The psychometric quality and clinical usefulness of three pain assessment tools for elderly people with dementia.

In view of the need for valid, reliable, and clinically useful scales to assess pain in elderly people with dementia, this study evaluated the psychometric properties of translated versions of the PAINAD, PACSLAC, and DOLOPLUS-2 scales. In an observational study design, two raters simultaneously assessed the nursing home residents (n=128) for pain during influenza vaccination and care situations. The PACSLAC was valued as the most useful scale by nurses. Cronbach's alpha was high (>.80) for the total scale at T2 and T3 and adequate for the 'Facial expression' and 'Social/personality/mood' subscales. IC scores for the 'Activity/body movement' and 'Physiological indicators/eating/sleeping changes/vocal behaviors' subscales were low. It demonstrated good validity and reliability, although the scale should be further refined. This refinement should increase homogeneity. The PAINAD showed good psychometric qualities in terms of reliability, validity, and homogeneity (alpha ranged .69-.74 at T2 and T3) (except for the 'Breathing' item). The PAINAD scale had lower scores for clinical usefulness in this sample. The Dutch version of the DOLOPLUS-2 was considered more difficult to use but showed acceptable psychometric qualities in terms of the issues assessed, except for the 'psychosocial reactions' subscale. IC of the DOLOPLUS were adequate for the total scale (alpha ranged .74-.75) and almost all subscales (alpha ranged .58-.80). Findings of this study provide evidence of validity and reliability of the three pain assessment scales. Now that a pain scale is available, future studies also need to focus on its implementation in nursing practice.