This study sought to develop and validate the ABCAM instrument to measure the decision factors related to the use of CAM among cancer patients. The conceptual model of ABCAM was guided by TPB. It was developed through the literature review, qualitative research, expert review, pilot testing, and quantitative psychometric analysis. The final instrument consists of 15 items measuring three domains related to the attitudes and beliefs predictive of CAM use: expected benefits, perceived barriers, and subjective norms. The scores appear to be reliable and valid in our study population. As hypothesized, CAM users reported higher expected benefits, lower perceived barriers, and more positive subjective norms associated with CAM than those who did not use CAM.
In comparison to existing questionnaires [13
], the ABCAM is the only one we know that has gone through the process from development to validation in cancer patients. The theoretical model and content of our scale had similarities to the scale developed by Hirai et al., however, the perceived negative outcomes of CAM as measured by Hirai et al. did not include the barriers related to CAM use, which our instrument improves upon. Additionally, all three domains of the ABCAM instrument, including perceived benefits, perceived barriers, and subjective norms, demonstrated higher internal consistency than those reported by Hirai et al. [34
Our study showed that higher scores of perceived benefits were associated with CAM use among cancer patients. Previous research has shown that cancer patients often use CAM because perceiving it will improve their physical and emotional health, enhance their quality of life, strengthen their immune system, reduce symptoms, and have a positive effect on cancer [10
]. Perceived positive outcomes of CAM use were associated with higher CAM use among a sample of Japanese cancer patients in a prior study [34
]. It is important to note that while immune enhancement was a response endorsed by participants, this item cross-loaded to social norm which did not get retained in our final shortened instrument because it did not contribute to the unique factor structure of the instrument. This further suggests the belief that CAM improving one's immune system appears to be socially constructed.
The literature suggests that some of the barriers toward the use of CAM include lack of knowledge, perceived ineffectiveness, cost, time constraint, access to the provider, and perceived side effects of CAM therapies [10
]. As expected, our study showed that cancer patients who used CAM demonstrated lower perceived barriers as compared to non-CAM users. The domain of perceived barriers represents the construct of perceived behavioral control in the TPB. It is important to note that some of barriers listed are experienced by individuals but they are probably structural barriers (e.g., cost, and access) as well. Therefore, these barriers may be beyond the control of many individuals and will require policy change, insurance coverage, and design of an integrative health care delivery system to ultimately influence change.
Prior studies found that CAM users were more likely to be of female sex, younger age, higher socioeconomic status (e.g., education, and income), and white race [1
]. Our barrier domain may help understand what specific barriers are experienced among different sociodemographic groups. As evidence accumulates regarding the potential efficacy of some of the CAM therapies in cancer symptom management, this understanding may help reduce the potential disparity in CAM integration. Using our instrument may help quantify the level and significance of these barriers and to guide interventions to target them.
Subjective norms play an important role in patients' intended and actual health behaviors. Patients are more likely to use CAM if it is recommended by their family/friends and/or their health care providers [34
]. Our study revealed that CAM users had more positive subjective norms than non-CAM users. This suggests that social approval or disapproval may play an important role in influencing patients' use of CAM therapies; however, our items of family/friend influence cross-loaded between expected benefits and social norm and thus were removed from the final instrument. Consistent with prior qualitative research [25
], our data further strengthens the evidence that family/friends' opinions help shape an individual's expected benefit of CAM use; thus, its social normative effect cannot be separated from patients' expected benefits derived from the therapy. Another possible explanation is that cancer patients often consider the opinion of their treating specialist as most important and follow their advice [56
]. As our instrument is investigated in future research, we can tease out how sources of social influence may shape expectations of therapeutic benefits as well as decisions to use a particular therapy.
The limitations to this study need to be acknowledged. First, our qualitative interviews were conducted with breast cancer patients in the context of decision making about acupuncture; the content of the instrument may not be complete. However, our questionnaire items were also supplemented from the existing literature and then discussed among content experts and patients with other cancers during cognitive interviews. Second, our instrument was guided by TPB as a conceptual framework and well captured the domains in TPB, but like any conceptual model, it may not fully capture other important constructs such as preferences for natural therapies, holistic health view, and finding hope [39
]. Additionally, we created a brief instrument that can be incorporated into future cancer epidemiology and health service research; thus, the format of ABCAM is not a traditional TPB instrument. Third, our CAM use was based on self-report and may not capture all of the CAM therapies used by individuals; however, 60.9% use is in the range of what is reported in existing literature [3
]. Forth, nonparticipation bias is always a concern in an epidemiology study. Our 83% participation rate is acceptable in survey research, but cannot rule out the potential for selection bias. Lastly, our study was conducted in a large academic cancer center, and future research, including community cancer practices, is needed to increase the generalizability of this study.
In conclusion, this study provided the initial evidence that the ABCAM produced a reliable and valid score for measuring the behavioral predictors of CAM use. Future research is needed to demonstrate additional aspects of reliability and validity (e.g., confirmatory factor analysis; test-retest reliability; sensitivity to change). In addition, prospective research is needed to determine whether these attitudes and beliefs—expected benefits, perceived barriers, and subjective norms—predict both intended and actual use of CAM among cancer patients. Ultimately, this instrument will help elucidate how demographic, socioeconomic, and cultural issues may relate to these attitudes and beliefs, thereby influencing CAM use in the context of cancer care. Such understanding is necessary to guide the appropriate integration of CAM into the conventional health system to improve the health and wellbeing of diverse populations of cancer patients.