Methods: Prospective, longitudinal study of English speaking adults presenting to otolaryngology clinics at one of three hospitals for a suspected or newly diagnosed head and neck cancer, who faced a treatment decision. Data gathered: audiotapes of physician consultations, audio diaries kept by the patient on a portable cassette recorder, and a post treatment decision semi-structured interview.

Results: 25 patients were enrolled, 22 completed post decision interviews (1 died, 2 lost to follow up). Mean age was 58, 14/22 patients were male. Discussions of choice in the consultation: 2/22 patients' consultations with physicians had no discussion of treatment choices when treatment choices were available. 20/22 had consultations with physicians in which choices of some sort were discussed – either treatment (eg: radiation and/or chemotherapy vs. surgery vs. palliation), or timing and type of surgery. Palliation was framed in passive terms (eg: let nature take its course). All patients recalled and understood their treatment choices, if offered. Role in decision-making: 10/22 patients described feeling unable to question treatment recommendations because of: intimidation, fear of abandonment by the physician, or lack of knowledge about medicine. ‘Choice' in decision-making: 10/20 patients who were offered choices, even after describing their choices, stated they felt they had ‘no choice', either because they were facing death, or because the illness was cancer. Common characteristics of those feeling they had ‘no choice' were a) one of the treatment choices offered was palliation (all patients offered palliation felt they had no choice), and b) severe pain (9/10 of those patients described the presence of serious physical discomfort due to their cancer, either in their diaries or during their interviews).

Conclusions: The decision-making process for cancer treatment is distinct from decision-making in some other medical situations. The twin drivers of pain and fear of death in cancer treatment decision-making cannot be overestimated, and have a substantial effect on patients during decision-making. However, patients in this study showed good recall of treatment options. Problematic power differentials during consultations, and framing effects for treatment options, remain common. Application of these findings to decision support tool development would improve their likelihood of success.