Psycho-social factors of terminal illness were explored with 36 terminally ill hospice patients. These findings were compared with staff accounts to assess whether staff held different perceptions of terminal illness from patients. The dynamic nature terminal illness was explored by re-interviewing thirteen of the original patient-staff dyads 4-6 weeks later. Psychological responses to cancer have been shown to be affected by a variety of psycho-social factors, including age, length of illness and social support. The experience of terminal illness may be different from previous stages of cancer, possibly due to the patient's realisation of the nearness of death (Yalom, 1980). Models of dying highlight the patient's perspective and social environment in the determination of the dying process. Possible communication difficulties between patient and staff and the use of 'proxy data' in clinical and research settings highlights the need to explore patient and staff perspectives more fully. A measure which described positive and negative experiences of terminal illness was developed (POTIS) and administered with the Hospital Anxiety and Depression Scale to terminally ill patients within a hospice. Staff were identified with a patient and their ratings of the patient's experience gathered. Thirteen patients/staff pairs were re-interviewed using the above procedure. Results indicated that staff rated patients as being more anxious than patient's own ratings. There was no difference between staff and patient ratings of depression. Cluster analysis was used to describe natural groups occurring within patient responses to the POTIS and patient demographic data used to describe these groups. Small sample numbers limited conclusions regarding change of perceptions over time. Issues regarding staff and patient perceptions were discussed in terms of the clinical and research use of proxy data. Possible problems with the internal reliability of measures and sample bias which occur with this population were discussed.