Studies suggest that pain is under-recognized and under-treated in those with severe dementia. Identifying pain is the first step in its effective management. Few studies have investigated the utility of behavioural pain tools in those with advanced dementia. Participants were nursing home residents with advanced dementia who were observed on three occasions using a pain assessment tool (PAINAD). Following further assessment, an appropriate management plan was formulated for those thought to be in pain. Participants who scored above the cutoff (two) on the PAINAD scale, but were felt not to be in pain, formed the false positive group. The pain and false positive groups were reassessed at 1 and 3 months. Seventy-nine participants completed the study, with 39 participants scoring above two on the PAINAD. Of these, only 13 were assessed as being in pain. The other 26 participants who scored above the cutoff on PAINAD were not felt to be in pain. Instead, their behaviour had a psychosocial explanation, often to do with a lack of understanding as to what was happening to them. The sensitivity of PAINAD was 92%. In those with pain, a significant decrease was demonstrated in the PAINAD scores on intervention following treatment for pain (p=0.008). PAINAD is a sensitive tool for detecting pain in people with advanced dementia, but has a high false positive rate, frequently detecting psychosocial distress rather than pain. PAINAD can be used to assess whether pain management strategies have been successful.