Objectives:Unrelieved pain is present in a majority of terminal cancer patients. However, the treatment of pain in palliative and hospice care is affected by the lack of validated pain assessment. The goal of this study was to evaluate differences in pain evaluation between terminal cancer patients and physicians and evaluate the pain levels as a survival biomarker.Materials and Methods:Patients were evaluated every 7 days for a total of 4 assessments. Physicians evaluated patients' pain on an numeric rating scale (NRS) scale after clinical examination, after which the patients completed NRS, Quality of Life Questionnaire Core 15 Pal (QLQ-C15-PAL), and Edmonton Symptom Assessment System (ESAS) questionnaires.Results:On average, physicians minimally underestimated the pain level in patients (3.47 vs. 3.94 on an NRS scale). Pain was overestimated in 28% and underestimated in 46% of the patients. However, half of all underestimation was clinically meaningful, compared with 28% of the overestimation. For patients with an NRS score of 7, pain underestimation was both clinically and statistically significant (5.56 vs. 8.17). Pain ratings exhibited a very small correlation to survival (up to r=-0.22), limiting their use as a survival biomarker.Discussion:Although physicians can accurately assess mild pain in terminal cancer patients in the hospice setting, the underestimation of pain is still clinically significant in almost a quarter of patients, and especially pronounced in patients with higher levels of pain and in female patients. Hence, validated pain assessment is a necessity in hospice care, with the choice of pain evaluation tool dependent on patient and physician preference.