The Psychogeriatric Assessment Scales (PAS): psychometric properties in French and German speaking populations.

OBJECTIVES

To investigate the structure and measurement properties of the Psychogeriatric Assessment Scales (PAS) in a community sample, to study the performance of French and German translations of the PAS, and to determine the power of PAS scales to discriminate dementia and depression diagnosed to DSM-IV criteria. Design and measures The measures were gathered as part of a large community survey. Responses to PAS items were obtained in the course of the administration of the 'Canberra Interview for the Elderly'. Demographic variables, the mini-mental state examination and measures of premorbid intelligence robust to current intellectual impairment were also gathered in the course of the survey.

SETTING

A community survey in the cities of Zurich and Geneva, Switzerland.

PARTICIPANTS

Subjects were residents of the two cities aged 65 years or older. There were 465 participants from Zurich and 456 from Geneva. Interviews with an informant who knew the subject well were also obtained. Subjects from Zurich were interviewed in German, subjects from Geneva in French.

RESULTS

Confirmatory factor analyses of PAS scales showed close agreement between sites and with results obtained in the sample used to construct the instrument. Correlations with demographic and validating variables also showed this pattern. The PAS cognitive impairment and cognitive decline scales were shown to be good discriminators of DSM-IV dementia while the depression scale discriminated DSM-IV major depressive episode.

CONCLUSIONS

This study supports the PAS as a valid and reliable summary of the status of older persons with respect to the impairments, changes and symptomatology that lie at the core of dementing and depressive disorders. This study not only supports the results found in the original development sample, but demonstrates that it performs well against DSM-IV diagnostic criteria and amongst speakers of French and German.