We estimate that 17.4 percent of all U.S. adults used CAM to treat an illness and 27.4 percent used CAM for health promotion in 2007. Using our predefined mutually exclusive User categories, 10 million adults used CAM to treat an illness (4.4 percent of U.S. adults) only, 32 million adults (14.3 percent of U.S. adults) used CAM for health promotion only, and 29 million adults (13.0 percent of U.S. adults) used CAM for both purposes (). The NCCAM Manipulative and Body-Based Therapies domain was the most common form of CAM used as treatment for an illness (used by 8.9 percent of adult NHIS respondents) and the Mind–Body domain was the most common form of CAM used for health promotion (used by 14.8 percent of adult NHIS respondents) (). Among specific CAM modalities, Chiropractic or Osteopathic Manipulation was the most common modality used as treatment (used by 6.8 percent of the adult population for treatment) and the most common modalities used for health promotion were relaxation techniques and herbals and nonvitamin supplements (both of which were used by over 11 percent of the adult population for health promotion).
The Sociodemographic Characteristics of Complementary and Alternative Medicine Users versus Non-Users
By U.S. region, Health Promotion Users were overrepresented in the West (accounting for 25 percent of Health Promotion Users compared with 20 percent of Non-Users) and were relatively underrepresented in the South. Treatment Users were more heavily represented in the Midwest than Health Promotion Users (accounting for 30 percent of Treatment Users and 24 percent of Health Promotion Users, p = .01).
All CAM Users were more likely than Non-Users to be female (); women accounted for 48 percent of Non-Users but 52 percent of Treatment Users, 57 percent of Health Promotion Users, and 61 percent of Mixed Users. Across all CAM User categories, CAM use was also associated with higher educational level and higher likelihood of being Non-Hispanic white when compared with Non-Users.
Health Promotion Users were significantly more likely to be white, married, and educated than Treatment Users (e.g., 15 percent of Health Promotion Users had completed a graduate degree compared with only 8.8 percent of Treatment Users, p<.001).
Health Status and Health Behaviors
Health Promotion Users were significantly healthier than all other user types (72 percent of Health Promotion Users reported “excellent or very good health” compared with only 54 percent of Treatment Users, p<.001) and had lower obesity and functional limitation rates (). On the other hand, Treatment Users had significantly higher rates of functional limitation (47 percent of Treatment Users reported a functional limitation compared with 27 percent among Health Promotion Users, p<.001) and the functional limitation rate among Health Promotion Users approximated that of Non-Users.
The Health Status and Health Behaviors of Complementary and Alternative Medicine Users versus Non-Users
CAM Users in general were more active than Non-Users; CAM use was associated with greater levels of physical activity in all three categories examined. Among CAM User types, Health Promotion Users had the highest reported physical activity rates.
Health Services Utilization
All CAM Users were more likely to have visited a general practitioner, a medical specialist, and a mental health practitioner in the previous 12 months than Non-Users (). Treatment Users had significantly higher rates of use of all conventional health utilization measures when compared with Health Promotion Users (p<.001 for all measures), whose use of health services approximated that of Non-Users.
Utilization of Conventional Health Services among Complementary and Alternative Medicine Users versus Non-Users
Logistic Regression Analyses
Each of the dichotomized health status and health behavior measures in was submitted to logistic regression to assess the effect of the type of CAM use (Treatment Users versus Health Promotion Users) while controlling for the regional and sociodemographic characteristics in . The type of CAM use was strongly predictive, p<.001, of health status, obesity, and functional limitations. On the other hand, the type of CAM use was not significantly predictive (p>.05) of smoking status, alcohol status, having been tested for HIV, or for having received the flu vaccination.
For the dichotomized physical activity measures in and the health service utilization measures in submitted to logistic regression to assess the effect of the type of CAM use while controlling for the regional and sociodemographic characteristics, health status, and health behaviors, the type of CAM use was predictive, p<.01, of greater physical activity on all physical activity measures and of greater use of all health services utilization measures.