Use of complementary and alternative medicine by pediatric rheumatology patients




To assess how often the patients in The Hospital for Sick Children's rheumatology clinic use complementary and alternative medicine (CAM), and the type of CAM used.


Cross-sectional descriptive survey study in a tertiary care referral center.


We found that 90 (64%) of 141 respondents used at least 1 form of CAM; 45 (50%) of these 90 used more than 1 form. Duration of illness was positively associated with the use of CAM; the difference in the mean length of illness between users (4.12 ± 3.60 years) and nonusers (2.82 ± 3.23 years) was 1.3 years (P = 0.04). Of the 41 respondents with more than 1 illness, 32 used CAM, compared with 57 of 96 who had only 1 illness (P = 0.036).


Health care providers must be aware of their patients' possible use of CAM, especially those with more than 1 illness.


Interest in complementary and alternative medicine (CAM) in Western countries has substantially increased in the past decade. Patients and their families seem to more frequently seek their health practitioners' opinions about a variety of CAM modalities. This may be partly the result of the increased accessibility of health information on the Internet. A recent study (1) found that 91 (44%) of 205 Web sites about rheumatoid arthritis promoted the use of CAM.

One of the most frequently cited reasons for using CAM is to aid in the management of chronic illnesses and pain (2–5). In fact, a Canadian study (3) revealed that 56% of CAM users had chronic problems of the musculoskeletal system. Furthermore, a secondary analysis of the 1994–1995 National Public Health Survey (4) indicated that Canadians' use of CAM was positively associated with the number of diagnosed chronic illnesses.

There is evidence that parents are giving CAM to their children, especially to those with chronic illnesses. For example, 55% of children with asthma are reported to use CAM (6). For children with cancer, reported use of CAM has increased from 9% in 1977 to 46% in 1994 (7).

With the chronic nature of rheumatic disease, it is not surprising that parents and the patients themselves seek various methods of pain control and relief. Two separate surveys of adult and geriatric rheumatology outpatients found that 66% use CAM (8, 9). The frequency of CAM use is better documented in adult than pediatric patients with rheumatic disease. The only 2 studies (10, 11), published more than a decade ago, on the use of CAM for children with rheumatic diseases found that between 66% and 70% of these children use at least 1 form of CAM. These studies, however, involved small sample sizes.

If the majority of pediatric rheumatology patients are using CAM, perhaps its use should be considered a routine part of the physician's assessment of the patient. We recognize that patients may be reluctant to disclose their use of CAM at clinic visits and pondered whether this reluctance may be based on concerns that disclosure may prejudice medical care.

The purpose of our study was to determine 1) the prevalence of current use of CAM by children with rheumatic disease seen in a tertiary care rheumatology clinic, 2) the types of CAM they currently use, and 3) any differences in the reported use of CAM between patients responding to a survey administered in the outpatient clinic and those responding to a mailed survey in their homes. We hypothesized that the majority of pediatric rheumatology patients have used at least 1 form of CAM within the preceding 12 months. We also hypothesized that there would be higher use of CAM reported by those responding to the mailed survey.


For the purpose of this study, we developed this definition of CAM, based on previous reports (7,9,12): CAM is any supplement, therapy, or remedy used by a patient that has not been prescribed by the patient's rheumatologist, other physician, or registered dietitian, excluding once-a-day multivitamins and meal-replacement products, and including the act of consulting with an alternative health care practitioner (e.g., naturopath, homeopath, or chiropractor). It is important to note that although we excluded multivitamins, we did indeed consider self prescribed single vitamins and/or minerals as CAM.


This was a cross-sectional, descriptive survey study. Our population consisted of pediatric rheumatology patients attending an outpatient rheumatology clinic in a major tertiary care referral center. We used 2 sampling methods. The first method involved recruiting clinic patients consecutively while they waited to see their rheumatologist. The second method used a random number list to select a sample of current rheumatology patients from our computer database to whom the survey was mailed. The purpose of the 2 sample groups was to obtain a broad distribution of participants and to compare the results between the 2 groups.


We developed a survey to answer our research questions. It was designed to be nonthreatening, nonjudgmental, and anonymous, and was deliberately presented as a health care survey to avoid any unfavorable connotations that may be associated with the terms complementary and alternative medicine. The selection of CAM treatment modalities in the questionnaire was based on what has been reported in the literature. Twenty products and services were listed (Table 1), and respondents were instructed to mark all the items they had used within the past year without a prescription from a conventional medical health professional.

Table 1. Types of CAM used*
Type of CAMNo. of users (%)
  • *

    Complementary and alternative medicine (CAM) as defined by our survey.

  • 29/77 (38%) CAM users used more than 1 product.

  • 16/38 (42%) CAM users used more than 1 service.

Product (n = 77) 
 Vitamins33 (37)
 Minerals31 (34)
 Relaxation techniques20 (22)
 Copper bracelets/rings16 (18)
 Herbal remedies15 (17)
 Fish oil10 (11)
 Meditation6 (7)
 Other dietary supplements6 (7)
Services (n = 38) 
 Chiropractor19 (21)
 Other CAM practitioner15 (17)
 Homeopath9 (10)
 Manual healer8 (9)
 Iridologist5 (6)
 Naturopath5 (6)
 Acupuncturist4 (4)
 Reflexologist2 (2)
 Aromatherapy1 (1)
 Biofeedback1 (1)
 Hypnosis0 (0)
 Osteopath0 (0)

The survey was edited by a plain-language specialist to ensure the language was appropriate for a lay audience, and approved by the Research Ethics Board for use at The Hospital for Sick Children in Toronto, Ontario, Canada.

We pretested the survey for face validity and content validity before administering it to the sample groups. We determined the survey's face validity by having it reviewed by a variety of health care professionals, including 3 registered dietitians, 3 pediatric rheumatologists, and 1 social worker. A convenience sample of 5 patients and their families were asked to pilot the survey for readability and ease of administration. To ensure adequate content validity, we followed 3 steps: 1) a review of the literature, 2) a review of other tools used (7, 9), and 3) a detailed review of survey questions by an experienced chiropractor.

A total of 200 surveys were distributed to 2 different groups. For the first group (Group A), a convenience sample of 90 patients attending the hospital's rheumatology clinic was approached by 1 of the research staff members and informed of the study. The patient or his or her caregiver was asked to read the cover letter and complete the survey while waiting in the clinic. We did not specify who should fill in the survey. For the second group (Group B), 110 copies of the same survey were mailed to a random sample of the hospital's current rheumatology patients. The random sample was generated (within the same time frame, i.e., May 1999–September 1999, as that for the Group A collection) from the computer database of the rheumatology division. We modified Dillman's total design method (13) to ensure the highest response rate possible. Initially, the survey was mailed with a cover letter and a self-addressed stamped envelope for the return of the completed survey. As an incentive, we also included 3 extra postage stamps for the respondents' personal use. We then sent a reminder postcard 2 weeks later to all potential respondents.

To avoid duplication of the survey responders in the clinic, we verified that the potential respondent was not on the list of recipients of the mailed survey.


To determine the sample size, based on previous studies (9–11) for this group, we estimated that the proportion of CAM users in our population would be approximately 65%. To establish a precise estimate, we calculated that 180 subjects would be needed to provide a 95% confidence interval (95% CI) of 7%. Therefore, we aimed to recruit 90 subjects from each Group. We sent 110 surveys to Group B, assuming an 80% response rate from the random mailing.

Data analysis.

We primarily used descriptive statistics to analyze the data, but we also tested for significant differences between the 2 responder groups. We examined bivariate associations between responder variables (age, diagnosis, length of illness, sex) and outcomes of interest. Categorical variables were compared with the chi-square test and continuous variables with the Student's t-test. Differences were defined as statistically significant at 5% for all computations. The diagnoses were categorized into 3 groups: 1) the arthritis group (all arthritis subtypes); 2) the connective tissue disease group (systemic lupus erythematosus, dermatomyositis, scleroderma, vasculitis); and 3) the other group, including fibromyalgia, rare diagnoses such as sarcoidosis, and those illnesses not yet diagnosed.


All 90 outpatients approached in the clinic (Group A) completed the survey (response rate 100%). Of the 110 surveys mailed to Group B, 51 were returned (response rate 46.4%). In total, we analyzed 141 completed surveys, 78.3% of our original goal. The surveys were completed by parents (84%), patients (11%), and others (i.e., type of respondent not identified; 5%). There were no statistical differences in the demographic characteristics of users and nonusers of CAM (Table 2).

Table 2. Demographic characteristics of respondents
All (n = 141)CAM users (n = 90)
  • *

    Sex not indicated: 5 for all respondents and 2 for complementary and alternative medicine (CAM) users.

  • Mean duration of illness: 3.47 years for all respondents, 4.12 years for CAM users. More than 1 illness: 103 (73%) for all respondents; 71 (79%) for CAM users.

Mean age, years10.410.9
Sex, no. (%)*  
 Male41 (29)30 (33)
 Female95 (67)58 (64)
Diagnosis group, no. (%)  
 Arthritis69 (49)45 (50)
 Connective tissue disease25 (18)18 (20)
 Other41 (29)24 (27)
 Not indicated6 (4)3 (3)

We found that 90 of the 141 patients we surveyed (64%; 95% CI 55.9–71.7) currently use at least 1 form of CAM, and 45 (50%) of these 90 CAM users use more than 1 type of CAM. Of those children using products, 38% used more than 1 type; of those using services, 42% used more than 1 type (Table 1). The most common types of CAM used were vitamins and minerals. Respondents indicated that they used these vitamins: vitamin C (55%, 18/33), vitamin D (30%, 10/33), other (30%, 10/33), vitamin E (15%, 5/33), and vitamin A, Bs, and others not indicated (3%, 1/33); 10 respondents used more than 1 vitamin. Respondents indicated that they used these specific minerals: calcium (52%, 16/31), iron (35% 11/31), zinc and magnesium (10%, 3/31), and others (13%, 4/31); 4 respondents used more than 1 mineral. Other commonly used CAM included relaxation techniques, copper bracelets or rings, and herbal remedies. The types of CAM used less frequently (less than 10% of the time) were fish oil, other supplements, aromatherapy, reflexology, iridology, and naturopathy. No respondents reported using hypnosis or osteopathy.

We did not find significant differences between users and nonusers of CAM in age, sex, or diagnostic group. However, CAM users had a longer duration of illness (4.12 ± 3.60 years) than nonusers (2.82 ± 3.23 years); mean difference 1.3 years (P = 0.04).

Of the 137 patients who reported their number of illnesses, 41 had been diagnosed with more than 1 illness. These 41 children were more likely to be CAM users (32/41) than those with just 1 illness (57/96; P = 0.036).

Group A (clinic respondents) and Group B (mail respondents) differed significantly in their use of minerals (14/90 versus 17/51; P = 0.014) and in their use of chiropractic services (16/90 versus 3/51; P = 0.047).

Thirty-six (26%) of all 141 respondents reported using books, magazines, or other literature to find information on CAM; 45% (64/141) did not respond to this question. Forty-two (30%) of the 141 respondents used the Internet to find information about CAM; 26% (37/141) did not respond to this question.


We found that almost two-thirds of our study population had used at least 1 form of CAM within the 12 months preceding the survey; half of these users partook of more than 1 type. Our findings are similar to those of Southwood and colleagues (11) who found that 70% of the patients with juvenile arthritis surveyed were users of CAM, although not necessarily within the previous 12 months. Moreover, both a 1994 (9) and 2000 (8) survey of adult and geriatric rheumatology patients found that 66% use CAM. In their 1990 study, Southwood and colleagues (11) found that the most popular form of CAM was some form of copper band or bracelet, whereas in our study, vitamins and minerals were the most popular form of CAM. It is possible that some practitioners may not include vitamins and minerals in their definition of CAM, hence our results could be interpreted differently. For example, if we excluded those respondents who used only vitamins and/or minerals, the total number of CAM users would decrease to 73/141 (51.8%). Similarly, it could be debated that calcium is not a CAM; if we excluded respondents who used only calcium, the total number of CAM users would decrease to 78/141 (55.3%).

Our study suggests that the type of CAM used may have changed over the past decade, but the proportion of CAM users has remained remarkably consistent. Differences in the wording of the surveys themselves and in the use of CAM in different countries may also account for some of the differences between our studies: the survey of Southwood and colleagues (11) involved patients from Australia, New Zealand, and Canada.

We found that the most common types of CAM used were vitamins (37%) and minerals (34%), followed by relaxation techniques (22%). Because we excluded multivitamins from the definition of CAM, it would seem that a substantial use of vitamins and minerals is an interesting and novel finding. The only other study (9) that excluded multivitamins, that we are aware of, described the use of CAM in an adult rheumatology patient population. Clinicians must therefore be aware of the possibility of interactions between vitamins or minerals, or both, and treatment medications.

Neither of the previous studies of pediatric rheumatology patients and CAM use that we noted considered the number of diagnoses for each child as a variable (10, 11). We found that children with more than 1 illness were more likely to be users of CAM. This finding concurs with that of a recent publication (8) that assessed the use of CAM in geriatric rheumatology patients.

Contrary to what we expected, we found no appreciable difference in the proportion of CAM users between the 2 sample groups. We had postulated that those completing surveys in their homes might be more comfortable disclosing their use of CAM (owing to the anonymity of the process), and hence result in an increased number of reported users in that group. However, the only significant difference between the survey groups was in their use of use of minerals. This may reflect a reporting bias from those in Group B who returned surveys; perhaps only those patients who felt comfortable or even enthusiastic about disclosing their use of minerals returned the survey.

Despite the widespread use of CAM, its benefits, adverse effects, and potential interactions are not well known. Some modalities of CAM have the potential to adversely affect the patient's disease course and outcome. For example, the use of cat's claw has been implicated in the development of renal failure in a patient with systemic lupus erythematosus (14). It is possible that users of CAM may abandon a child's conventional treatment.

This study does have limitations. Inherent in self-administered questionnaires is the potential for misinterpretation of the questions themselves. However, our questionnaire was modified by a plain-language specialist and pretested with patients and families during its development. Given our sampling methods, we may have studied a group of children who were less likely to be CAM users. Certainly, our sample of participants excluded children who did not receive treatment from our tertiary care institution and children refusing conventional treatment. However, previous studies (6, 8, 12, 15) show that the majority of CAM users also receive traditional medical therapy and would likely be properly represented in our sample.

Our study is the only report of CAM use by pediatric patients with a variety of rheumatic disorders in a tertiary care setting. The findings of our and previous studies, including reports on adults and children with chronic disease, suggest that there is a high prevalence of CAM use across all age groups and in many countries.

In conclusion, we have demonstrated that there is a high prevalence of CAM use in pediatric rheumatology patients attending a tertiary care ambulatory clinic. We recommend that questions pertaining to the use of CAM should be a part of every medical visit (16). We highlight the importance of considering, if not assuming, that patients with chronic disease, and especially those with more than 1 illness, may be using CAM.


The authors wish to thank those colleagues who provided valuable input for this study and manuscript. A special thanks to K. Everett and everyone else who assisted with data collection. This article was prepared with the assistance of Editorial Services, The Hospital for Sick Children, Toronto, Ontario, Canada.