Functional disability and death wishes in older Europeans: results from the EURODEP concerted action

In a clinical study of depressed older adults, a past suicide attempt was more common among patients who had impaired instrumental activities of daily living [17]. Two of the above mentioned case–control studies on completed suicide addressed the issue of functional disability. Conwell and colleagues [7] utilised the Lawton Instrumental Activities of Daily Living (IADL) and found that low IADL increased suicide risk. The Swedish study by Waern and colleagues [12] showed that people aged 75 years and above who required help with cooking, cleaning and shopping had an increased suicide risk compared to those who could carry out these activities on their own.

It remains unclear whether functional disability is associated with suicidal ideation independently of depression. Thus far, results are inconclusive, with some cross-sectional studies showing an independent association [18, 19], and others not [20]. A large multicenter study may provide further knowledge.

The EURODEP concerted action is a collaboration among research groups involved in population-based studies on depression in late life [21, 22]. With a total study sample of over 22,000, EURODEP provides the power to examine factors related to depression in later life. Previously reported findings from EURODEP [23] suggested a dose–response relationship between disability and depressive symptoms that was consistent across Western Europe. Eleven centres also had information regarding death wishes for a total of 15,890 persons. The aim of the current study was to explore the association between functional disability and death wishes, and to determine whether such a proposed relationship would be independent of depressive symptoms.

Representative samples of older persons aged 65–104 years (n = 15,890) living in Amsterdam (n = 3,987), Berlin, (n = 488), Dublin (n = 1,012), Reykjavik (n = 772), Liverpool (n = 3,366), London (n = 637), Ähtäri (n = 1,035), Gothenburg (n = 447), Munich (n = 346), Verona (n = 202), Zaragoza (n = 3,598) took part in psychiatric examinations. Inclusion criteria varied somewhat among centres. Sampling was based either on municipality registers or on general practitioner registers. For further information concerning the interviews and participants, the reader is referred to the detailed accounts of Copeland [21], Prince [22] and Braam [23].

Several instruments were used to assess death wishes (Table 1). Eight of the participating centres used the Geriatric Mental State scale (GMS) [24] and one used the short version of the Comprehensive Assessment and Referral Evaluation (SHORT-CARE) [25]. The questions employed were identical in both instruments (Have you felt that life was not worth living? Have you ever felt that you would rather be dead? Have you ever felt you wanted to end it all? Have you ever thought of doing anything about it yourself?). One centre used the Paykel question [26] (Have you ever wished that you were dead—for example, that you would fall asleep and never wake up again?) and another used an item from the Zung Self-Rating Depression Scale (ZSDS) [27] (I feel that others would be better off if I were dead). A person was considered to have death wishes if the death wish/suicidal ideation item of the EURO-D scale was endorsed. As decision trees differentiating between suicidal ideation and attempts varied at different sites, a more detailed analysis of specific types of suicidal behaviour was not possible.

Perceived loneliness was assessed (yes/no) at ten of the participating centres. The specific questions employed are shown by centre in the ESM Appendix 1.

Most centres rated functional disability in accordance with the Katz scale [29]. Activities of daily living (ADL) are used to measure the individual’s ability to carry out everyday activities such as bathing, dressing, toileting, transfer, continence and feeding. To harmonise the ratings, total scores of the ADL scales were trichotomised into “no,” “intermediate”, or “high” levels of disability at each centre [23]. “High” level of disability was defined as the scores which fell in the highest tertile. Data regarding specific types of chronic conditions were not available at all centres, and ten centres had data on number of chronic conditions. Numbers were categorised as none, “one” and “two or more” [23].

The EURO-D harmonised scale was constructed by expert opinion with respect to correspondence between items on different depression scales (GMS [24], SHORT-CARE [25], ZSDS [27] and the Comprehensive Psychopathological Rating Scale (CPRS) [30]), and wherever possible by probabilistic modelling for some scales [22]. The EURO-D scale includes 12 items (depressive affect, pessimism, death wishes (as defined above), guilt, sleep problems, lack of interest, irritability, appetite problems, fatigue, reduced concentration, lack of enjoyment and tearfulness). Each item receives a rating of 0 (not present) or 1 (present). EURO-D scores can thereby range from 0 to 12, with higher scores reflecting greater depression symptom burden. For the purpose of this study, we removed the death wishes item from the EURO-D scale, yielding a maximum score of 11.

The Mini Mental State Examination (MMSE) [31] was used to assess cognitive functioning and was available in nine centres. In this study, MMSE score was analysed as a continuous variable. The diagnosis of dementia was based on the AGECAT algorithm [32] in nine centres. This algorithm has previously been validated against dementia diagnosis according to clinicians and against DSM-III-R criteria with satisfactory results [33, 34]. In the Gothenburg sample the diagnosis of dementia was based on DSM-III-R criteria, using tests of short- and long-term memory, abstract thinking, aphasia, apraxia, and agnosia [35]. Dementia was diagnosed using the Wilson Mental Capacity Scale [36] in Äthäri.

A generalised linear mixed model with a logistic link function was used to analyse the association between sex, age, education, marital status, perceived loneliness, functional disability, number of chronic diseases, MMSE and EURO-D (independent variables) and death wishes (the dependent variable). Age was added as a continuous variable, and results are reported as how a 10-year increase in age increases the prevalence of death wishes. Plausible interactions (sex and functional disability; sex and chronic condition; sex and marital status; functional disability and depression) were also added to the model and tested. As the interactions proved sensitive to the inclusion or exclusion of other independent variables, a more parsimonious model containing only main effects that proved to be stable was in the end chosen. Results are presented as odds ratios (OR) and 95 % confidence intervals (CI). Odds ratios are what is termed “subject specific” and should thus be interpreted as the effect that a predictor variable has on the odds of developing death wishes for any given centre. See Agresti [37] for a detailed explanation. Further, the median odds ratio (MOR) was calculated from the intercept variance. This measure can be understood as the effect that belonging to a certain centre will have on the odds of having death wishes. If two random persons from two different centres who share the same covariate values were to be picked, one would have higher odds of death wishes and the other lower odds. Using the person with the higher odds in the numerator and the person with lower odds in the denominator repeating this procedure for every possible combination of subjects-centres would result in a distribution of odds ratios. The MOR is the median of this distribution and can be compared to the fixed effect estimates, yielding a measure of the relative size of the unexplained variation in comparison to the effects that explanatory variables have. For a more detailed explanation, see for instance Merlo [38]. Models were estimated using proc Glimmix in SAS 9.3 (SAS Institute Inc.). Descriptive statistics were analysed in IBM SPSS Statistics, version 20 for Windows.