Open AccessThe Self‐Administered 24‐Item Geriatric Pain Measure (GPM‐24‐SA): Psychometric Properties in Three European Populations of Community‐Dwelling Older Adults
Author(s) -
CloughGorr Kerri M.,
Blozik Eva,
Gillmann Gerhard,
Beck John C.,
Ferrell Bruce A.,
Anders Jennifer,
Harari Danielle,
Stuck Andreas E.
Publication year - 2008
Publication title -
pain medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.893
H-Index - 97
eISSN - 1526-4637
pISSN - 1526-2375
DOI - 10.1111/j.1526-4637.2008.00497.x
Subject(s) - missing data , medicine , discriminant validity , imputation (statistics) , psychometrics , internal consistency , physical therapy , gerontology , clinical psychology , statistics , mathematics
Objective. To explore the feasibility and psychometric properties of a self‐administered version of the 24‐item Geriatric Pain Measure (GPM‐24‐SA). Design. Secondary analysis of baseline data from the Prevention in Older People—Assessment in Generalists' practices trial, an international multi‐center study of a health‐risk appraisal system. Participants. One thousand seventy‐two community dwelling nondisabled older adults self‐reporting pain from London, UK; Hamburg, Germany; and Solothurn, Switzerland. Outcome Measures. GPM‐24‐SA as part of a multidimensional Health Risk Appraisal Questionnaire including self‐reported demographic and health‐related information. Results. Among the 1,072 subjects, 655 had complete GPM‐24‐SA data, 404 had ≤30% missing GPM‐24‐SA data, and 13 had >30% missing GPM‐24‐SA data. In psychometric analyses across the three European populations with complete GPM‐24‐SA data, the measure exhibited stable internal consistency, good convergent, divergent and discriminant validity, and produced stable pain measurements. However, factor analysis indicated differences in the GPM‐24‐SA across sites with discrepancies mainly related to items of a single subscale that failed to load appropriately. Analyses including imputation for subjects with ≤30% missing data demonstrated psychometric properties comparable to complete data analyses suggesting that imputation in cases with ≤30% missing GPM‐24‐SA data provides sufficient information to generate a valid score. Conclusion. The GPM‐24‐SA is a promising tool for self‐administered assessment of pain in community dwelling older adults. However, because of incomplete response and uncertainty in factor structure, further refinement and psychometric evaluation of the GPM‐24‐SA is needed before it could be recommended for widespread use.