PURPOSE: To document self-reported pain descriptions throughout residency in a long-term care facility for a convenience sample of older adults (Canadian war veterans). DESIGN AND METHODS: The present study was a longitudinal, fixed-cohort study with anticipated attrition. Participants (n=33) were assessed every three months from admission until attrition or end of study (three years). The assessments included standardized measures of psychosocial functioning and a comprehensive pain questionnaire for residents who reported a pain problem. RESULTS: On average, respondents completed five assessments before attrition and reported pain in 69% of assessments. The likelihood of pain report decreased as the number of assessments increased. Cross-sectional analysis indicated that, at admission, pain duration was greater and nonverbal pain expression less prominent than later in residency. Longitudinal analysis indicated greater than 50% within-subject variability across three consecutive pain reports for nonpharmacological pain relief, pain expression and functional impact of pain. CONCLUSION: The present study highlights the need for patientcentred, longitudinal investigations of both the natural progression of pain and environmental factors that mediate pain management in the interests of improving pain control for this population.