Risk indicators of depression in residential homes
Eisses AM, Kluiter H, Jongenelis K, Pot AM, Beekman AT and Ormel J;
Commented by , 23 Aug 2004
Objective
To assess risk factors for depressive symptoms in elderly residents of residential homes.
Methods
This cross-sectional study used baseline data from another intervention study. 23 residential homes in the region of Drenthe in the Netherlands were approached for inclusion. Residents were requested to participate by letter; of those consenting, people with Mini Mental State scores below 15 and/or those with severe hearing or speech problems were excluded.
Measurements
Depression was rated with the 30-item Geriatric Depression Scale (GDS) and Risk factors for depression were assessed in 8 domains: age, sex, socioeconomic status, physical health, life events, social support, personality and family history/vulnerability to depression.
Statistics
Firstly, bivariate associations between GDS scores and putative risk factors for depression were sought by Pearson's correlation and t-tests, followed by step-wise multiple linear regression.
Results
11 out of 23 homes were willing to participate. After exclusions, data was collected on 479 residents (of a total of 597 potential participants from the 11 homes). 75 % were female, mean age 85.4 years.
12 out of 21 factors analysed were statistically significantly associated with depressive symptoms (defined by a score of > 10 on GDS): hearing impairment, no religious affiliation, incontinence, recent hospital admission, blindness, familial vulnerability, younger age, loneliness, pain, functional impairment, neuroticism, and lack of support.
Multiple regression analysis suggested that neuroticism, loneliness, functional impairment, higher educational levels and lack of social support were the factors most strongly associated with depressive symptoms.
Discussion
This study identified risk factors for depression, but the cross-sectional design precluded deductions based on causal pathways. Participants’ answers may have been coloured by low mood or recall bias. The authors did not attempt to make a clinical diagnosis or use a diagnostic classification such as ICD-10, relying instead on the raw score on the GDS and the catch-all category of "depressive symptoms".
Although nearly half the homes were not willing to participate, the residents included were assumed to be broadly representative of the 5 % of Dutch people who live in residential homes.
Notwithstanding these limitations, the main finding is that correlates of late life depression are comparable to those in the community, i.e. depression in residential homes has no specific nature or aetiological pathway.
Factors not associated with depression (or only weakly associated) included gender, age, recent life events, recent hospital admission and most surprisingly, chronic physical illness.
It is well established in the literature that the prevalence of major depression and depressive symptoms is particularly high in residential home dwellers. In assessing people in this high risk group, health professionals can be reasonably confident that
a) there is nothing unique about depressive symptoms in residential homes
b) residents who are particularly lonely or emotionally unstable, or have substantial functional impairment, have the highest risk of depression.