Psychometric evaluation of multidimensional pain inventory (Swedish version) in a sample of elderly people.

OBJECTIVES The aim was to psychometrically evaluate the Swedish version of the Multidimensional Pain Inventory (MPI-S) and the "brief screening version of MPI-S" for use in an elderly sample. METHODS This study comprised 175 people aged 76-99 years reporting pain and in need of help to manage daily living. The instrument's factor structures were investigated through factor analyses, convergent and discriminant validity were assessed through inter-scale correlations and correlations with items from SF-12. Reliability was assessed by Cronbach's alpha. RESULTS The full-length version of MPI-S did not, in general, show any satisfactory validity and reliability when used among elderly. It had acceptable convergent and discriminant validity, but the factor analysis did not show a good model fit. Low alpha values were found for most of the sub-scales. However, the brief screening version of MPI showed acceptable validity and reliability, except for rather low alpha values in sections 3 and 4. CONCLUSION The MPI-S instrument may not be very useful for measuring pain among frail elderly. The brief screening version may instead be a better alternative to the full version of the MPI-S. However, the small number of observations may be the reason to the lack of fit, and further studies are warranted.