treatment compliance refers to performing behaviours which follow a course of treatment that has been agreed upon by the patient and the health care professional. noncompliance can result in prolongatiorr of current health problems, may lead to more extensive treatments, may compromise the effectiveness of treatment and may increase the chance of recurrence. the present pilot study, carried out at the thunder bay regional cancer centre, assessed the compliance of 20 female breast cancer patients taking tamoxifen, an oral medication considered to have prophylactic benefits. measures were taken on desirability of control (burger & cooper, 1979), health locus of control (wallston, wallston & devellis, 1978), mood states (the profile of mood states, celia et al., 1987), quality of life (functional living index - cancer, schipper et al., 1984), and the patients' perceptions of the cancer centre, their relationship with their oncologist, and their own likely rate of compliance. further, half of the subjects were asked to complete a daily diary of their medication taking. each patients' physician was asked to provide information on diagnosis, present status, physical functioning rating and likely compliance rate. subjects were seen four times over a six-week period. at these meetings, the investigator counted the number of pills (tamoxifen) remaining in the bottje and asked the subject to complete the quality of life measure. subjects were not told this was a study of compliance. it was hypothesized that: 1. patients scoring lower on the quality of life measure would be more likely to be compliant than those whose quality of life was high. 2. patients who displayed an internal locus of control, who scored high on desirability of control, and whose relationship with their oncologist allowed for this internal/high desire control would be more compliant than similar patients whose relationship did not allow for this control. 3. patients showing signs of mood disturbance would be less compliant to the medication regimen. 4. patients regarding the centre highly would be more compliant. 5. patients on the regimen for a longer period of time would be less compliant than those just beginning the regimen. 6. patients' predictions about their own rate of compliance would accurately reflect actual compliance, while oncologists' predictions would not be accurate and would overestimate actual compliance rates. 7. patients who kept a daily medication diary would be more compliant than those not using a diary. the hypotheses were not supported by the data with the exception that oncologists' predictions of compliance were not related to actual compliance. however, age was significantly related to compliance in that older women were more likely to be compliant to the medication regimen. also, it appears that oncologists base their estimations of compliance on the length of time which has elapsed since the patient first began treatment - a factor showing no relationship to actual compliance rate. several tentative, albeit statistically nonsignificant, relationships provide compelling suggestions for future research. also, implications for the care of cancer patients are discussed.