sidered and valued. This is resonant with a baseline commitment to patient autonomy. Attending to patient preferences is not only morally laudatory, because it is part of honoring patient self-determination, but also may increase adherence and, by that means, efficacy as well.

Questions will inevitably arise about just how much evidence about a CAM therapy is needed before it becomes a part of the treatment options routinely mentioned or offered to patients. It can be argued that it is unethical to offer any CAM therapy until evidence from an RCT is complete. The stronger argument is that judgments about how much evidence is enough must be referenced to a diversity of factors, including the efficacies of conventional medical therapies; the hazards of the CAM modalities in question, if any; the extent to which these CAM modalities are preferred by patients; and the overall quality of whatever evidence exists. It should not be forgotten that RCTs of many conventional medical therapies that are routinely recommended and widely used have not been conducted. Posing the RCT as the sole test and necessary barrier for acceptance of a CAM therapy is questionable and may not be applicable with respect to certain CAM modalities. Requiring RCTs may undermine the commitment to medical pluralism.

As noted above, Adams and colleagues (2002) offer a helpful framework for addressing the therapeutic relationship in the absence of significant evidence, when there is no standard efficacious treatment or when conventional therapy has failed, and when the patient’s intention to use a CAM therapy is strong and persistent. When the implications of this framework are spelled out and the patient makes an informed, autonomous choice in favor a CAM therapy, it may, indeed, be unethical for the physician to withhold either treatment or an appropriate referral (Adams et al., 2002; Cohen, 2003). The rationale is that, as a general rule, “the personal beliefs and choices of other persons should be respected if they pose no threat to other parties” (Adams et al., 2002). If a patient with a precancerous condition, for example, seeks to pursue such therapies as meditation, colonics, yoga, and reiki rather than surgery for a limited time while continuing monitoring by a physician, it is ethically compelling for the physician to honor the patient’s core values and internal sense of integrity while compassionately discussing the physician’s perspective (Adams et al., 2002). The physician who cannot in good conscience support this choice can document informed refusal of care and needs to consider the caring commitment that is part of the ongoing relationship with the patient (Adams et al., 2002). Such a physician may feel that it is medically incorrect, given his or her training and experience, to elect to support the patient’s preference for CAM; and even though the physician is compassionate, he or she may be unable to administer or support care effectively. The ethical obligation of nonabandonment requires that if the physician opts to withdraw from