As previously reported, four out of 10 United States adults used CAM therapies in the prior 12 months3
. We found, similar to previous reports61,63
, that CAM users in general reported more health problems in the prior year as evidenced by an increased number of clinical conditions included in the Charlson Comorbidity Index and Kessler Score. CAM users also reported higher numbers of visits to health care offices and emergency rooms and days spent in bed in the prior year.
In contrast to other findings that associated CAM use with worse health64,65
, we found CAM use was associated with better current health status as well as improved health over the prior year. Our findings present an interesting paradox in that the respondents using CAM were more likely to have chronic illness, as evidenced by the high CCI and K6 scores, yet also were more likely to report that their health status was excellent and better than the prior year. One interpretation of this finding is that the current ‘excellent’ health status reflects what the respondents felt at the moment of being interviewed for the survey while their answers to the questions on chronic conditions reported what the respondents had experienced in the prior 12 months. Since the timeframes for these questions differed, the responses could be consistent with one another. Alternatively, the respondents’ perceptions of health may be affected by patients’ expectations after their investment in CAM or a sense of empowerment or optimism related to the CAM use of interest4,66
Our analytic approach used in this study is novel. In reviewing the literature on the relationship between health status and CAM use, most previous studies have modeled CAM use as a dependent (or response) variable and have included health status or change in health status as independent variables. Since it is reasonable to assume the CAM use and clinical conditions reported by the respondents took place before the respondents reported ‘current health status’ and ‘health improvement’, we took the opposite approach where health status and change in health status were our dependent variables and CAM use was included among our independent covariates. In this way, we were specifically adjusting for the statistical effects of other factors on the likelihood of the health status and its improvement over time, as well as identifying potential confounders of the relationship between CAM use and health status and health improvement.
While our decision to categorize health status as excellent compared to other responses may have affected our results, our sensitivity analysis confirmed our finding that CAM use is associated with health status and change in health over the prior year.
Due to the observational nature of the database we analyzed, our finding does not determine causation and it is worth noting that, in general, CAM effectiveness research, at best, has been contradictory. Most studies for botanicals have been negative67
and while there are positive reports of small benefits for acupuncture68,69
, most studies show no difference between acupuncture and sham acupuncture21,70
. Studies of mind–body therapies seem to be more positive and suggest benefits for reducing blood pressure71
, preventing falls72
, low back pain73
, and irritable bowel syndrome74
. Alternatively, hypotheses on the relationship of CAM to health benefits could be explained by anthropological research which shows that participation in healing rituals can confer subjective perceptions of benefit irrespective of any changes in pathophysiology or symptomatology66
Our study has several limitations. Many of them are inherent in survey research75
. Questions are subject to varied interpretations by respondents of different cultures and social and educational backgrounds; thus subjective answers, such as health status being good, fair, or poor, may be reported inconsistently by subjects of different backgrounds and may be affected by expectation and other factors. Recall bias and a limited set of CAM therapies affect prevalence estimates. For example, modalities such as deep breathing exercises may not be generally viewed as CAM therapy; failing to include this therapy would likely lead to a biased estimate of CAM prevalence. The self-reported symptoms, conditions, and health status may not meet standard clinical definitions. The absence of data on quantity, duration, and timing of CAM use limits our ability to distinguish the characteristics of one-time users from more frequent ones and to ascertain any dose response treatment effects on health status. Finally, since the survey was administered only in English and Spanish, it may have under-represented certain immigrant populations.
Although CAM use is increasingly popular, response to CAM is complex and not readily understood; research on its effectiveness is still in developmental stages. Methodological constraints, such as small sample size, inadequate controls, and poor specificity of eligibility criteria and interventions, have plagued the field and hampered the interpretation and generalizations of results76–78
. Our findings, however, suggest that, on a population basis, CAM use may have implications for better health status and health improvement over time. Clearly, large-scale randomized controlled studies are required to establish a causal relationship between CAM treatments and their effects on health status. Our results suggest that such studies are needed.