupuncture, biofeedback, chiropractic, energetic healing, homeopathy, hypnotherapy, magnets, nonchiropractic manipulation, message, specialized movement therapies such as qi gong and yoga, and music therapy).
At least 50% of the responding physicians had referred patients to 8 of the therapies (ie, acupuncture, behavioral medicine, biofeedback, counseling/psychotherapy, dietary prescriptions, electromagnetic applications such as transcutaneous and percutaneous electrical nerve stimulation, exercise, and massage). When direct clinical use of the therapies was combined with referral (ie, when an individual physician either administered the therapy or referred the patient to someone else to administer it), 9 of the 22 modalities had been used by more than 50% of the respondents, with counseling/psychotherapy (85%) and exercise (81%) heading the list.
While there were 9 therapies for which more than 50% of the clinicians had reported at least some degree of clinical use (ie, either through personal administration or referral), it is important to note that there were considerably more of these modalities that were not used in any substantive form by this group of physicians. There were also 9 therapies, in fact, for which 75% of the sample had reported no clinical use (in descending order these were meditation [24%], prayer and spiritual direction [23%], nonchiropractic manipulation [23%], hypnotherapy [19%], herbal medicine [14%], music therapy [9%], magnets [5%], energetic healing [5%], and homeopathy [4%]). Interestingly, none of these modalities were considered part of legitimate medical practice by a substantial proportion of the sample, nor did as many as 50% of the respondents report sufficient knowledge to discuss any of them with their patients.
In summary, there appeared to be a major dichotomy among the 22 listed therapies, with one group having achieved a certain degree of acceptance and the other group definitely not having achieved anything approaching mainstream acceptance. In general, physicians' knowledge of a therapy (as defined by reporting to possess enough knowledge to discuss it with patients) and whether or not they considered it to be a part of legitimate medical practice were relatively closely related to one another, as they were to clinical use and referral patterns.
A multiple regression was then performed to ascertain the extent to which clinical use of CAM therapies among members of the American Board of Rheumatology was related to demographic, professional, and attitudinal characteristics of the sample. The dependent variable in this analysis was the total number of therapies for which the respondent reported having either personally administered to patients or referred patients to other clinicians for administration. The independent variable set included (1) the demographic and professional characteristics of age, sex, and number of hours per week spent in clinical practice; (2) the total number of CAM therapies that the respondents thought were a part of legitimate medicine; and (3) the total number of these therapies that the physicians considered themselves to be sufficiently knowledgeable about to discuss with their patients.
Overall, 41% of the variation in the overall self-reported clinical use of the 22 CAM therapies was shared with the set of 5 predictors (Table 3). Only the number of board certifications was not an individually significant correlate of clinical use. The strongest predictor was, not surprisingly, beliefs in the legitimacy of the CAM therapies. Even after this variable was statistically controlled for, however, knowledge and hours spent in clinical practice were positively related to use. Female physicians were also marginally more likely to engage in the clinical use of CAM therapies (P = .052), while there was a negative relationship (P = .008) between physician age and referral or administration of the CAM therapies. Interestingly, however, a secondary analysis (not shown) indicated that this latter relationship was mitigated by the fact that physicians 55 years of age or older were actually significantly more likely to personally administer certain therapies (most notably counseling and behavioral medicine techniques), while being less likely than their younger colleagues to refer patients to other practitioners.
This survey suffered from a number of limitations. The response rate, while respectable for a physician survey, could have been higher. Its interpretation also depends upon self-reported data, and there is always a question regarding exactly what does and does not constitute a CAM therapy. Many types of psychologically based therapies (a number of which, such as counseling, psychotherapy, relaxation techniques, and behavioral medicine, are difficult to separate into distinct, mutually exclusive modalities), for example, have almost certainly crossed the boundary into conventional medical practice, while exercise therapy is actually a component of the American College of Rheumatology's clinical guidelines for osteoarthritis of the knee and hip.46 Other therapies, such as biofeedback and transcutaneous electrical nerve stimulation, may well be in the process of making this transition from complementary to conventional medicine. Finally, the responses generated by this survey cannot be assumed to be representative of rheumatological practice, since the present sample of clinicians are obviously better trained and probably more knowledgeable than clinicians in general.
With these caveats in mind, however, the present results do appear to reflect an openness among rheumatologists toward a number of CAM treatment modalities that they consider to be a part of legitimate medical practice and, to a lesser extent, about which they report possessing enough knowledge. These results, along with the weaker age and sex relationships, are generally consonant with those reported previously (our surveys), although the latter variables are not always correlated with physicians' use of CAM therapies.19 Finally, although the clinical administration of CAM therapies by conventional physicians has not been well studied, some researchers have found a relationship between a tendency to accept CAM or to refer patients to CAM providers and experience/knowledge of the therapies involved,21,28,30,36 while others have not.24,26
These results are also remarkably consistent with those of a smaller survey (Table 4) conducted recently using a random sample of the membership of the International Association for the Study of Pain,44 especially given the fact that the 2 sets of respondents were probably treating different types of patients for different conditions. This survey also comprised a large proportion of board-certified physicians (86%) and achieved a similar response rate (53%).
Among the 14 CAM therapies that were common to the 2 surveys, the rank ordering of the use (as defined by referral or personal administration) for these therapies was quite similar (ρ = 0.88, P<.001), and the 2 samples had, on average, administered or referred patients to approximately half of these nonpharmacological (nutraceuticals were not included in the pain specialists' survey) therapies (6.6 of the 14 among the rheumatologists; 7.4 among the pain specialists).
What neither survey addresses, of course, is whether patients actually benefit from exposure to these therapies. More investigations targeted at rigorously assessing their efficacy is therefore urgently needed, both by practitioners who must make referral decisions and by patients who need to know what viable treatment options are available to them.
In the meantime, while this evidence is accumulating, opinions understandably differ regarding the appropriateness of using some of these individual therapies. The present results can thus be interpreted as reflecting a very real and appropriate commitment among this highly qualified group of clinicians toward the treatment of a variety of patient conditions for which completely viable pharmacological options do not yet exist. It will be interesting to see whether the use of CAM therapies among mainstream rheumatologists will increase over time, the way that public acceptance and use of alternative therapists have.
Accepted for publication August 7, 2001.
This study was supported in part by grant 1P50AT0008401 from the National Center for Complementary and Alternative Medicine, National Institutes of Health, Bethesda, Md.
Corresponding author and reprints: R. Barker Bausell, PhD, Complementary Medicine Program, University of Maryland School of Medicine, Kernan Hospital Mansion, 2200 Kernan Dr, Baltimore, MD 21207 (e-mail: firstname.lastname@example.org).