Health-related quality of life in older depressed psychogeriatric patients: one year follow-up

Depression is one of the most common mental disorders among older adults [1], and the most common mood disorder in late life, which has led to increasing global concern as the number of older adults rises [2, 3]. About 4 to 10 % of older adults in community settings suffer from major depressive disorders [3, 4]. Risk factors for depression among older adults include female gender, physical morbidity, and an impaired level of physical and cognitive functioning [5]. Depression has been associated with an increased risk of mortality [5, 6], poorer outcome of treatment of physical disorders [7, 8],and health related quality of life in older adults [9].

Health related quality of life (HQoL) relates to the perceived effects of mental and physical health on each individual’s ability to live a fulfilling life [10, 11]; thus, it is an important outcome measure in old age [12‐14] and a multidimensional concept [10, 15]. It exists no consensus on the definition of HQoL or the questionnaires to use [16]. In the present study, HQoL was understood as a simple index formed by an algorithm where the importance of five health problems (mental and physical) were taken into account and a visual analog scale where experiences of overall health were rated [17].

Knowledge about change in HQoL after treatment of depression among older adults is an important outcome-measure of relevance for those with depression as well as their family, health care planners, and health care providers. A recent review, identified five longitudinal studies of depressed psychogeriatric in- and outpatients [9] with no follow up after discharge [18‐21] or up to 3 months after discharge [22]. Higher severity of depressive symptoms at baseline was related to poorer HQoL at discharge [18, 20] and 3 month follow-up [22]. One study have reported that the HQoL was significantly better in older adults with remission compared to older adults without remission [19]. However, this was an outpatient study without a follow-up period after discharge. Studies using a follow-up after discharge are needed to better understand how remission of depression may influence change of HQoL over time.

It is known that older patients with depression may both have poorer baseline and follow-up physical health and HQoL than older adults without depression [21, 23]. However, as far as we know, if HQoL in patients with remission after treatment using a long-term follow-up perspective is comparable to those without depression when adjusting for physical and mental health differences has yet to be studied [9]. This is highly warranted and clinically relevant information.

The aim of the present study was first to describe HQoL by the EQ-5D Index and EQ-VAS in patients with depression when first admitted to a psychogeriatric unit and then 1 year after admission. Secondly, we wanted to explore if and how improved HQoL 1 year after admission was explained by the remission of depression at follow-up, adjusting for socio-demographic variables and baseline health conditions. Thirdly, we aimed to compare HQoL in the group of patients with remission from depression 1 year after admission with a group of older community living adults without depression, adjusting for differences in socio-demographic and health conditions.

This study applied data from a 1 year follow-up study to explore the associations between changes of HQoL in hospitalized psychogeriatric patients with depression and remission at follow-up. Furthermore, HQoL data from a cross-sectional study among older community-living adults without depression (reference group) was compared with HQoL data from older patients with remission from depression at follow-up.

The baseline characteristics of the 108 depressed hospitalized older adults with follow-up information are presented in Table 1. There were no significant differences between those not participating and those participating at follow-up with respect to gender, level of education, history of vascular disease, poor physical health, previous history of psychiatric disease, baseline MMSE, and IADL and MADRS scores. Those not participating were older (mean79.1 SD 6.5 years; p < 0.01). The mean duration (SD) from hospitalization to study inclusion was 5.2 (4.7) days, mean duration of hospitalization was 68.2 (48.9) days, and mean duration of follow-up from inclusion was 418.1 (48.2) days. At follow-up, 37 (of 108, 34.2 %) patients had depression, and thus 71 had remitted from depression.

To our knowledge, this is the first 1 year follow-up study of health-related quality of life (HQoL) among hospitalized psychogeriatric depressed patients. We found that HQoL by EQ-5D Index and EQ-VAS was more improved in those with remission from depression at follow-up than those without remission, even after adjustment for socio-demographics and baseline health conditions. The HQoL in those with remission at follow-up was comparable with the HQoL in a reference group of older adults when adjusting for differences in socio-demographic and health conditions.

The finding in the present study of a significantly improved EQ-5D Index and EQ-VAS in those with remission compared to those with depression at follow-up is analogous to a previous study of psychogeriatric depressed older out-patients [19] and of depressed adults receiving primary care treatment [35]. However, the previous studies had either no or a shorter follow-up.

In the present study the adjusted explained variance for the regression models was about 27 %, similar to the explained variance in a previous study of EQ-5D Index in adults receiving primary health care treatment [35]. In that study, remission of depression, physical health, depression severity, and ability to work were related to overall EQ-5D Index at a 6 month follow-up [35]. In the recent study of hospitalized depressed patients, as in the previously mentioned study of depressed adults receiving primary health care treatment, the improvement of HQoL was not associated with socio-demographic variables such as gender, age, and education [35].

In adjusted analyses of improved HQoL, we found that improved EQ-VAS score was associated with worse baseline MADRS score, while improved EQ-5D Index at follow-up was not explained by baseline MADRS. We do not have a firm explanation for this difference, but the two HQoL measures are slightly different since the EQ-5D Index uses an algorithm that includes the valuations of the general population sample’s response to the questions asked, while EQ-VAS does not [17]. Older adults with more severe baseline depressive symptoms may experience and report changes of EQ-VAS differently than those with less severe baseline depressive symptoms. In line with this, an observational study of HQoL found that older people with depression at baseline were more likely to report improved HQoL after 12 months compared to those without baseline depression [51]. At least, it is not expected that older adults with more severe depressive baseline symptoms accept their health difficulties, lower their expectations, and adjust their internal standards to level out the discrepancy between the possible and the actual situation to a greater degree (“response-shift”) [52] than those with less severe depressive symptoms [9, 53].

HQoL relates to the perceived effects of both mental and physical health [10, 11], and the present study found that those with good baseline physical health were more likely to have improved EQ-5D Index and EQ-VAS at follow-up compared to those with poor physical health, independent of baseline MADRS score and remission of depression at follow-up. Our result is in line with a longitudinal study of EQ-5D Index in adults receiving primary care treatment [35] and a two-year observational study of community-living older adults reporting that higher physical comorbidity was associated with poorer HQoL at follow-up both in depressed and non-depressed older adults [23].

A simple, direct comparison of EQ-5D Index and EQ-VAS between those with remission and a reference sample of older adults showed that HQoL in the reference sample was better than in the patients with remission. However, those in the reference sample were younger and had higher education, better cognitive and instrumental functioning, better physical health, and lower depression symptom scores on MADRS. Adjusting for these conditions, there was no longer a significant differences in health related QOL by EQ-5D Index or EQ-VAS between the two groups. However, EQ-5DIndex may have a ceiling effect in persons with very good health [54]. The participants in the reference group had better physical health than the patients with remission from depression, but the reference group cannot be characterized as very healthy. Thus, we hold it less likely that a ceiling effect is the cause of our finding of no HQoL difference between the reference group and the patients with remission.

Even if the study has a number of strengths, such as using well-known, internationally accepted diagnostic procedures for depression and having information about variables of potential importance to the outcome, like socio-demographic and medical health variables, some limitations need to be addressed. Firstly, the loss of patients at follow-up (25 %) was large, and could have biased the results. However, those who died or were not followed-up had similar distribution of gender, level of education, history of vascular disease, poor physical health, history of psychiatric disease, and baseline MMSE, IADL, and MADRS scores. Secondly, we do not have detailed information about the treatment given (ie the antidepressant medications, ECT and/or milieu therapy). Thirdly, the reference group was not matched to the group of older adults with remission. In addition to better physical and cognitive health, the reference group was younger and had better education than the patients with remission from depression. The comparison took these differences into account. However, there might be rest confounding which the statistical analyses were not able to handle. In addition, other conditions of importance for HQoL than those we have studied might be of importance; eg major life events, social support, and social coherence [55‐57]. Thus, interpretation of the results should be done with caution. Fourthly, due to the recruitment process of the non-depressed participants using advertisements in the local newspaper, home nursing offices, senior centers, and various volunteer organizations, the non-depressed group of participants should not be regarded as representative of the general community-living older people without depression.

Finally, the present study employed the tariffs produced by Dolan et al. [17, 28] to calculate the EQ-5D Index. A Swedish study of depressed adults [35] and Norwegian studies of HQoL in older adults [30, 58] used the same tariffs. It has been questioned if EQ-5D Index calculated by use of valuation of health conditions in general populations is the most appropriate approach [35]. A suggested alternative would have been to base the evaluation of individuals who actually have the health state under study [35]. However, EQ-5D is meant to be a generic HQoL measure and such an approach would make a comparison between groups with different health conditions more difficult. Conversely, the simplicity of the Index may more poorly pick up some aspects of depression that have importance for HQoL compared to a longer, generic HQoL measure that assesses the physical, mental, and social dimensions of HQoL separately. However, the EuroQol Group Questionnaire was used since it is not demanding for older adults with restricted capacity and depression; it is cognitively relatively simple to fill in [54]. Thus, an additional longer HQoL measure was not used, but we analyzed both the data from the EQ-5D Index and the EQ-VAS, giving mainly the same results.

Older hospital patients with depression who experienced remission one year after admission gained HQoL and their HQoL was comparable with the HQoL in a reference group of older adults without depression when adjusting for differences in socio-demographics and health conditions. These findings are important patient health-outcomes and can have importance for health care planners.

EQ-5D Index, summary index of the EuroQol Group EQ-5D descriptive system; EQ-VAS, EuroQol Group visual analog scale score; GMHR, general medical health rating; HRQoL, health-related quality of life; MADRS, Montgomery-Aasberg depression rating scale; MMSE-NR, Mini-Mental-State Exam-Norwegian Revised Version