ABSTRACT
OBJECTIVES: The study aims to evaluate the psychometric properties of the German version of a scale for the assessment of pain in advanced dementia (PAINAD-G). DESIGN: Cross-sectional study. SETTING: Eight nursing homes. PARTICIPANTS: Ninety-nine residents in 8 nursing homes diagnosed with Alzheimer's disease (68.3%) or other types of dementia (31.7%) participated after informed consent was obtained from their proxies. Nurses in charge of the residents observed their pain behavior over a 2-minute period while performing routine nursing activities--once in the morning and once in the evening. MEASUREMENTS: PAINAD-G relies on the observation of 5 behavioral categories indicative of pain: breathing, vocalization, facial expression, body language, and consolability. RESULTS: Psychometric analyses revealed good internal consistency of the scale (Cronbach's alpha = 0.85). Inter-rater stability amounted to r = 0.80 and retest reliability to r = 0.90. Principal component analysis allowed the extraction of one factor that accounted for 63.5% of the cumulative factor variance. Validity data shows that PAINAD-G scores were higher in residents assumed to suffer from pain in comparison to those without pain. On the other hand, the level of pain rating did not correspond with the PAINAD-G scores. Residents rated to suffer from pain showed more pain behavior with increased cognitive deterioration. Measures that indicate nonpain disorders did not correlate with the PAINAD-G scores. DISCUSSION: PAINAD-G is a 1-dimensional scale that demonstrates good reliability. The outcome supports the assumption that the scale actually measures pain.
