Care staff training in detection of depression in residential homes for the elderly.
Eisses AMH, Kluiter H, Jongenelis K, Pot AM, et al.;
Commented by , 16 Jun 2005
Aims
To determine the effect of staff training on the detection, treatment and outcome of depression in residents of 10 residential homes in Holland.
Method
A randomised controlled trial was conducted in 10 residential homes for older people in The Netherlands. The intervention, directed at care staff, consisted of training in using a standardised screening instrument (the Behaviour Rating Scale for Psychogeriatric In-patients), plus review of findings of screening at a team meeting; compared with controls, who received treatment as usual. The effects of the intervention, eg., successful detection, were assessed separately.
Participants included all consenting residents; exclusion criteria were profound deafness, aphasia, or Mini Mental State Score < 15.
Residents were assessed at baseline and follow-up by blinded raters with the 30-item Geriatric Depression Scale (GDS), and using a cut-off of 10/30 as indicative of depressive illness. Care staff were asked to rate at baseline and follow-up, whether residents were depressed or not depressed.
Results
5 matched homes were assigned to the experimental group, 5 to the control group. 426 residents were included at baseline: 113 were excluded and 52 declined to participate. There were no significant baseline differences between control and experimental groups. At follow-up, data was available on 173 residents in the experimental group, 187 in the controls. 77 % of residents were female, 11 % had a score > 10 on the GDS.
In the experimental group, 27 residents had depressive symptoms at baseline, compared to 34 in the control group.
11/33 residents in the control group received treatment compared to 1/26 in the experimental group. 7 out of 12 in the experimental group improved at follow-up, compared with 3 out of 19 in the control group. Analysis of those who dropped-out and those who were assessed twice revealed that the mean score on GDS at baseline was significantly higher in those who dropped-out.
Care home staff rated residents as depressed or not depressed: overall, their ratings did not correlate well with GDS scores in either experimental or control groups.
The authors definition of a favourable response to depression was if the GDS score had fallen at follow-up: on this parameter, 58 % of the experimental group improved vs 15 % of the control group.
Dr Seymour's comments
This is a rather complex study in a field that is notoriously difficult to research. As in other studies, there were a high number of drop-outs, multiple potential confounding factors, and difficulties with outcome measures. The number of residents with depression identified in this study was surprisingly small compared to previous prevalence studies.
Depression in residential homes is a major public health problem, contributing to morbidity and mortality in older people and often leading to a miserable death. Any attempt, therefore, to improve detection or treatment is to be welcomed. The intervention in this study improved care staff’s detection of depression, without wrongly rating non-depressed residents as depressed (ie., picking up false positives).
The other main finding of this study was that the treatment rate of residents with depression increased after the intervention. Therapeutic nihilism in this patient group is therefore not justified.