[Psychometric properties of the Activities Daily Life Scale (ADL)]. / Caractéristiques psychométriques de l'échelle des activités de la vie quotidienne (AVQ).

OBJECTIVE: Deficits in social functioning are an important core feature of mental health. Recently in France, the Activities Daily Life (ADL) scale has been proposed by the French authorities to assess social functioning for all hospitalized patients in a psychiatric ward. The perspective is to use this scale in the financing and organization of mental health services in France. The ADL scale is a 6-item (dressing/undressing, walking/mobility, eating/drinking, using toilets, behaviour, relationships/communication) heteroquestionnaire completed by a health care professional at the beginning of each hospitalization, assessing functioning of patients suffering from mental health diseases. However, limited consensus exists on this scale. The psychometric properties of the ADL scale have not been assessed. There is a pressing need for detailed examination of its performance. The aim of this study was to explore ADL psychometric properties in a sample of hospitalized patients in a psychiatric ward. METHOD: We retrospectively analyzed data for all episodes of care delivered to hospitalized patients in a psychiatric ward in our French Public Hospital from January 1, 2008 to June 30, 2008. The study involved retrospective review of administrative and medical databases. The following data were collected: age, gender, diagnoses based on the International Classification of Diseases - 10th version, ADL scale and Assessment of Social Self-Sufficiency scale (ASSS). The psychometric properties were examined using construct validity, reliability, external validity, reproducibility and sensitivity to change. Data analysis was performed using SPSS 15.0 and WINSTEP software. RESULTS: A total of 1066 patients completed the ADL scale. Among them, 49.7% were male, mean age was 36.5 ± 10.8, and 83.5% were single. Schizophrenia, schizotypal and delusional disorders (40.0%), mood disorders (27.9%) and mental and behavioural disorders due to psychoactive substance use (12%) were the most common diagnoses. Factor analysis with varimax rotation identified a 2-factor structure accounting for 82% of the total variance. The first dimension (ADL 1) comprised four items and represented personal care activities. The second dimension (ADL 2) comprised two items and represented social functioning. A floor effect was reported for ADL 1 and its unidimensionality was not satisfactory: two items showed an INFIT statistic outside the acceptable range. Internal consistency was satisfactory for the two dimensions: each item achieved the 0.40 standard for item-internal consistency. The correlation of each item with its contributory dimension was higher than with the other (item discriminant validity). Cronbach's alpha coefficients ranged over 0.70 in the whole sample. Concerning external validity, positive correlations were not systematically found between ADL and ASSS dimensions. The score of ADL 1 had medium to high correlations with four dimensions scores of the ASSS, while the score of ADL 2 were not at all or weakly correlated with ASSS dimension scores. Globally, ADL did not cover sentimental life and social relationships. There were statistical associations between ADL and age or gender: women and subjects older than 60 had a higher level of dependency. We didn't find any association with marital status or diagnoses. The ADL scale presented a good reproducibility but was not sensitive to change. CONCLUSION: The psychometric properties of the ADL scale were not sufficient for several parameters such as validity or sensitivity to change, contrary to other available French scales. The use of a heteroquestionnaire rather than a self-administered questionnaire should be discussed by professionals and the French authorities. These results should be taken into account in the use of the ADL scale for the economic and administrative management of psychiatry. Further research should be conducted to confirm these results.