Exploring the use of complementary and alternative medicine (CAM) among people with cancer: A mixed-methods approach to instrument development

Abstract

Purpose: Complementary and alternative medicine (CAM) use by individuals with cancer is widespread. By definition, CAM are not an integral part of conventional medicine, and until scientific research has established the safety and efficacy of some individual CAM, an exploration into factors influencing the use of CAM among individuals with cancer is warranted. This thesis presents a two-phased exploratory sequential mixed-methods study which aimed to develop and test an instrument to explore factors influencing the use of CAM among individuals with cancer. Two central research questions guided the study: (a) what factors influence individuals with cancer to use, or not use CAM; and (b) how valid and reliable is the newly developed instrument in measuring factors that influence CAM use by individuals with cancer?

Method: In-depth interviews were initially conducted to explore the phenomenon of CAM usage with: individuals with cancer; oncology health professionals; and CAM practitioners in Phase I. The PRECEDE model was used to guide and categorize the qualitative findings. Categories developed from the qualitative findings were then used to generate items and inform the development of a new instrument to measure factors influencing CAM use in the cancer population. During the instrument development process, guidelines from the European Organization for Research and Treatment of Cancer (EORTC) were utilized. Through this process, a new instrument entitled Factors Influencing the Use of Complementary and Alternative Medicine (FICAM) Scale was developed and pilot-tested. The FICAM Scale was then administered to a sample of adult individuals with cancer to test its psychometric properties in Phase II. Data were analyzed in three ways: exploratory factor analysis to identify factors that influence the use of CAM, analysis of the scale reliability, and correlational hypothesis testing to establish construct validity.

Results:In Phase I, a total of 35 individual in-depth interviews were conducted with 3 key-groups of participants: (i) individuals with cancer who were both CAM users and non-CAM users (n = 17), (ii) oncology health professionals (n = 9), and (iii) CAM practitioners (n = 9). Participants indicated every individual’s experience with cancer was unique and therefore the spectrum of factors influencing their use of CAM was diverse across the illness trajectory. Building upon the PRECEDE Model, 12 categories of factors influencing CAM use were inductively developed from the content analysis of the in-depth interviews. Qualitative data findings were then used to generate a pool of 157 items for use in development of the new instrument. Items were reviewed by a group of 19 panel members and refined following pilot-studies with a sample of 19 individuals with cancer.

The outcome was a 59-item FICAM Scale. Subsequently, the FICAM Scale, along with four established instruments, was administered to a consecutive sample of 125 individuals with cancer (response rate – 82.5%) in Phase II. Exploratory factor analysis yielded a three-factor solution of 45-items – belief in CAM, belief in conventional medicine, and barriers to CAM - explaining 45.93% of the variability of factors influencing CAM use. Cronbach alphas for the FICAM subscales ranged from .744 to .951. The FICAM Scale also demonstrated construct validity through inter-item correlations and corrected item-total correlations. Test-retest reliability revealed that the FICAM Scale was stable, with the Intra-Class Correlation (ICC) coefficient for its subscales ranging from .822 to .967. Participants who indicated a positive belief in CAM were more likely to use CAM when compared with those who indicated a positive belief in conventional medicine. Convergent validity of the FICAM Scale was supported by multiple relationships between its subscales and previously established instruments. Results of independent samples t-tests indicated that there were statistically significant differences in the FICAM scores among participants, according to gender, age and stage of cancer.

Conclusion:This is the first rigorously developed instrument to measure factors influencing CAM use in individuals with cancer. The study utilized an exploratory sequential mixed-methods approach incorporating qualitative data from in-depth interviews to inform and guide the development of a new quantitative survey instrument. Using the PRECEDE Model as its theoretical framework to inform the structure of items, and the accepted methods in instrument development as outlined by the EORTC, helped to ensure the development of a psychometrically sound scale. Preliminary testing on the newly developed FICAM scale has provided sufficient evidence to warrant further testing. Understanding motivations for CAM use is vital as disclosure to conventional medical practitioners is not routine and may significantly impact the efficacy of cancer treatments. The FICAM Scale has the potential to help and inform practice and research about CAM communication and use.