Psychometric Evaluation of a Pain Intensity Measure for Persons with Dementia.

Objective: The goal of this study was to conduct initial psychometric analyses of a seven-item pain intensity measure for persons with dementia (PIMD) that was developed using items from existing pain observational measures.

Design and Methods: We evaluated validity by examining associations with an expert clinician's pain intensity rating (ECPIR) and an established pain observation tool (Mobilization Observation Behaviour Intensity Dementia [MOBID]). We also examined correlations between the PIMD and known correlates of pain: depression, sleep disturbances, agitation, painful diagnoses, and caregiver pain reports. We examined the differences between PIMD scores for "at rest" and "during movement" observations. We assessed reliability by calculating Cronbach's alpha and estimating inter-rater reliability using intraclass correlations (ICCs). Finally, we examined whether six additional "recent changes in behavior" items improved the PIMD's ability to predict expert clinicians' pain ratings.

Setting: Sixteen nursing homes located in Alabama, Georgia, Pennsylvania, and New Jersey.

Participants: One hundred ninety residents with moderate to severe cognitive impairment, mean age of 84 years, 49.5% female, and 70% white.

Results: PIMD during movement scores were highly correlated with the ECPIR and overall MOBID scores. As expected, there were large differences between at rest and during movement PIMD scores. Associations of PIMD with known correlates of pain were generally low and statistically nonsignificant. Internal consistency was supported with a Cronbach alpha of 0.72 and an inter-rater ICC of 0.82 for during movement PIMD scores.

Conclusions: Initial evaluation of the PIMD supports its validity and reliability. Additional testing is needed to evaluate the tool's sensitivity to changes in pain intensity.