Depression and Quality of Life in Older Persons: A Review

The World Health Organization (WHO) has predicted that by 2020 depression will become the third leading cause of disability worldwide [1]. Depression in older persons (≥60 years) is prevalent in community living settings [2,3,4,5,6,7,8] and even more prevalent among older individuals who have been hospitalized due to serious physical diseases or institutionalized due to reduced physical and/or cognitive functioning [9,10,11,12]. Known risk factors for depression are female gender [1,13], older age [14,15], poorer coping abilities [16], physical morbidity [2,4,9,17,18,19,20,21,22], impaired level of functioning [2,5,6,9,13,18,23,24,25,26,27,28,29], reduced cognition [2,3,8,20,30,31,32,33,34,35], and bereavement [13,36]. Depression has been associated with an increased risk of mortality [2,37], and poorer outcome of treatment of physical disorders [4,10]. In addition, depression may influence quality of life (QOL) negatively [38,39,40].

The WHO defines the concept of QOL as ‘individuals' perception of their position in life in the context of the culture and value systems in which they live and in relation to their goals, expectations, standards and concerns' [41]. QOL is a multidimensional and multilevel concept. The QOL concept is often divided into three levels, where global QOL is at the highest level in a hierarchy, followed by generic health-related QOL (HQOL) at the next level, and disease-specific HQOL (not included in this review) at the third and lowest level [42]. Global QOL may include general life satisfaction (LS) and covers general feelings of well-being (WB) [38,42] and other aspects such as economic situation, health, social and/or spiritual aspects of life [43]. Generic HQOL usually includes domains such as physical, psychological, social, and environmental evaluations of life [38], with both positive and negative aspects [38,41]. Therefore, generic HQOL is a more comprehensive concept than the current health status of an individual.

Several studies worldwide have explored the relationship between depressive symptoms or a depressive disorder and QOL in older persons, but as far as we know, there is no review. A review can offer a summary of existing quantitative research with a quality assessment of each study to contribute to a better understanding of the relationship between depression and QOL. The wide range of definitions of QOL complicates the research field and may make comparisons between studies difficult. Furthermore, the wide variety of assessment instruments for both depression/depressive symptoms and QOL [2,3,4,5,6,7,8,10,18,20,21,25,26,34,44,45,46,47,48,49,50,51] increases the need for a review of the existing research. Given these challenges, the aim of the present study was to review the literature on the association between depression and QOL in older persons.

We conducted a systematic, computerized search in the MEDLINE, PubMed, PsychINFO, EMBASE and CINAHL databases (end date March 9, 2014). We used the terms ‘depress' (with truncation, which included all words that contained depress, such as depression, depressed, depressive etc.), AND ‘older persons' OR ‘aging' OR ‘elder care' OR ‘geriatric patient' OR ‘geriatric psychiatry' OR ‘geriatric psychotherapy' AND ‘quality of life' OR ‘life satisfaction' OR ‘well-being'. According to database-specific rules, CINAHL headings, key words, and MeSH terms were combined, as in ‘depression in old age' OR ‘depression in the elderly'. In addition, reference lists were screened to find studies that were otherwise not detected by the systematic searches. Studies were included in the review if they met the following criteria:

• mean age of studied subjects ≥60 years,

• a quantitative design,

• depression was classified according to established diagnostic criteria [Diagnostic and Statistical Manual for Mental Disorders (DSM); International Classification of Diseases (ICD)] or assessed and defined by a specific depression instrument,

• at least 1 assessment instrument for global QOL or generic HQOL,

• an assessment of a relationship between depression and the concept of QOL was undertaken in the same individuals,

• the study was published in a scientific referee-based journal written in English.

Studies were excluded from the review if they were theoretical, reviews, editorials, comments or disseminations.

The titles and abstracts of 953 record hits were screened; 26 additional records were located by examining reference lists to identify relevant publications not detected by the computerized search. After screening titles and abstracts, 523 studies were kept for full-text screening and evaluation based on the inclusion and exclusion criteria. The search, screening, and full-text screening were performed by 2 researchers (H.S. and A.-S.H.). Detailed information about the studies that were identified, screened, assessed for eligibility, and included in this review is presented in the PRISMA flow diagram (fig. 1) [52].

Based on theoretical considerations and methodological aspects, the quality of each of the studies was assessed according to predefined criteria [53,54]. Eight quality criteria (fig. 2) were used. The first five were scored as 0 or 1, whereas the three latter were scored as 0, 1, or 2 because valid and reliable information on depression and the QOL assessment was considered to be of leading importance in the evaluation covering the main focus of this review. Thus, the quality score varied between 0 and 11.

Choosing a summary cutoff score for ‘high quality' remains arbitrary [55] but is usually within a threshold between 50 and 70% of the maximum obtainable points [16,53,56], and an a priori cutoff value for ‘high quality' above 60% was established. A study was considered to be ‘high quality' when it scored ≥7 points, and ‘low quality' when it scored ≤6 points of the maximum obtainable 11 points. This simple method recommended in the Cochrane Handbook for Reviews was selected to ascertain the validity of the review [55,57].

In total, 52 studies had a cross-sectional design, and 22 studies had a longitudinal design (tables 1, 2). The concept of QOL was defined in 25 studies. The results of the quality assessment showed that one longitudinal study received 11 points (1.4%). Of the 74 studies, 32 cross-sectional (43.3%) and 21 longitudinal (95.5%) were high quality, according to our evaluation. Studies that received ≤6 points lacked four or more of the quality criteria.

No studies reported difficulties administrating the QOL instruments. In all, 45 studies reported that they excluded individuals with cognitive impairment or a diagnosis of dementia, of which 30 were cross-sectional studies (58.8%) [11,12,17,19,24,30,31,32,33,37,58,59,60,62,63,64,65,66,69,72,73,75,76,77,78,88,89,90,91,92] and 15 were longitudinal studies (68.2%) [2,3,4,8,10,18,20,21,25,26,34,45,46,48,49]. Cognitive function was equally assessed in clinical studies (n = 11) and in community-based studies (n = 12). The Mini-Mental State Examination (MMSE) was used in 23 studies [2,8,10,12,17,19,20,25,26,31,32,33,37,48,58,62,66,69,73,78,88,89,90] for exclusion purposes. To exclude participants, the cutoff score for the MMSE varied considerably between the studies, i.e. a short form was used in one study with a cutoff ≤5. Otherwise, the range in the original MMSE was ≤9-28 (tables 3, 4).

Of the 74 studies, 34 studies were clinical studies and 39 studies were community-based epidemiological studies. One study was carried out simultaneously in hospitals, primary health-care settings and in the community (tables 3, 4). The most frequent setting in the clinical studies was psychogeriatric hospitals (inpatients, 4 studies; outpatients, 7 studies) and medical hospitals (inpatients, 6 studies; outpatients, 3 studies). The remaining clinical studies (14 studies) were carried out in primary care settings such as GP practices, nursing homes, long-term care or assisted living facilities. The mean age of the participants in the studies varied between 62.8 and 86.5 years. The proportion of females for all studies was 70% or higher in 26 (35.1%) of the studies we reviewed.

High-quality cross-sectional studies (32 studies) were conducted in clinical settings (16 studies), hospitals (11 studies), and primary health care (5 studies). The remaining 16 studies were community based. The high-quality longitudinal studies (21 studies) were conducted in clinical settings in 11 studies, in hospitals (8 studies), and primary health-care settings (3 studies), while the remaining 10 studies were community based.

The 74 published studies showed notable differences in the usage of diagnostic procedures and assessments of depression/depressive symptoms across regions of the world (tables 3, 4). While structured clinical interviews were commonly used in North America (USA and Canada; 43.5%) and Asia (including Russia and the United Arab Emirates; 36.8%), they were less frequently used in Europe (23.8%). The use of QOL assessment instruments across regions varied. While the use of the WHO Quality of Life Scale (WHOQOL-BREF, WHOQOL-100, and WHOQOL-OLD) instruments dominated in Europe (8 of 21 studies), LS instruments dominated in the USA (13 of 23 studies), whereas in South America (Mexico and Brazil) almost all studies used the Medical Outcomes Study short form (SF-36; 6 of 7 studies). In Asia, no instrument seemed to dominate.

All continents were represented among the 74 studies. Studies from North America (31.1%), Asia, the Middle East and Russia (28.4%), and Europe (27%) were most common, followed by South America (9.5%) and Oceania (Australia; 2.7%). One study (1.4%) included 20 countries from four continents [80].

In all, 18 different instruments were used to assess depression/depressive symptoms including self-report instruments, observational inventories, structural interviews and/or diagnostic evaluation (DSM/ICD; table 5). More than 1 assessment instrument for depression/depressive symptoms was employed in 22 studies. The Geriatric Depression Scale (GDS) [94,95] was used in 39 studies and was the most common assessment instrument, while 10 different instruments were used only once for the specific study.

In all, 24 different instruments were used to assess QOL (table 6). We categorized these assessment instruments according to the QOL concept hierarchy, i.e. global QOL including WB and general LS and secondly generic HQOL. QOL was assessed at two concept levels in 6 studies [10,33,45,78,84,91], and one study used a previously unknown assessment instrument that we categorized as a global QOL assessment [50]. The global QOL and generic HQOL assessment instruments were employed equally often, i.e. in 38 and 42 studies, respectively.

Only 24 (32.4%) of the studies used assessment instruments that had been specifically developed for older persons (≥60 years), such as the Life Satisfaction Index (LSI), the Philadelphia Geriatric Morale Scale (PGC), the WHO Quality of Life Assessment for Older Adults (WHOQOL-OLD), CASP-19, the Purpose in Life Test (PIL), the Life Purpose Questionnaire (LPQ), and the Salamon-Conte Life Satisfaction in the Elderly Scale (LSES).

Thirty-one studies assessed global QOL with instruments that evaluated LS, of which 14 relied on the LSI (LSR, LSI-A; short version LSI-Z; tables 3, 4). The concept of WB was less frequently used (11 studies). In total, 9 assessment instruments of global QOL were used in one study each, 5 of which we categorized as a WB instrument. Eleven studies used >1 assessment instrument to assess global QOL and generic HQOL. One study used a combined WB and generic HQOL assessment instrument.

The most frequently used generic HQOL assessment instrument (21 studies) was the Medical Outcomes Study General Health, including older (MOS; 6- and 20-item versions) and newer versions (SF-8, 12, 20, and 36). Generic HQOL was assessed as the only QOL concept in 35 studies.

Fifteen of the high-quality studies with a cross-sectional design (a total of 32 studies) used global QOL assessment instruments (including assessments of LS in 14 of the studies, and WB and LS in 1 study). Seven studies recruited persons in clinical settings (1 study in medical hospital; 6 in primary health care), and 6 studies recruited persons from the community. Persons with cognitive impairment were excluded in nearly all of these (12 of 15 global QOL studies). Generic HQOL assessment instruments were used in 17 studies, where 11 studies recruited persons in clinical settings (5 studies in psychogeriatric hospitals; 5 studies in medical hospitals; 1 study in primary care), and 6 studies recruited persons in the community. Persons with cognitive impairment were excluded in 11 of these studies. In addition, 2 studies used global QOL (LS) and generic HQOL instruments. Both of these studies recruited persons from the community and excluded individuals with cognitive impairment.

Global QOL assessment instruments were used in 9 of the 21 high-quality methodological studies with a longitudinal design (including assessment of LS in 4 studies, WB in 2 studies and both LS and WB in 3 studies). Of these, 3 studies recruited persons in clinical settings (2 studies in a psychogeriatric hospital; 1 study in primary care) and 6 studies recruited persons from the community. Persons with cognitive impairment were excluded in 3 studies. Generic HQOL assessment instruments were used in 10 longitudinal high-quality studies, where 6 studies recruited persons in clinical settings (5 in psychogeriatric hospital; 1 in medical hospital; 1 in primary care), and 4 studies recruited persons in the community. Two studies used both global QOL (LS or WB) and generic HQOL assessment instruments and recruited persons from the clinical and community setting. Persons with cognitive impairment were excluded in all of the studies that used any of the generic HQOL instruments (12 studies), with one exception [50].

In the 53 studies with a quality score of ≥7 points (tables 1, 2), the main finding was that the severity of depression was associated with poorer QOL. This association appeared to be stable over time and independent of whether a global QOL or a generic HQOL assessment instrument was employed (tables 3, 4).

All of the high-quality cross-sectional studies that solely assessed global QOL (11 studies) reported a negative association between depression and global QOL. A higher depression symptom score was associated with poorer global QOL. Depressed persons had poorer QOL than nondepressed persons.

In the high-quality longitudinal studies that solely assessed global QOL (9 studies), a depressive disorder or a high depression symptom score at baseline was related to poorer QOL at follow-up. An improvement in QOL was seen in fully and not fully recovered depressed persons compared to those with persistent depression.

The high-quality cross-sectional studies that assessed generic HQOL (19 studies) reported that a depressive disorder and a higher depressive symptom score were associated with poorer generic HQOL. Older persons with depressive symptoms and additional physical comorbidity had poorer generic HQOL than those without any comorbidity, independent of the depressive symptom load.

In the high-quality longitudinal studies (10 studies) that relied solely on generic HQOL, a depressive disorder and a higher depressive symptom score were consistently associated with poorer generic HQOL. Persons with a depressive disorder at baseline had poorer generic HQOL at follow-up than nondepressed individuals, and the severity of depressive symptoms at baseline had a negative effect on an improvement in generic HQOL at follow-up. Depression with physical comorbidity at baseline was associated with poorer generic HQOL at follow-up. Depressed persons with two physical conditions or more had poorer generic HQOL compared with persons with fewer physical conditions.

Generally, most of the generic HQOL domains were affected in a negative way by a depressive disorder or a severe degree of depressive symptoms, except for physical functioning (role limitations due to physical health) and in the mental domain (emotional, psychological functioning) [3,10,21,25,45,82]. Two studies found that general health and vitality (and energy level) were not affected by depression [2,21] while spiritual, body pain, and satisfaction with living arrangements domains were not affected by depression in 1 study each [2,25,45].

The 74 studies we examined used a large number of assessment instruments assessing depression or depressive symptoms and QOL. Most of the instruments for assessing depressive symptoms were self-report instruments. This diversity in the use of instruments hinders comparisons between studies and limits the potential to summarize data into estimates of the relationship between depression and QOL. Nevertheless, as one would expect, the main findings were that depression (at both the symptom and the disease level) was associated with poorer QOL, and that this association appeared to be stable over time and independent of whether global QOL or generic HQOL were studied. Furthermore, the high-quality cross-sectional studies reported that depressed persons had poorer QOL than nondepressed persons. The high-quality longitudinal studies established that depression at baseline predicted poorer QOL at follow-up. A higher baseline depression symptom score was related to poorer QOL and improvements in QOL were less likely to be detected at follow-up.

There is considerable agreement that QOL should be seen as a multilevel and multidimensional concept, but there is an absence of consensus on how to define QOL and dimensions of generic HQOL [38,41]. Some of the operationalized constructs partly overlap [38], which results in different QOL constructs being assessed. To interpret the findings of the studies, it is important to explicitly define which QOL concept has been used and use a well-tested assessment instrument that covers the chosen definition. In this review, we found that QOL was defined or conceptualized explicitly in only about one third (25/74) of all the studies and that about one sixth (12/74) of the QOL assessment instruments were used only once.

Our review revealed that the choice of assessment instruments seemed to be made for cultural or geographical reasons, and did not appear to be related to a theoretical framework. Assessment instruments of generic HQOL dominated in Europe and South America, while assessments of global QOL dominated in the USA. There were also geographical differences as to how to assess depression. Structural clinical interviews to diagnose depression were most commonly used in North America and Asia, but were seldom used in Europe.

Furthermore, the negative association between depression and QOL was consistent in high-quality cross-sectional and longitudinal studies, independent of the type of sample studied. In the longitudinal studies of psychogeriatric patients, global QOL and generic HQOL domains were negatively affected by the severity of depression. Recovering from depression after treatment resulted in higher QOL, and the QOL increased even in patients who did not fully recover from the depressive episode. As time goes by, older persons may accept their loss of health and function due to biological and psychosocial changes to some extent, and thus may lower their expectations and adjust their internal standards to level out the discrepancy between the possible and the actual situation (‘response shift') [140]. Consequently, QOL may be rated higher at follow-up even if an individual's health has not improved substantially. However, our review revealed that patients with more severe depression at baseline before treatment were less likely to experience improved QOL. We do not have a firm explanation for this, but it may be that persons with severe depression do not have the internal resources or capacity to adapt or adjust over time in the same way. It is evident that poor resources and coping strategies are associated with depressive disorder or severity of depressive symptoms in older persons [16].

The cross-sectional and longitudinal studies of medical and primary health-care patients also showed that the severity of depression was related to poorer global QOL and generic HQOL at baseline and follow-up. The prevalence of depression in older medical inpatients and patients in primary health care has been reported to be high in several studies [12,30,45,60,75]. Depression or depressive symptoms are the most common comorbidity in older persons with physical health difficulties [15,141,142]. Because depression affects QOL negatively regardless of medical health, it is important to detect depression and treat depressed patients. Thus, it is highly recommended that the health personnel in specialist and primary health-care settings have a dual treatment perspective, including both mental and physical health.

The negative association between depression and QOL was confirmed in numerous cross-sectional and longitudinal epidemiological community studies. In the longitudinal studies, as presented in table 4, the follow-up period varied from 3 months to 6 years, and there were mostly 2 assessments, with a variation from 2 to 7. In addition to depression, the studies also looked at and controlled for a wide range of risk factors for poor QOL. In the studies that had long-term follow-up, factors other than the risk factors considered by the studies might also influence QOL, such as a functional decline, stressful life events [36], locus of control [16], or a response shift in the participants' view of standards and expectations for life [49]. Despite this, the findings in the longitudinal community studies we reviewed are unambiguous: depression affects QOL negatively over time. In addition, since depression may be overlooked due to atypical symptomatology [26,30,143] or mistaken as grief over loss of health or close persons [143], it is important for health-care professionals working with older persons, health-care administrators, and health-care planners to address depression in older persons.

Our review does have its limitations. First, the literature search was mainly conducted by one reviewer. However, the computer search strategy was discussed with the co-authors, and a scientific librarian assisted in the search. Second, the search was limited to articles published in English. Thus, there may be studies in other languages with results that are different from those we reported. Third, the examination and synthesis of the outcomes were complicated due to conceptual differences in the definition of QOL and differences in study designs, sample compositions, instruments used to assess depression and QOL, settings, length of follow-up, and adjustment variables. This heterogeneity may cause validity and reliability problems, and a meta-analytic approach in evaluating the studies statistically had to be omitted due the variability in these factors. Fourth, in about 60% of the studies included in our review, it was explicitly stated that the authors had excluded persons with some degree of cognitive impairment. Assessment of QOL in persons with some degree of cognitive impairment may be difficult when using general global QOL and generic HQOL instruments; thus, disease-specific assessment instruments should be developed and used. However, the exclusion criteria for the definition of cognitive impairment varied considerably between studies. Thus, we cannot generalize the findings of this review to groups with cognitive impairment. Fifth, the approach in selecting quality criteria was chosen after advice from the Handbook for Reviews [55], but the criterion of sample size may be debatable since sample size by itself does not directly tell about the quality of the study. However, a low number of participants (<100) can influence the validity of the study due to low statistical power [53].

Sixth, only one third of the studies defined the concept of QOL. As previously stated, the lack of a theoretical foundation for QOL and the multitude of instruments make it difficult to compare study results [38,41] and draw firm conclusions. For example, global QOL most often covers a range of appraisals including economic, health, social and/or spiritual aspects of life [43] and may be expressed as overall QOL, general LS, or general feelings of WB [38,42]. Some studies seem to have defined WB not as a concept of global QOL, but as the diametrical opposite of depression. Consequently, the assessment instruments used in these studies are one-dimensional scales where the best outcome is high levels of WB, and severe depression is the worst outcome either by the use of a visual analogue scale [144] or an index of several items [129]. Thus, these studies did not examine the relationship between depressive symptom score and WB as a global assessment of QOL [145], and therefore they were not included in the review.

This review reports findings from cross-sectional and longitudinal studies and suggests a clear and consistent relationship between depression and poorer QOL in older persons in clinical and community settings. However, the diversity of assessment instruments used in the various studies limits direct comparison between studies and the potential to summarize data as estimates of the relationship between depression and QOL. There is also a need for additional studies that review the relationship between depression and QOL in older persons with cognitive impairment.

The authors declare that they have no competing interests.