Introduction

European population is undergoing an unprecedented ageing process, which is taking place as a joint effect of increased life expectancy and reduced fertility: the percentage of persons aged 60+ increased from 9.2% in 1990 to 11.7% in 2013 and is projected to reach 21.1% by 2050 [1]. Increased life expectancy leads to increased prevalence of non-communicable diseases and to a relevant increase on the burden associated to these conditions: in fact, the increase in years lived with disability in the last two decades was 55.4% for non-communicable diseases, 7.6% for communicable disease and 0.3% for injuries [2].

In such a context, healthy ageing is becoming an important pillar of research and an objective for policy-makers. Healthy ageing is defined as “the process of developing and maintaining the functional ability that enables well-being in older age” [3]. Healthy ageing is expected to impact on quality of life (QoL): thus, understanding the features of QoL, and of its determinants, in a healthy ageing perspective is of primary relevance. Studies addressing determinants of QoL generally focus on a limited number of domains, such as the presence of multi-morbidities [4,5], visual impairment [6] and obesity [7], behavioral issues, such as higher levels of alcohol use [8,9], smoking [10,11] or active lifestyle [9,12]. In addition to this, social factors have also been shown to influence QoL in the ageing process: examples of this include social and family relationships [13,14] and socioeconomic status [15,16].

Research addressing the impact of a larger number of factors on QoL is, however, almost lacking. In particular, little research has been devoted to the analysis of the impact of built environment (i.e. the part of the environment which results as an effect of human activity) on QoL [17], while some indications of its impact on the health state have been found: living in walkable environments are associated with increased physical activity, lower prevalence of overweight, lower depression and less reported alcohol abuse [18]; better-quality built environment is associated with lower prevalence of psychiatric symptoms [19]; worse-quality built environment is associated with poorer self-rated health [20]. The aim of this paper is to comprehensively identify the determinants of QoL in a large population study sample of persons aged 18–50 and 50+ that were enrolled on occasion of the COURAGE in Europe Project (Collaborative Research on Ageing in Europe). We included a wide set of candidate determinants, such as demographic data, chronic conditions, health and health habits information, as well as social networks and built environment variables. We relied on the definition of QoL endorsed by the WHO’s WHOQOL group: QoL as the 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 [21].

Methods

Results

The final sample with complete information across all variables, described in Table 1, comprised 5639 participants, mean age 46.3 (SD 18.4). Most of respondents were from urban contexts and the sample was balanced for gender. Responders from Poland reported lower QoL scores. Please refer to S1 Table for the complete distribution of variables by country.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Distribution of variables by country.

https://doi.org/10.1371/journal.pone.0159293.t001

Table 2 reports the results of the hierarchical regression analysis. The final model accounted for 45% of the variation (49.4% in Finland, 47.7% in Poland and 40.2% in Spain) and the most relevant contribution, in terms of explained QoL variation, was given by sociodemographic data (Model 1: 17.9% explained QoL variation), chronic conditions (Model 3: 4.6% additional explained QoL variation) and SNI (Model 6: 4.6% additional explained QoL variation). Please refer to S2 Table for the full hierarchical regression models to predict quality of life.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Hierarchical regression models to predict quality of life.

https://doi.org/10.1371/journal.pone.0159293.t002

The following predictors had the most significant direct association with higher QoL scores: living in Finland contrasted to living in Poland, having a good social network and living in a house that is perceived as usable and with low risk of accidents. The following predictors had the most significant inverse association with higher QoL scores: age, lower education levels (primary or secondary) contrasted to higher education, being an active smoker contrasted to being a long-life never smoker, being depressed, self-reporting pain determining moderate to extreme difficulties compared to no pain, moderate to extreme learning difficulties compared to no difficulties, feelings of sadness of any extent, self-reporting to be moderately emotionally affected by one’s own health condition contrasted to not being affected and, finally, self-reporting severe/extreme distant vision problems.

Some variables were basically consistently associated to QoL in the three countries: these include variables with an inverse association with QoL such as smoking status, bodily aches and pain, being emotionally affected by health problems, as well as variables directly associated, such as good social network and living in a house that is perceived as usable and with low risk of accidents. Other variables, on the contrary, had a specific role across countries: age was significantly associated to QoL only in Poland, alcohol consumption only in Spain, angina was significantly associated only in Finland and depression only in Spain, self-reported feelings of sadness were significantly associated both in Finland and Poland, but not in Spain.

Discussion

With this study, we reported on the determinants of QoL using data referred to 5639 persons from Finland, Poland and Spain. Our results show that the most important factors were sociodemographic variables (age, education level and living in Finland), negative health habits (smoking status and physical inactivity), presence of chronic conditions, particularly depression, presence of disabling pain, of learning difficulties and of visual problems, a wide and rich social network, and living in usable house that is perceived as non-risky.

To our knowledge, this is one of the few studies addressing a wide range of determinants of QoL from a population study. Thompson and colleagues [32] addressed the effect of several variables over composite measures of physical and mental QoL in U.S. adults aged 50 and over. They found that medical care costs, smoking and leisure-time physical activity (i.e. practicing physical activities in addition to work-related ones) had large effects on both mental and physical aspects of QoL, particularly among those with functional limitations. Layte and colleagues [33] addressed the determinants of QoL in an Irish sample of subjects aged 50 and over: they found that mental health, social participation and physical health explain more than 60% of the variation of QoL, with mental health issues having the strongest independent power. The results of our study are in line with those herein described, but expand their value by adding some data, such as education and the characteristics of respondents’ homes, by the fact that we controlled for a large number of variables and by the fact that we enrolled subjects from 18 years on.

Our results can be broadly divided in two areas: the determinants of QoL that are not, or little, likely to be modified if not in the long run, and those that might be modified with short-medium term interventions. In the first group, variables such as age and education level–that however might change among the younger respondents–and visual problems are included; the second group is composed of health habits such as smoking, alcohol consumption, physical inactivity, presence of pain as well as by less frequently addressed variables like social network and the features of the built environment.

The relation between individual QoL and age is controversial. In some studies, age was found among the predictors of decreased QoL, most likely as a result of the effect of chronic conditions in older individuals [5,34], while such a relationship was not confirmed in other studies [4,35]. Our results show that age determined worse QoL when controlling for a wide set of variables, and we did not find a relevant impact of comorbidities like in previous studies: it is to be noted that only angina and depression remained significant in our final model. Severe distant vision problems had a strong negative impact on QoL, which basically confirmed previous research findings [36,37]. Similarly to what shown in other studies, we also found that distant vision, and not near vision, impact on QoL [38] and that positive relationships have a mitigating role [39]; on the contrary, we did not find any effect of hearing impairment that was previously shown to impact on QoL [40].

Education has been sometimes used as a way to address socioeconomic status (SES) inequalities connected to health and QoL: our findings show that lower educational levels are associated to diminished QoL compared to having high school or academic degrees. Similar results were found among participants to the American Health and Retirement Study and to the English Longitudinal Study of Ageing [41], as well as among community-dwelling elderly individuals (aged 65+) from Switzerland [42]. As previously reported [16] people with low SES seem to face a double burden: first, increased levels of health impairments and, second, lower QoL once health is impaired. Our data show that lower education levels, presence of health state problems–specifically connected to pain and emotional difficulties–and presence of chronic conditions such as angina and depression determine lower QoL. Angina and depressive symptoms were already found to be associated with and impaired QoL [43,44], and pain is frequently reported in association with depression or low mood: a recent study showed that these two symptoms have a distinct impact over reduced QoL [45], while other hypothesized that they show a bidirectional interaction and are correlated [46,47]: our results seem to be consistent with the hypothesis of a distinct impact of pain and depressive symptoms on reduced QoL. Our data did not address the nature of pain, which can therefore have any origin and can be associated to prevalent conditions, e.g. musculoskeletal disease like neck or back pain, or neurological conditions such as headache disorders. Presence of musculoskeletal pain was in fact associated to lower QoL, with higher pain severity and multiple locations being associated to worse QoL [48–50]. Pain from headache disorders were shown to influence QoL through both pain intensity and frequency of headaches, with more frequent headaches determining worse QoL [51,52].

As reported in a recent review, a negative relationship between smoking and QoL exists and the magnitude of association is consistent with the amount of cigarettes smoked [53]. Smoking cessation was found to significantly improve QoL, which was found to be better between 2 and 5 years after quitting [11].Our data seem to support these findings: contrasted to never smoking, smoking status determined worse QoL, while no predictive power was found for previous smoking. However, we did not include information on the duration of non-smoking period, therefore this has to be taken just as an hypothesis. Smoking has relevant implications, not only connected to QoL, but also to general health and preventive policies. Compared to non-smokers, smokers were more likely to have problems with mobility, self-care and daily activities, and were also more likely to report pain, anxiety and depression [11]. Smoking is a risk factor that tends to present with other factors, such as alcohol consumption and physical inactivity [54,55], which is partly consistent with our results. Low levels of physical activity were in fact found to be associated to reduced QoL in a large number of studies [56,57], as well as to obesity or overweight [58,59]. In contrast with these results, our data did not support the association between obesity, increased WC and QoL, despite a more than half of our sample was obese or overweight and one-third had high-risk WC. The issue of alcohol consumption deserves a separate comment, since a moderate alcohol consumption was associated to better QoL compared to being abstainer: this results is consistent with what previously found in other studies [60,61].

Social network and social support have been shown to have a positive impact on health-related quality of life [62,63]. The difficulty with the measurement of this construct lies in the multiplicity of sub-concepts that are part of it and that include social network size, density (i.e. the extent to which the members are connected to each other), homogeneity (i.e. how much individuals are similar to each other in a network) and boundedness (i.e. the degree to which they are defined on the basis of traditional group structures such as kin, work and neighbourhood) [64] and, generally, research focused on only one of these aspects. The COURAGE in Europe SNI, on the contrary, allows for the determination of social networks’ support on the basis of the structural component of the social network of an individual: each component of the network (spouse, parents, other relatives, neighbours, friends and co-workers) provides some kind of support (emotional, instrumental, appraisal and informational), and the frequency of contacts finally provides information on intensity of contacts and on the level of involvement [30]. Therefore, the weighted SNI used in our analysis, by addressing all of the relevant components of social network, enabled us to provide reliable information on the positive effect of a wide and rich social network towards higher QoL.

Finally, our results are completely novel with regard to the positive impact on QoL of living in a house that is perceived as usable and non-risky. Research addressing the impact of built environment features on QoL is almost lacking and, to our knowledge, two previous studies addressed such a relationship. The first refers to a sample of citizens of Bogotá (Colombia) aged 60 or more, in which perceived features of built environment, namely street noise, safety of public areas and street crossings were found to be significantly associated with QoL [65]. In a second study from Italy, the features of built environment were restricted to living in an urban or rural context only, and significant group differences were found only among male respondents [66]. Such a way to address features of built environment is largely partial, as it provides no indication on the degree to which the context in which the persons live has any facilitating or hindering effect. On the contrary our approach, in which different scales measuring perceived environment characteristics were used, enabled us to open to a broader perspective on the association between environmental factors and quality of life: the findings of our study suggest that policies that promote and ensure in-house safety could influence QoL, also controlling for several determinants.

Taken as a whole, our results pointed out similarities and differences in the impact of determinants in the three countries. Some variables–namely smoking status, presence of bodily pain, being emotionally affected by health conditions, as well as having good social network and living in a house that is perceived as usable and with low risk of accidents–had an impact that was basically consistent across countries. For other variables, some differences in their impact across countries were pointed out, in particular connected to age, chronic conditions such as angina and depression, and self-reported feeling of sadness.

Previous studies suggested that age is a predictor of decreased QoL as a result of the effect of chronic conditions [5,34], while in other studies this association was not found [4,35], and our results are in between: if we consider the whole group, the association was confirmed. However, if we take single countries, we see a relevant association for age in Polish population, but not for example among the Spanish, consistently with the results of Garin and colleagues [4]. The same is to be said for angina and depression: they were already found to be associated with and impaired QoL [43,44], but this association was mostly derived from studies performed in clinical populations. In our study, depression was significantly associated to worse QoL only among the Spanish subgroup, thus confirming results previously addressed in other studies where presence of depression was associated to reduced QoL in Spanish populations [67,68]. What it is interesting to note is that among the Polish and Finnish subgroup, feelings of sadness were associated to worse QoL, but not the diagnosis of depression. This remarks the difference between the two variables, and opens to understand the reasons for this trend. We can hypothesize that differences exist in the organization of services in the three countries or that there is higher perception of stigma: a report from Poland seems to point out this element as a relevant one at work, within families and general community members [69].

Some limitations need to be acknowledged. First, the cross-sectional design, that does not allow the investigation of causal relationships. Second, health conditions were self-reported, which could lead to identification bias. Such an approach has been shown to be generally adequate when matched to physicians’ statements on the presence of a disease [70]: anyway, to reduce the risk of reporting biases, generic reference to broad illness group was used, and therefore no definition of disease severity could be applied. Third, data on health state description and on difficulties in daily activities were referred to the previous thirty days, but no indication on the duration of the problem (i.e. how many days in the last thirty?) was included: this might have exerted some effect on variables such as presence of pain, sleep problems or concentration. Finally, we recommend a cautious approach when interpreting the comparison across countries since they should be mostly regarded as a descriptive one. The reason for this is that the three samples are representative of the population of each country, but the way in which respondents were selected, and weights created, was country-dependent: for this reason, no differences across countries were calculated.

In conclusion, we reported information on the determinants of QoL in a sample of community dwelling adults and older adults from Finland, Poland and Spain. Our results shed light on the importance of identifying modifiable risk factors that act as determinants of QoL, and provide public health indications that could support concrete actions at country level. In particular, smoking cessation, increasing the level of physical activity, identifying and addressing pain-related problems, improving social network ties and applying universal design approach to houses and environmental infrastructures could potentially increase QoL of ageing population.

Supporting Information