Item distribution, structural validity and internal consistency of the German version of the sleep disorders inventory for people living with dementia in nursing homes

The Sleep Disorders Inventory (SDI) is a frequently used proxy measurement to assess sleep disturbances for people living with dementia. This study examined the item distribution, structural validity and internal consistency of the German SDI version for people living with dementia in nursing homes. This cross-sectional secondary analysis used baseline data collected in 2021 from the MoNoPol-Sleep trial and additional cross-sectional data collected between November 2022 and November 2023. Inclusion criteria for people living with dementia were as follows: at least one sleep problem according to the SDI, a diagnosed dementia, or a Dementia Screening Scale Score of 3 or more, and at least two weeks living in the nursing home. The analysis included descriptive statistics for the SDI item distribution, an exploratory factor analysis for the evaluation of structural validity and the calculation of Cronbach’s Alpha for the analysis of internal consistency. For data analysis, 300 people living with dementia were included. The calculated item distribution showed that, with the exception of Item seven, all items demonstrated acceptable item difficulty for both frequency and severity ratings. The exploratory factor analysis showed a one-factor model with 56.6% explained total variance. The factor “Sleep disturbances at night” (five items) explained 34% total variance. The items “Sleep disturbances during the day,” and “Other night-time behaviours that bother you” showed no adequate loading to a possible second factor. The latter item does not capture additional sleep disturbances but rather provides a more detailed description of the behaviour of people living with dementia. Cronbach’s alpha was 0.71 for the identified factor. The SDI is internally consistent and structurally valid to measure sleep disturbances at night. Only one item is available for recording daytime sleepiness. Further instrument development should start at this point and examine whether additional items are needed to adequately capture daytime sleepiness. Item seven may be omitted in future data collections. In addition, further research is needed to evaluate test-retest reliability, inter-rater reliability, criterion validity and sensitivity to change.