BACKGROUND

Depression in nursing-home residents is often under-recognised. We aimed to establish the effectiveness of a structural approach to its management.

METHODS

Between May 15, 2009, and April 30, 2011, we undertook a multicentre, stepped-wedge cluster-randomised trial in four provinces of the Netherlands. A network of nursing homes was invited to enrol one dementia and one somatic unit per nursing home. In enrolled units, nursing-home staff recruited residents, who were eligible as long as we had received written informed consent. Units were randomly allocated to one of five groups with computer-generated random numbers. A multidisciplinary care programme, Act in Case of Depression (AiD), was implemented at different timepoints in each group: at baseline, no groups were implenting the programme (usual care); the first group implemented it shortly after baseline; and other groups sequentially began implementation after assessments at intervals of roughly 4 months. Residents did not know when the intervention was being implemented or what the programme elements were; research staff were masked to intervention implementation, depression treatment, and results of previous assessments; and data analysts were masked to intervention implementation. The primary endpoint was depression prevalence in units, which was the proportion of residents per unit with a score of more than seven on the proxy-based Cornell scale for depression in dementia. Analyses were by intention to treat. This trial is registered with the Netherlands National Trial Register, number NTR1477.

FINDINGS

16 dementia units (403 residents) and 17 somatic units (390 residents) were enrolled in the course of the study. In somatic units, AiD reduced prevalence of depression (adjusted effect size -7·3%, 95% CI -13·7 to -0·9). The effect was not significant in dementia units (0·6, -5·6 to 6·8) and differed significantly from that in somatic units (p=0·031). Adherence to depression assessment procedures was lower in dementia units (69% [SD 19%]) than in somatic units (82% [15%]; p=0·045). Adherence to treatment pathways did not differ between dementia units (43% [SD 33%]) and somatic units (38% [40%]; p=0·745).

INTERPRETATION

A structural approach to management of depression in nursing homes that includes assessment procedures can reduce depression prevalence in somatic units. Improvements are needed in depression screening in dementia units and in implementation of nursing-home treatment protocols generally.

FUNDING

The Netherlands Organization for Health Research and Development.