Quality of life in community-dwelling Dutch elderly measured by EQ-5D-3L

In 2050 more than 25% of the Dutch population will be aged 65 years and older [1]. With an ageing population and rising health expenditure, preventing illness in elderly becomes more important. Measuring the health of the population is therefore important to guide health policy decisions, normally aimed at improving health and reducing socioeconomic differences in health. The EQ-5D-3L is one of most commonly applied generic health-related quality-of-life instrument [2, 3], and is suitable for measuring health status within an elderly population [4, 5]. There are six publications (i.e., [6‐11]) reporting quality-of-life data using the EQ-5D instrument for Dutch elderly. Stolk et al. [6] reported EQ-5D-3L-index and EQ-visual analogue scale (VAS) scores for three age-classes (i.e. 65–69, 70–75 and ≥75 years) and both sexes based on 254 respondents. Furthermore, Timmers et al. [7] reported the EQ-5D-3L index values for elderly in two age-classes of 50–69 and 70–97 years based on 1311 respondents and using data from the second Dutch National Survey of General Practice (DNSGP-2) conducted in 2001 [9]. Further used Hoeymans et al. [9] the DNSGP-2 cohort to present the EQ-domains for elderly (1008 respondents) in three age-classes (i.e. 60–69, 70–79 and ≥80) and both sexes. Szende et al. [8] reported quality-of-life (i.e. EQ-domains, EQ-5D index and EQ-VAS scores) for two age-classes (i.e. 65–74 and ≥75 years) based on 450 respondents, using data from the European Study of the Epidemiology of Mental Disorders (ESEMeD) study collected between 2001 and 2003 [12]. Further used König et al. [10] the ESEMeD study, to report on the EQ-domains and EQ-VAS scores of advanced elderly (i.e., ≥75 years) in six European countries, including the 164 Dutch responders aged 75 and older. A more recent study (data collected in 2012), reported EQ-5D-5 L index scores for the age-groups 60–70 and ≥70 years based on 281 respondents [11]. Important determinants for health identified in earlier studies were age (e.g. [8, 13‐19]), gender (e.g. [8, 13‐19]) and education (e.g. [8, 13, 14, 16, 19]). Detailed comorbidity data were in most of these studies lacking.

The aim of this study was to evaluate health status and associated factors in community-dwelling elderly in the Netherlands, using the EQ-5D-3L instrument.

The CHO-CAP questionnaire was distributed to 72,074 CAPiTA-participants. The questionnaire and informed consent were received from 48,634 subjects, corresponding to a response rate of 67.5% of those invited to participate (see Additional file 1: Figure S1).

The majority of participants were male (57%), were aged between 65 and 74 years (69.9%), were low educated (41.9%), lived in the west (36.6%) and had no comorbidities (54.8%), see Table 1 and Additional file 1: Table S1. Women, the very elderly, and lower educated elderly were underrepresented in the study population (Table 1). Non-responders had more comorbidities than those included in the study (see Additional file 1: Table S1).

The proportion of elderly with no problems, some problems and severe problems of functioning by domain and by age-group and sex is shown in Fig. 1a with underlying data in Additional file 1: Table S2. The population profile by age-group and sex, expressed as percentage reporting problems in one or several domains is presented in Fig. 1b and in Additional file 1: Figure S2. In most age-classes, pain/discomfort was the most frequently reported complaint, ranging from 24.1% (age-group 65–69) to 38.0% (age-group 80–84). Relatively few respondents had problems with self-care, ranging from 1.4% (age-group 65–69) to 5.6% (age-group 80–84). In the eldest elderly (≥85 years), mobility was the most frequently reported problem (52.9%), while anxiety/depression was least frequently reported as problem (8.1%). The proportion of persons reporting problems, and the number of domains with problems is rising with increasing age (see Figs. 1a–1b). In all age-groups and for all five domains, women reported more problems than men (see Figs. 1a–1b).

Mean population EQ-VAS and mean EQ-5D-index values were 82 and 0.92, respectively. Male respondents had both, higher EQ-VAS and higher EQ-5D-index values than female respondents, namely 82 versus 81 and 0.93 versus 0.90 respectively, see Fig. 2 and Additional file 1: Tables S3 and S4. Both mean EQ-VAS and EQ-5D-index values decreased significantly with increasing age from 84 to 0.94 in the 65–69 age-group to 76 and 0.86 in the elder elderly, respectively (Fig. 2 and Additional file 1: Tables S3 and S4).

Next to female gender (β = −0.40) and increasing age (β: −0.003), lower education (+0.011 for medium/+0.20 for high education), smoking (−0.021), having a comorbidity (e.g. asthma: − 0.16; diabetes using insulin: −0.027) and/or a previous cardiovascular event (stroke: −0.56; myocardial infarct −0.29) and living in the southern region (−0.005) were negatively associated with EQ-5D index values (see Table 2). Similar findings were found for EQ-VAS. For more details see Table 2.

Within elderly people, large quality-of-life differences exist. The proportion reporting problems was rising with increased age, as was also the proportion of persons reporting problems in more than one domain. For most age-groups and both sexes, pain/discomfort was the most frequently reported complaint. Similar to other studies, self-care problems were reported least frequent [9, 19, 26]. Only in the oldest elderly, mobility problems were more frequently reported than pain/discomfort, while anxiety/depression was the least frequent reported problem. Similar to other studies [8‐10, 13‐19] female sex was associated with more problems on each of the five EQ-5D domains, and consequently also with a lower EQ-5D-index compared to men. Increasing age, female sex, lower education, smoking, having a comorbidity and/or a previous cardio-vascular event, but also geographic factors such as living in the southern region of the Netherlands and living in an urban area, were all associated with lower EQ-5D health status.

The use of national tariffs, different response styles due to social and cultural background and different reference levels all influence the final EQ-5D-indices score [10, 13, 15, 27]. According to Szende et al. [8] it is mainly the prior living standards of a country explaining the observed cross-country differences in general health. Therefore, a direct comparison with other countries is hampered by these important cross-country differences in background. König et al. [10], who compared the health status of elderly in six European countries, namely Belgium, France, Germany, Italy, the Netherlands, and Spain, noted that the Netherlands was the country with the lowest proportion of respondents reporting any problems. This corresponds to the findings of Konerding et al. [13] who studied the health status in adult type 2 diabetes patients in six European countries, and who noted that the Dutch respondents reported fewer problems in four of the five domains, and only Finland reported fewer problems with depression and/or anxiety. In our study, the tendency of reporting problems, increased with age and was associated with being female and having a low education, similar to other studies (e.g. [6, 8‐10, 13‐16, 19, 27]). Pain/discomfort was the most frequently reported problem, but in the eldest elderly problems with the mobility domain occurred most frequently, similar to findings in a Dutch study of Hoeymans et al. [9]. Anxiety/depression was reported least frequent like, in other studies [8‐10]. Overall, our large sample of respondents appeared to report slightly fewer problems in most domains of EQ-5D compared to previous studies conducted in Dutch elderly [6‐8]. The relative largest differences were found for anxiety/depression. According to Szende et al. [8] 9.9% and 12.5% of the 65–74 years and ≥75 years old reported to be moderately or extremely anxious or depressed, and according to Hoeymans et al. [9] this was 11.8% (65–79 years) and 13.6% (≥80 years). In our population only 6% reported to be moderately or extremely anxious or depressed, similar to König et al. [10]. Comparing the mean EQ-5D-index scores found in other Dutch studies, our estimates were slightly higher [6‐8] which may be explained from the fact that our sample consisted of a relatively fit and healthy population participating in a clinical trial, with more males and higher education level compared to the general population.

Social and cultural background differences mostly found between countries [10, 13, 15, 27] could be confirmed to exist also within a small country such as the Netherlands. In particular, respondents living in Southern Netherlands tend to have a slightly lower health status than those living in other regions. This is a consistent finding in many studies on within country differences in health status [28]. The different cultural and social background and history of the Southern part of the country seems to have an impact up to today.

Strength of our study is the large study population of elderly persons (i.e. ≥65 years), allowing stratification by gender and five age-groups. A further strength of the current study is the data availability of prevalent chronic comorbidities and/or a previous cardio-vascular event.

One of the limitations of our study was that we had a sample with a higher percentage of men than all other studies and furthermore our population had a higher education level than the general population. By definition, immunocompromised persons were excluded, as this was an exclusion criterion in the CAPiTA-trial [21, 22]. Our study population was therefore probably fitter and healthier than the general Dutch elderly population.

Within community-dwelling elderly large differences in health status exist. Impairment increases rapidly with age, but health status is also associated with socio-demographic variables and comorbidities.