Background
Behaviours such as unhealthy diet, smoking, excessive alcohol intake, and physical inactivity are a major public health issue due to their association with higher risks of morbidity and mortality [
1‐
4]. A strong socioeconomic gradient exists in health-risk behaviour [
5,
6], and they are markedly more prevalent among residents of deprived neighbourhoods than among those of non-deprived neighbourhoods [
7,
8]. Large differences have been found above and beyond personal characteristics such as sex, age, ethnic background, cohabitation status, educational level, and employment status [
7], but the underlying factors that may explain the independent association between neighbourhood deprivation and health-risk behaviours remain poorly understood [
7‐
9]. Studies have suggested that stress could explain the link between neighbourhood deprivation and health-risk behaviours [
9‐
12], thus acting as an aggravating factor that increases health-risk behaviour among residents of deprived neighbourhoods. Residents of deprived neighbourhoods have an increased risk of perceived stress compared to residents with similar sociodemographic characteristics and socioeconomic status (SES) in non-deprived neighbourhoods [
9,
13]. Living in deprived neighbourhoods may lead to increased stress through factors such as overcrowding, high crime rates, perceived danger, poor transportation, poor housing, disrepair, limited services, poor infrastructure, and a lack of social support [
13‐
18]. Additionally, neighbourhood deprivation is an independent source of stress, over and above individual SES and other factors [
17]. Stress is associated with health-risk behaviour [
19‐
27], possibly in large part because people often cope with feelings of stress through risky, but often pleasurable, behaviours [
10,
20,
21,
28‐
33] such as eating high-fat foods, smoking, and drinking alcohol [
10]. Furthermore, the motivation for physical activity may be limited when experiencing stress [
10].
There is only limited research into whether perceived stress can influence health-risk behaviour among residents of deprived neighbourhoods. A study of women in deprived neighbourhoods in Australia found cross-sectional and longitudinal associations between perceived stress and both reduced leisure-time physical activity and more frequent fast food consumption [
34]. Further research into perceived stress and health-related behaviours in deprived neighbourhoods is warranted, including comparisons with the general population [
18]. There appear to be large socioeconomic differences in health-risk behaviours among residents in deprived neighbourhoods [
7], and it seems plausible that perceived stress can increase socioeconomic differences in health-risk behaviour. Whether perceived stress has an aggravating effect on the associations between SES and health-risk behaviour in deprived neighbourhood remains uncertain, as no published studies have examined this. A better understanding of health-risk behaviours in the context of stress and coping may be useful for health promotion interventions aimed at reducing health-risk behaviour in deprived neighbourhoods and may reduce social inequality in health in general.
The aim of this study was threefold: First, to compare the prevalence of perceived stress among residents in deprived neighbourhood with that in the general population. Second, to determine whether perceived stress was associated with health-risk behaviour (including their co-occurrence) among residents of deprived neighbourhoods. Third, to examine whether the association between SES and health-risk behaviour among residents of deprived neighbourhoods was modified by perceived stress.
Discussion
This study examined associations between perceived stress, SES, and health-risk behaviours among residents of deprived neighbourhoods, and whether perceived stress modified the associations between SES and health-risk behaviours. Consistent with prior research, we found a significantly higher risk of perceived stress among residents of deprived neighbourhoods compared with the general population [
9,
13]. The study found that perceived stress in deprived neighbourhoods was significantly associated with low intake of fruit or vegetables, daily smoking, physical inactivity, and the co-occurrence of health-risk behaviours. However, perceived stress was not associated with high-risk alcohol intake. Additionally, perceived stress modified the associations between SES and physical inactivity and between SES and having two or more health-risk behaviours in the deprived neighbourhoods. Perceived stress was strongly associated with physical inactivity and having two or more health-risk behaviours among residents with medium/high SES compared to residents with low SES.
The findings of this study support similar studies of associations between stress and health-risk behaviour in other populations [
19‐
27]. The association between stress and low fruit or vegetable intake has also been found by prior research [
23,
43]. In a study by Mouchacca et al., stress was found to predict higher fast food consumption among women from deprived neighbourhoods in Australia [
34]. In addition, a systematic review reported less healthy eating patterns among people in lower social positions who had higher stress levels [
24]. Unhealthier eating habits of people with higher levels of stress may be due to unhealthy foods being perceived as a ‘comfort’ or ‘reward’ in coping with stress [
44].
Our findings are in line with several studies showing that high levels of perceived stress are associated with increased smoking levels, smoking initiation, and a reduced likelihood of quitting smoking [
19‐
21,
25‐
27,
45]. Among others, Kaplan et al. found that smokers with high levels of perceived stress tended to smoke more and were less confident that they would be able to quit smoking [
45]. There is strong evidence that cigarette smoking can be a coping mechanism that provides a respite from stressful physical environments such as overcrowding, low-quality housing, traffic, and neighbourhood noise [
46,
47].
The absence of an association between perceived stress and high-risk alcohol intake in deprived neighbourhoods is supported by Ng and Jeffery’s study, in which no association between high perceived stress levels and the level of alcohol intake was found [
20]. Other studies have, however, reported a positive association between stress and high-risk alcohol intake [
25‐
27,
48]. These inconsistent results could be due to the different methods used to measure stress and alcohol intake or differences in the studied populations. It is worth mentioning that the correlation between socioeconomic factors and alcohol intake in Denmark is very modest when compared to other Western countries [
49]. In Denmark, high-risk alcohol intake is more prevalent in younger people with basic school as their highest level of education, whereas the most risk-prone in older age groups are those with the highest education [
50]. The Danish alcohol culture can also be characterised as relatively ‘wet’, where drinking is a fundamental factor of social life for many in the general population [
51]. However, this is not the case in deprived neighbourhoods [
7].
The association between perceived stress and physical inactivity among residents of deprived neighbourhoods was unsurprising, as previous studies have reported that stress is associated with reduced physical activity and increased sedentary behaviours [
19,
20,
26,
27]. However, Steptoe et al. found no differences in exercise frequency or duration with changes in perceived stress [
48]. Reduced participation in physical activity can be caused by stress due to difficult life circumstances and lack of coping, which may go beyond self-care and health-promoting behaviours such as physical activity [
26]. In addition, many people respond to stress by engaging in less physical activity because they may consider sedentary physical activity rewarding in the short term, despite evidence that physical activity can reduce stress over time [
20,
26,
52,
53].
Our finding on the association between perceived stress and the co-occurrence of health-risk behaviours is in line with Fine et al., who reported that persons with high mental distress are twice as likely to have three or four health-risk behaviours than none or one health-risk behaviour [
54]. Our findings highlight that deprived neighbourhood residents with perceived stress constitute a risk group for the co-occurrence of health-risk behaviours.
In the research on the association between stress and health-risk behaviour, there is often the perception that health-risk behaviour can be a tool to cope with stress [
10,
20,
21,
28‐
33] because behaviours such as smoking and eating high-fat food give immediate pleasure [
20,
33]. However, research cannot establish with certainty that health-risk behaviours are used as coping tools for stress [
30]. Our results showing higher odds for health-risk behaviour among residents with perceived stress offer indirect support for health-risk behaviour as a coping tool. In Kaplan et al.’s focus group study of residents in low-income communities, the participants described a direct causal pathway between stress and poor health as well as an indirect pathway through health behaviours, including uncontrolled eating, smoking, and physical inactivity [
45]. The study participants articulated various theories about the links between stress and health-risk behaviour, such as self-medication, adaptive behaviour, discounting the future, depletion of willpower, and competing priorities [
45].
Our study also found that disposable income, economic deprivation, and strain had strong associations with perceived stress, which may partly explain why residents of deprived neighbourhoods have a higher risk of perceived stress than the general population. Wilkinson has suggested that the poor suffer doubly from deprivation: besides the direct material effects, it also affects their health through psychosocial channels [
55]. Being at the bottom of society’s hierarchy may lead to stress from feelings of bitterness based on invidious social comparisons [
56], and the perception of social inequality can be an incentive for health-risk behaviour [
28]. In the same way, living in a neighbourhood which is deprived not only in absolute terms, but also relative to nearby neighbourhoods and to society in general, can induce feelings of exclusion and stigmatisation, and residents in deprived neighbourhoods may resort to health-risk behaviours to cope with these perceptions [
57].
The findings of this study have important implications for public health practice. Health promotion interventions would benefit from incorporating stress reduction strategies to address health-risk behaviour in deprived neighbourhoods. We recommend interventions to help residents of deprived neighbourhoods to cope with stress without resorting to health-risk behaviour—especially in regard to fruit and vegetable consumption, smoking, and physical activity. These interventions should help residents to acknowledge the short-term nature of the rewards of health-risk behaviours and to find other more effective and less harmful ways of coping with stress. Previous research has found that activities such as physical activity, relaxation techniques, talking to others, or making time for social activities were effective in managing stress [
30,
58]. Lipschitz et al. found that persons with poor stress management had more health-risk behaviours than those with effective stress management [
59]. The study also found a relationship between improved stress management over six months and decreasing health-risk behaviour [
58,
59]. Furthermore, as mental health has shown to be associated with health behaviours, we recommend the adoption of general mental health promotion initiatives in deprived neighbourhoods [
60]. Practitioners should assess perceived stress and refer their patients to stress reduction facilities when appropriate. Knowledge about perceived stress among residents of deprived neighbourhoods can be used to identify vulnerable groups who need special attention in health promotion interventions aimed at deprived neighbourhoods.
The idea that perceived stress caused by neighbourhood deprivation can increase the risk of health-risk behaviour arises mainly from cross-sectional studies associating neighbourhood deprivation with perceived stress [
10,
13]. The results of this study provide a base for future longitudinal studies to examine causal associations between perceived stress and health-risk behaviour in deprived neighbourhoods.
Future research should also examine pathways between perceived stress and health-risk behaviour in deprived neighbourhoods to gain a deeper understanding of the mechanisms through which perceived stress causes health-risk behaviour. This should include identifying of modifiers and mediators of this association, which could provide important target areas for future interventions. The influence of perceived stress on health-risk behaviours may depend, for example, on the availability of buffering resources and stress-modifying factors such as social support or personality characteristics [
61,
62]. In general, socioeconomically deprived groups have fewer buffering resources, making them more prone to stress and therefore more likely to cope via health-risk behaviour than higher socioeconomic groups [
61‐
63]. Further research into interpersonal and intrapersonal coping resources for stress in deprived neighbourhoods would be useful, e.g. on social networks that can provide emotional support.
Strengths and weaknesses of the study
One of the main strengths of this study is that it is based on a large sample of residents living in 12 deprived neighbourhoods with a response rate of 63%. This is noteworthy, as residents of deprived neighbourhoods are less likely to participate in health research [
64,
65] and tend to be underrepresented in health profile surveys [
19,
66]. It is the first study on perceived stress and health-risk behaviour carried out in a Danish context. Further, the study is unique in simultaneously examining four central health-risk behaviours (low intake of fruit and vegetable, smoking, high-risk alcohol intake, and physical inactivity) as well as the co-occurrence of health-risk behaviours.
Our study has some limitations. Non-response in DNHPS could affect the results regarding the level of perceived stress; if the most stressed people tend not to participate, the level of perceived stress may be underestimated. The respondents are considered to be representative of the residents of deprived neighbourhoods in Denmark, however, so we do not consider non-response to be a serious bias in relation to the observed associations. The cross-sectional data used for this study precluded the assessment of the directions of causations in the relationships between perceived stress and health-risk behaviour. The results should therefore be interpreted in the light of existing theories and research [
27]. The association between perceived stress and health-risk behaviour could also operate in the reverse direction to that examined in the present study. As smoking, drinking, and physical inactivity can increase stress indicators, health-risk behaviour may also cause or worsen mental health problems [
32]. A further limitation of our study is the measurement of perceived stress using an index based on four items. This might have certain drawbacks compared with Cohen’s 10-item PSS, which has been demonstrated to be a valid and reliable instrument on its own [
38]. However, we were unable to include further items in the perceived stress index due to limitations in the length of the questionnaire used in the DNHPS
. The four questions chosen were found to provide the best measure of perceived stress in deprived neighbourhoods. Another limitation is the use of two slightly different formulations of the second question in the measurement of perceived stress in the DNHPS and the DHMS. However, we do not consider that this has any major impact on the results. Furthermore, we do not know if some of the participants in the DNHPS also participated in the DHMS, as we do not have access to information about the respondents’ personal identification number in the DNHPS [
7]. Different modes of data collection were used for the DNHPS (primarily telephone interviews) and the DHMS (paper or web-based self-administered questionnaires). In a narrative review by Bowling (2005), face-to-face and telephone interviews generally had a higher risk of social desirability bias than self-administered questionnaires [
67]. This could cause social desirability response bias in regard to level of perceived stress and health-risk behaviour.
Further, our study was based on self-reported data, which may have led to information bias. People tend to overestimate their physical activity, which can be an indicator of social desirability bias [
68].