Measurement properties of the Nottingham Health Profile and Short Form 36 health status measures in a population sample of elderly people living at home: Results from ELPHS

Abstract

Objectives. To investigate reliability and validity of the Nottingham Health Profile (NHP) and Short Form 36 (SF‐36) in an elderly population living at home. Design. Interview survey of an age‐sex stratified random sample of people age 65+. Methods. In all, 481 people were sampled from 78 general practices across East Anglia. The NHP and SF‐36 were administered, in random order, alongside measures of physical performance, morbidity, symptoms, activities of daily living (ADL) and the Hospital Anxiety and Depression Scale (HAD). Repeat interviews were completed by 93 individuals after one month. Results. Neither questionnaire showed bias in test‐retest reliability at 1 month; the change in scale scores showed greater variability for the SF‐36 than the NHP. Internal consistency was comparable with Cronbach's alpha in the range of .7 to .85, except for NHP social isolation (.61) and SF‐36 physical functioning (.94). Substantial ceiling effects (> 20% of cases) were found for all six NHP scales and for four of eight SF‐36 scales (role limitations owing to physical and emotional problems, social functioning and bodily pain). Comparable scales of both questionnaires exhibited a significant and comparable relationship with actual physical performance and selfreported ADL, morbidities, symptoms and HAD (p < .001 in all cases). Conclusions. This study confirms that some SF‐36 scales are more sensitive to minor levels of morbidity in elderly people, but the SF‐36 has not made substantial improvements over the NHP in test‐retest reliability, internal consistency or validity for this population. The two questionnaires appear to have equally good psychometric properties in this age group.