TY - JOUR
AU - Lasgaard, Mathias
AB - Abstract Background Research suggests that loneliness and social isolation (SI) are serious public health concerns. However, our knowledge of the associations of loneliness and SI with specific chronic diseases is limited. Purpose The present prospective cohort study investigated (a) the longitudinal associations of loneliness and SI with four chronic diseases (cardiovascular disease [CVD], chronic obstructive pulmonary disease [COPD], diabetes mellitus Type 2 [T2D], and cancer), (b) the synergistic association of loneliness and SI with chronic disease, and (c) baseline psychological and behavioral explanatory factors. Methods Self-reported data from the 2013 Danish “How are you?” survey (N = 24,687) were combined with individual-level data from the National Danish Patient Registry on diagnoses in a 5 year follow-up period (2013–2018). Results Cox proportional hazard regression analyses showed that loneliness and SI were independently associated with CVD (loneliness: adjusted hazard ratio (AHR) = 1.20, 95% confidence interval [CI; 1.03, 1.40]; SI: AHR = 1.23, 95% CI [1.04, 146]) and T2D (loneliness: AHR =1.90, 95% CI [1.42, 2.55]; SI: AHR = 1.59, 95% CI [1.15, 2.21]). No significant associations were found between loneliness or SI and COPD and cancer, respectively. Likewise, loneliness and SI did not demonstrate a synergistic effect on chronic disease. Multiple mediation analysis indicated that loneliness and SI had an indirect effect on CVD and T2D through both baseline psychological and behavioral factors. Conclusion Loneliness and SI were independently associated with a diagnosis of CVD and T2D within a 5 year follow-up period. The associations of loneliness and SI with CVD and T2D were fully explained by baseline psychological and behavioral factors. Loneliness, Social isolation, CVD, Diabetes, COPD, Cancer In recent years, researchers and political stakeholders have described loneliness and social isolation (SI) as serious public health concerns due to growing evidence of their association with negative health outcomes [1–3]. However, our knowledge of disease-specific outcomes and the factors contributing to these associations remain limited. Moreover, most previous studies have explored the impact of loneliness and SI separately, while there is evidence to suggest that psychosocial factors cluster together to predict chronic disease outcomes [4, 5]. The present study filled those gaps, exploring (a) loneliness and SI as predictors of four chronic diseases, (b) the synergistic association of loneliness and SI with chronic disease, and (c) the psychological and behavioral explanatory factors that link loneliness and SI with chronic disease. Loneliness and Social Isolation: Two Distinct Constructs Despite being related, loneliness and SI are distinct constructs. Loneliness is a subjective, unpleasant emotional state resulting from a perceived discrepancy between desired and achieved levels of social contact [6]. Loneliness occurs when a person feels that he or she has fewer social contacts than desired (quantity) or because he or she feels that current social relationships lack quality characteristics (i.e., intimacy, openness, and confidentiality) [6]. SI, in contrast, reflects the objective characteristics of an individual’s social situation and refers to absence of social contacts and social relationships [7]. Studies have demonstrated that loneliness and SI are moderately correlated [8, 9], but nevertheless distinct constructs. Moreover, studies suggest interplay between loneliness and SI: SI may act as a precursor to loneliness, for instance, in the case of loss [10, 11], but SI may also be the consequence of long-term loneliness [12] because people withdraw from social relationships to avoid perceived threat [13]. Hence, the association between loneliness and SI seems to be reciprocal. Thus, although the terms have different meanings, they are interconnected and the experience of either is potentially detrimental. Loneliness and Social Isolation as Risk Factors for Negative Health Outcomes Meta-analyses have shown that loneliness and SI predict the onset of cardiovascular disease (CVD) [14] and increase the risk of premature death [1]. However, less is known about the associations of loneliness and SI with other somatic diseases [3, 14, 15], including cancer and diabetes. A few cross-sectional studies have found associations between loneliness and lung disease [16, 17], as well as between loneliness and SI and diabetes [17–19]. Poor social relationships (i.e., low level of social capital and SI) have also been linked with lung, breast, and colon cancer [20], whereas loneliness has been linked with cancer-related gene expressions in postmortem brain tissue [21, 22]. However, cross-sectional associations do not imply causation, and disease-specific outcomes of loneliness and SI remain largely unknown. Meta-analyses exploring the associations of loneliness and SI, respectively, with CVD and mortality have found them to be equally strong, with increased risk for both CVD [14] and premature death [1]. However, in the meta-analyses, few studies (e.g., [23–26]) examined both loneliness and SI. Hence, it is difficult to elucidate their separate effects in relation to outcomes and whether the presence of both risk factors increase risk as compared to the presence of only one risk factor. To further complicate matters, findings are mixed with results suggesting that loneliness is a stronger predictor of mortality than SI [23] but also that SI is a stronger predictor than loneliness [24, 27, 28]. Furthermore, a recent study on the effects of loneliness and SI on CVD found that only loneliness predicted incident CVD [29]. Another recent study found the associations of loneliness and SI with acute myocardial infarction (AMI) and stroke to be overall similar, even though SI was the only predictor of mortality following AMI and stroke [30]. Taken together, our knowledge of the associations of loneliness and SI with specific chronic diseases is limited. Moreover, knowledge of potential differences and synergistic effects with respect to the associations of loneliness and SI with poor health is missing. As such, our knowledge of the associated risk of each with health largely exists in parallel literature bases because loneliness and SI are rarely included in the same studies as predictors of health outcomes [31]. Only a few large-scale studies have investigated loneliness and SI in mutually adjusted analysis [23, 29, 30] and very few studies have investigated potential synergistic effects [30, 32]. Hence, our knowledge of the potential synergistic effect of loneliness and SI on chronic disease outcomes is sparse, leaving a gap in this area of research [1, 32, 33]. A Generic Model of Explanatory Factors Several explanatory factors underlying the associations of loneliness and SI with negative health outcomes have been suggested. Summarizing theoretical proposals, three generic pathways—a psychological, a behavioral, and a biological—can be identified [34], with several explanatory factors contributing to each of these pathways (Fig. 1). Due to the scope of the present study, the biological pathway will not be discussed further. The psychological pathway is generally viewed as an assembly of different psychological factors [34], and studies have found that factors such as perceived stress [18, 35, 36] and depression [27, 36] explain part of the association between loneliness and negative health outcomes. The behavioral pathway consists of factors related to lifestyle—dietary habits, alcohol intake, and sleep patterns [12, 34]—and studies have also found these factors to explain part of the associations [18, 35]. Fig. 1. Open in new tabDownload slide A conceptual model of explanatory factors. Fig. 1. Open in new tabDownload slide A conceptual model of explanatory factors. It has been proposed that maladaptive social hypervigilance is of special importance to these pathways. As such, hypervigilance has been proposed to produce elevated stress and diminish self-regulatory capacities hindering self-promoting health behaviors (e.g., physical activity and healthy dieting) in lonely individuals [12, 37]. Moreover, loneliness might cause the brain to remain somewhat vigilant during sleep, resulting in poor sleep patterns that ultimately affect the body’s restorative processes [38]. Others have suggested that provision of social support, social influence, social engagement, and attachment, as well as access to resources and material goods might constitute important mechanisms, which influence the psychological and behavioral pathways between social relations and poor health [39, 40]. With respect to potential confounders and moderators of this association, studies have indicated that sociodemographic factors affect the association of loneliness and SI with poor health [2, 18, 27, 41]. Notwithstanding these theoretical proposals and empirical studies, a gap remains in our knowledge of the explanatory factors underlying the associations of loneliness and SI with negative health outcomes both in general and with respect to specific diseases. Moreover, with respect to potential differences in the indirect pathways of loneliness and SI, our knowledge is sparse. However, a previous study has indicated that some differences do exist [27]. Further investigation is warranted [33]: better understanding of potential explanatory factors could help facilitate the design of appropriate and effective interventions to prevent or reduce excess health risk in lonely or isolated people [33]. Aims The present study aimed to (a) test parts of the generic model (Fig. 1) by examining loneliness and SI as independent predictors of CVD, diabetes mellitus Type 2 (T2D), chronic obstructive pulmonary disease (COPD), and cancer and to (b) clarify potential synergistic effects of loneliness and SI with chronic disease. Lastly, (c) the study aimed to examine the extent to which baseline psychological factors (e.g., perceived stress and negative affect) and behavioral factors (e.g., daily smoking, high alcohol consumption, unhealthy diet, physical inactivity, insufficient sleep duration, and obesity) explain the associations of loneliness and SI with chronic disease. Because of the generic nature of the model, we did not develop disease-specific hypotheses. Due to the scarcity of research on this topic and the contradictory findings, we did not develop clear expectations regarding potential differences in the associations of loneliness and SI with chronic disease. However, we hypothesized that both loneliness and SI increase the risk of chronic disease when examined simultaneously. Moreover, we hypothesized that various factors may help explain the associations of loneliness and SI with chronic disease. Methods The present prospective cohort study used data from the 2013 Danish Health and morbidity survey “How are you?” and data from the National Danish Patient Registry. The “How are you?” survey is conducted every 4 years in Denmark. The present study is based on 2013 data from 24,687 individuals, 35–79 years of age, residing in the Central Denmark Region (as per January 1, 2013) [42]. The study comprises data from the Central Denmark Region, one of the five Danish administrative regions, which is home to approximately 23 % of the Danish population of 5.7 million inhabitants. The study population’s demographic composition (sex, age, and civil status) was similar to that of the Danish population [43]. The sample was identified through the Civil Registration System using the unique personal identification number (CPR number) assigned to Danish citizens at birth. Since all individuals living in Denmark have a unique personal identification number, both respondents and nonrespondents in the survey can be linked on an individual level to different central registers. The respondents received the questionnaire (in paper/electronically) in January 2013. Three reminders were sent to increase the response rate, which was 66%. Using the CPR number, we linked respondents from the survey with individual level longitudinal register data from three and a half years preceding the survey (January 1, 2010–May 31, 2013) and 5 years after the survey was conducted (June 1, 2013–December 31, 2018). In Denmark, data are collected for administrative and scientific purposes in various registers. In the present study, data from The Danish National Patient Register [44] on diagnoses were merged with survey data. Moreover, data from the Central Person Register were used to exclude cases (respondents who died prior to May 31, 2013 [n = 37]. Calibrated weights were applied to enhance the representativeness of the study by accounting for differences in selection probabilities and response rates. The weights were constructed by Statistics Denmark using a model-based calibration approach [45] and are based on register data on sex, age, municipality of residence, highest completed educational level, income, marital status, ethnic background, number of visits to the general practitioner, hospitalization, occupational status, owner/tenant status, and protection from inquiries during statistical and scientific surveys for all individuals living in Denmark [46]. Predictor Variables (Survey Data) Loneliness was assessed using the Three-Item Loneliness Scale [T-ILS] [47]. The T-ILS contains the following questions: How often do you feel isolated from others? How often do you feel you lack companionship? How often do you feel left out? The sum of the items (ranging from 3 to 9) provides a global measure of loneliness, with higher scores indicating greater loneliness. The T-ILS has good psychometric properties, including good internal consistency (α = .76 in this sample), a high concurrent, and discriminant validity and correlates strongly with the UCLA loneliness scale [47]. In order to compare the estimates of loneliness and SI, the total score was dichotomized as also done in other large-scale studies [27, 29] with a score above 5 reflecting loneliness [2, 48]. The sum score was used in sensitivity analyses. SI was measured using a simple index based on questions regarding social contact. The index was inspired by the detailed Social Network Index created by Berkman and Syme [49]. The following five indicators were included: (a) living alone (yes = 1 and no = 0); (b) less than monthly contact with family with whom one does not live (yes = 1 and no = 0); (c) less than monthly contact with friends (yes = 1 and no = 0); (d) less than monthly contact with colleagues/fellow students outside the workplace or school (yes = 1 and no = 0); and (e) less than monthly contact with neighbors or the local community (yes = 1 and no = 0). Using the five indicators, we generated a sum score ranging from 0 to 5, with higher scores indicating greater SI. A score between 3 and 5, corresponding to a maximum of two areas of social interfaces, was treated as an indicator of SI. Several larger population-based studies have used a similar approach [24, 27]. The sum score was used in sensitivity analyses. Explanatory Factors (Survey Data) Psychological factors Perceived stress was measured using the 10-item Perceived Stress Scale [50]. Each item is rated on a five-point Likert scale with higher scores reflecting greater stress (0–40). Negative affect is defined as feelings of emotional distress, especially by the common variance of factors, such as anxiety, sadness, and unpleasant emotions [51], and was assessed using the two items Have you been troubled by low spirit, depressive feelings, and unhappiness (in the past 2 weeks)? and Have you been troubled by anxiety, nervousness, and unease (in the past 2 weeks)?. Each item is rated on a three-point Likert scale (yes, very troubled; yes, a little troubled; and no) with higher scores indicating greater negative affect (2–6). The crude measure demonstrated good internal consistency (α = .79). Behavioral factors Daily smoking: If respondents agreed indicated that they smoked daily, they were classified as daily smokers. High alcohol consumption: Respondents who stated to drink more than 14 units per week for men and more than 7 units per week for women were classified as high risk alcohol consumers [52]. Unhealthy diet: Using the Diet Quality Score [53], respondents having an unhealthy diet characterized by a low amount of fruit, vegetables, fish, and a high amount of saturated fat were identified. Physical inactivity: Respondents were classified as physically inactive if they were physically active of moderate to high intensity no more than 30 min weekly. Insufficient sleep duration: Respondents who stated to be having less than 6 hr of sleep or more than 9 hr of sleep per weekday was classified as having insufficient sleep duration [54, 55]. Obesity: respondents who had a body mass index ≥30 were classified as obese [56]. Sociodemographic factors (survey and register data) Three demographic variables were included: gender (register data), age (register data), and educational attainment (survey data). Using the Danish version of the International Standard Classification of Education, we classified educational level as low (0–10 years), medium (11–14 years), and high (≥15 years). Outcome Variables (Register Data 2010–2018) All data on diagnoses were obtained from the Danish National Patient Register using International Statistical Classification of Diseases and Related Health Problems 10th Revision (appear in brackets below). Register data from the Danish National Patient Register from the 3 year period prior to the survey were used to identify respondents diagnosed with one of the four chronic diseases prior to 2013. Respondents with a diagnosis within the four groups of chronic diseases in the 3 year period before 2013 were excluded from the respective analyses where they were used as outcome measures. CVD was defined as ischemic heart diseases (DI20–DI25), heart failure (DI150), peripheral artery occlusive disease (DI170–DI174), and stroke (DI160–DI164). Those conditions are considered the most common CVDs in the population [57]. T2D. Diabetes mellitus type 2 (DE11). COPD. Chronic obstructive pulmonary disease (DJ44). Cancer was defined as cancers in the digestive organs like the esophagus, pancreas, and liver (DC15–DC26); cancers in the respiratory system like the lungs, thymus, and trachea (DC30–DC39); breast cancer (DC50); cancers in the kidney and urinary system (DC64–DC68); and cancers in the mouth and oral cavity (DC00–DC14). Thus, only cancers related to health behaviors were included [58]. Statistical Data Analyses The significance level was set at p < .05. Adjusted hazard ratio (AHR) and odds ratio (OR) are presented with 95% confidence intervals (CIs). The statistical analyses were performed using STATA, version 15. Step 1. Cox proportional hazard regression analysis The STATA command phtest was used to assess Schoenfeld residuals, and no violations were observed (p > .05). The tested model with all variables (predictors, explanatory factors, and outcomes) is presented as a figure in Supplementary Material 1. Using Cox proportional hazard regression analysis, partially AHRs were calculated for the association between the two risk factors (i.e., loneliness and SI) and each specific chronic disease (i.e., CVD, T2D, COPD, and cancer). In order to investigate synergistic effects, adjusted Cox proportional hazard regression analyses were conducted in which loneliness, SI, and the interaction between loneliness and SI were entered into the same model. Loneliness and SI were included as predictors in all models. Moreover, models were adjusted for gender, age, and educational attainment. Lastly, using Cox proportional hazard regression analysis, fully AHRs were calculated for the association between the two risk factors (i.e., loneliness and SI) and each specific chronic disease (i.e., CVD, T2D, COPD, and cancer) when all sociodemographic and explanatory factors were added to the model. Step 2. Multiple mediation analysis The STATA command “khb,” which allows for the inclusion of a combination of different types of variables (dichotomous and continuous), was used to conduct multiple mediation analyses [59] based on a logistic regression model. The command decomposes the total effect (Path C) into the direct effect (Path c; the effect of the independent variable on the dependent variable when controlling for mediating variables; i.e., explanatory factors) and the indirect effect (Path a and Path b; the effect of the independent variable on the dependent variable through mediating variables). For an illustration, see Supplementary Material 1. Moreover, the analysis calculates the extent to which each mediator contributes to the indirect and total effect. Multiple mediation analyses were conducted to determine the indirect effect of loneliness and SI through the explanatory factors on each disease for which a significant association was detected in Step 1. Each explanatory factor was tested individually a priori. All factors that demonstrated a significant indirect effect were included in the multiple mediation models. Models were adjusted for gender, age, and educational attainment. Moreover, each model was adjusted for loneliness and SI, respectively, and total follow-up time in days (0–2,040) as the khb method does not handle survival data. Sensitivity analysis Using the sum score of T-ILS and the SI index as predictors of loneliness and SI, sensitivity analysis was conducted in order to examine whether the results of the main analyses remained the same. Moreover, the effect of gender was examined by adding the interaction between loneliness and gender and the interaction between SI and gender as predictors of chronic disease in two separate analyses. The models were adjusted for age, educational attainment, and loneliness and SI, respectively. Ethics The present study was approved by the Danish Data Protection Agency (record number 1-16-02-888-17). Information about the survey was provided to potential participants in an invitation letter emphasizing that participation was voluntary. The participants’ voluntary completion of the survey questionnaires constituted implied consent. According to Danish law, no formal ethical approval of survey and register-based studies is required from an ethics committee or other research oversight. Results Table 1 provides descriptive statistics of the study population at baseline. In total, 3,566 participants (18%) were lonely at baseline, whereas 2,339 (11%) participants were characterized as socially isolated. Seven hundred ninety-four participants (5% of the full sample) were both lonely and socially isolated. Thus, 26% of participants who were lonely at baseline also reported being socially isolated, and 40 % of the participants who were socially isolated also reported loneliness. The correlation between loneliness and SI was moderate (rt = .39). In total, 1,651 (6% of the full sample) of the study participants were diagnosed with CVD in the 5 year follow-up period, whereas 460 (1.8% of the full sample) participants were diagnosed with T2D. Only 370 (1.5% of the full sample) participants were diagnosed with COPD in the 5 year follow-up period, whereas 852 (3.2% of the full sample) participants were diagnosed with cancer. A matrix of the zero-order correlations of all included predictors and explanatory factors is provided in Supplementary Material 2. Table 1. Descriptive statistics of the study population at baseline . All (N = 24,687) . . . . n . % . M (SD) . Gender  Men 11,901 49.9  Woman 12,786 50.1 Age – – 54.65 (12.00) Educational attainment  Low 4,606 18.87  Medium 12,842 52.81  High 6,588 28.31 Perceived stress – – 10.95 (7.01) Negative affect – – 2.6 (1.01) Daily smoking 4,237 18.65 High alcohol consumption 9,581 6.88 Unhealthy diet 2,681 11.94 Physical inactivity 4,159 18.22 Insufficient sleep duration 3,714 16.30 Obesity 4,044 16.69 Lonely 3,566 17.68 3.68 (1.19) Socially isolated 2,339 11.44 2.21 (1.05) Lonely and socially isolated 749 4.49 Chronic disease (3 years prior to survey)  CDV 1,448 5.5  T2D 626 2.53  COPD 326 1.30  Cancer 677 2.55 . All (N = 24,687) . . . . n . % . M (SD) . Gender  Men 11,901 49.9  Woman 12,786 50.1 Age – – 54.65 (12.00) Educational attainment  Low 4,606 18.87  Medium 12,842 52.81  High 6,588 28.31 Perceived stress – – 10.95 (7.01) Negative affect – – 2.6 (1.01) Daily smoking 4,237 18.65 High alcohol consumption 9,581 6.88 Unhealthy diet 2,681 11.94 Physical inactivity 4,159 18.22 Insufficient sleep duration 3,714 16.30 Obesity 4,044 16.69 Lonely 3,566 17.68 3.68 (1.19) Socially isolated 2,339 11.44 2.21 (1.05) Lonely and socially isolated 749 4.49 Chronic disease (3 years prior to survey)  CDV 1,448 5.5  T2D 626 2.53  COPD 326 1.30  Cancer 677 2.55 All percentages are weighted based on register data to represent the population of the Central Denmark Region, 2013. COPD chronic obstructive pulmonary disease; CDV cardiovascular disease; SD standard deviation; T2D diabetes mellitus Type 2. Open in new tab Table 1. Descriptive statistics of the study population at baseline . All (N = 24,687) . . . . n . % . M (SD) . Gender  Men 11,901 49.9  Woman 12,786 50.1 Age – – 54.65 (12.00) Educational attainment  Low 4,606 18.87  Medium 12,842 52.81  High 6,588 28.31 Perceived stress – – 10.95 (7.01) Negative affect – – 2.6 (1.01) Daily smoking 4,237 18.65 High alcohol consumption 9,581 6.88 Unhealthy diet 2,681 11.94 Physical inactivity 4,159 18.22 Insufficient sleep duration 3,714 16.30 Obesity 4,044 16.69 Lonely 3,566 17.68 3.68 (1.19) Socially isolated 2,339 11.44 2.21 (1.05) Lonely and socially isolated 749 4.49 Chronic disease (3 years prior to survey)  CDV 1,448 5.5  T2D 626 2.53  COPD 326 1.30  Cancer 677 2.55 . All (N = 24,687) . . . . n . % . M (SD) . Gender  Men 11,901 49.9  Woman 12,786 50.1 Age – – 54.65 (12.00) Educational attainment  Low 4,606 18.87  Medium 12,842 52.81  High 6,588 28.31 Perceived stress – – 10.95 (7.01) Negative affect – – 2.6 (1.01) Daily smoking 4,237 18.65 High alcohol consumption 9,581 6.88 Unhealthy diet 2,681 11.94 Physical inactivity 4,159 18.22 Insufficient sleep duration 3,714 16.30 Obesity 4,044 16.69 Lonely 3,566 17.68 3.68 (1.19) Socially isolated 2,339 11.44 2.21 (1.05) Lonely and socially isolated 749 4.49 Chronic disease (3 years prior to survey)  CDV 1,448 5.5  T2D 626 2.53  COPD 326 1.30  Cancer 677 2.55 All percentages are weighted based on register data to represent the population of the Central Denmark Region, 2013. COPD chronic obstructive pulmonary disease; CDV cardiovascular disease; SD standard deviation; T2D diabetes mellitus Type 2. Open in new tab Step 1: Regression Analyses Loneliness (AHR = 1.30, 95% CI [1.10, 1.54] and SI (AHR = 1.24, 95% CI [1.03, 1.50]) at baseline independently increased the risk of being diagnosed with CVD by 30% and 24%, respectively (Table 2). Subsequent analysis demonstrated that the interaction between loneliness and SI was not significant (AHR = 1.20, 95% CI [0.80, 1.80]; Table 3). Both loneliness (AHR = 1.98, 95% CI [1.47, 2.67]) and SI (AHR = 1.56, 95% CI [1.11, 2.16]) independently increased the risk of T2D by 98% and 56%, respectively. Subsequent analysis demonstrated that the interaction between loneliness and SI was not significant (AHR = 0.72, 95% CI [0.37, 1.39]; Table 3). Neither loneliness nor SI predicted being diagnosed with COPD. However, both loneliness and SI increased the risk of COPD with an AHR of 1.26 (loneliness: 95% CI [0.90, 1.77]; SI: 95% CI [0.84, 1.88]; Table 3). Neither loneliness nor SI predicted being diagnosed with cancer independently. Table 2. Loneliness and social isolation as predictors of chronic disease in partially adjusted Cox proportional hazard regression analysis . CVDa . . T2Db . . COPDc . . Cancerd . . . HRe (95% CI) . p . HRe (95% CI) . p . HRe (95% CI) . p . HRe (95% CI) . p . Loneliness 1.30 (1.10–1.54) .002 1.98 (1.47–2.67) .000 1.26 (0.90–1.77) .181 1.13 (0.89–1.44) .321 Social isolation 1.24 (1.03–1.50) .021 1.56 (1.11–2.16) .009 1.26 (0.84–1.88) .264 1.01 (0.76–1.34) .933 . CVDa . . T2Db . . COPDc . . Cancerd . . . HRe (95% CI) . p . HRe (95% CI) . p . HRe (95% CI) . p . HRe (95% CI) . p . Loneliness 1.30 (1.10–1.54) .002 1.98 (1.47–2.67) .000 1.26 (0.90–1.77) .181 1.13 (0.89–1.44) .321 Social isolation 1.24 (1.03–1.50) .021 1.56 (1.11–2.16) .009 1.26 (0.84–1.88) .264 1.01 (0.76–1.34) .933 All estimates are weighted based on register data to represent the population of the Central Denmark Region, 2013. CDV cardiovascular disease; CI confidence interval; COPD chronic obstructive pulmonary disease; HR hazard ratio; SD standard deviation; T2D diabetes mellitus Type 2. aRespondents with CVD 3 years prior to the survey in 2013 were excluded from the analysis. bRespondents with T2D 3 years prior to the survey in 2013 were excluded from the analysis. cRespondents with COPD 3 years prior to the survey in 2013 were excluded from the analysis. dRespondents with cancer 3 years prior to the survey in 2013 were excluded from the analysis. eAdjusted for gender, age, and educational attainment. Open in new tab Table 2. Loneliness and social isolation as predictors of chronic disease in partially adjusted Cox proportional hazard regression analysis . CVDa . . T2Db . . COPDc . . Cancerd . . . HRe (95% CI) . p . HRe (95% CI) . p . HRe (95% CI) . p . HRe (95% CI) . p . Loneliness 1.30 (1.10–1.54) .002 1.98 (1.47–2.67) .000 1.26 (0.90–1.77) .181 1.13 (0.89–1.44) .321 Social isolation 1.24 (1.03–1.50) .021 1.56 (1.11–2.16) .009 1.26 (0.84–1.88) .264 1.01 (0.76–1.34) .933 . CVDa . . T2Db . . COPDc . . Cancerd . . . HRe (95% CI) . p . HRe (95% CI) . p . HRe (95% CI) . p . HRe (95% CI) . p . Loneliness 1.30 (1.10–1.54) .002 1.98 (1.47–2.67) .000 1.26 (0.90–1.77) .181 1.13 (0.89–1.44) .321 Social isolation 1.24 (1.03–1.50) .021 1.56 (1.11–2.16) .009 1.26 (0.84–1.88) .264 1.01 (0.76–1.34) .933 All estimates are weighted based on register data to represent the population of the Central Denmark Region, 2013. CDV cardiovascular disease; CI confidence interval; COPD chronic obstructive pulmonary disease; HR hazard ratio; SD standard deviation; T2D diabetes mellitus Type 2. aRespondents with CVD 3 years prior to the survey in 2013 were excluded from the analysis. bRespondents with T2D 3 years prior to the survey in 2013 were excluded from the analysis. cRespondents with COPD 3 years prior to the survey in 2013 were excluded from the analysis. dRespondents with cancer 3 years prior to the survey in 2013 were excluded from the analysis. eAdjusted for gender, age, and educational attainment. Open in new tab Table 3. The synergistic effect of loneliness and social isolation as a predictor of chronic disease in partially adjusted Cox proportional hazard regression analysis . CVDa . . T2Db . . COPDc . . Cancerd . . . HRe (95% CI) . P . HRe (95% CI) . P . HRe (95% CI) . p . HRe (95% CI) . p . Neither loneliness nor social isolation (reference group) – – – – – – – – Only loneliness 1.24 (1.02–1.51) .029 2.15 (1.54–3.00) <.001 1.26 (0.86–1.84) .227 1.18 (0.91–1.54) .207 Only social isolation 1.16 (0.93–1.46) .189 1.79 (1.21–2.65) .004 1.26 (0.78–2.05) .348 1.09 (0.79–1.49) .596 Loneliness × Social isolation 1.20 (0.80–1.80) .369 0.72 (0.37–1.39) .329 0.99 (0.42–2.34) .982 0.79 (0.41–1.52) .485 . CVDa . . T2Db . . COPDc . . Cancerd . . . HRe (95% CI) . P . HRe (95% CI) . P . HRe (95% CI) . p . HRe (95% CI) . p . Neither loneliness nor social isolation (reference group) – – – – – – – – Only loneliness 1.24 (1.02–1.51) .029 2.15 (1.54–3.00) <.001 1.26 (0.86–1.84) .227 1.18 (0.91–1.54) .207 Only social isolation 1.16 (0.93–1.46) .189 1.79 (1.21–2.65) .004 1.26 (0.78–2.05) .348 1.09 (0.79–1.49) .596 Loneliness × Social isolation 1.20 (0.80–1.80) .369 0.72 (0.37–1.39) .329 0.99 (0.42–2.34) .982 0.79 (0.41–1.52) .485 All estimates are weighted based on register data to represent the population of the Central Denmark Region, 2013. CDV cardiovascular disease; CI confidence interval; COPD chronic obstructive pulmonary disease; HR hazard ratio; SD standard deviation; T2D diabetes mellitus Type 2. aRespondents with CVD 3 years prior to the survey in 2013 were excluded from the analysis. bRespondents with T2D 3 years prior to the survey in 2013 were excluded from the analysis. cRespondents with COPD 3 years prior to the survey in 2013 were excluded from the analysis. dRespondents with cancer 3 years prior to the survey in 2013 were excluded from the analysis. eAdjusted for gender, age, and educational attainment. Open in new tab Table 3. The synergistic effect of loneliness and social isolation as a predictor of chronic disease in partially adjusted Cox proportional hazard regression analysis . CVDa . . T2Db . . COPDc . . Cancerd . . . HRe (95% CI) . P . HRe (95% CI) . P . HRe (95% CI) . p . HRe (95% CI) . p . Neither loneliness nor social isolation (reference group) – – – – – – – – Only loneliness 1.24 (1.02–1.51) .029 2.15 (1.54–3.00) <.001 1.26 (0.86–1.84) .227 1.18 (0.91–1.54) .207 Only social isolation 1.16 (0.93–1.46) .189 1.79 (1.21–2.65) .004 1.26 (0.78–2.05) .348 1.09 (0.79–1.49) .596 Loneliness × Social isolation 1.20 (0.80–1.80) .369 0.72 (0.37–1.39) .329 0.99 (0.42–2.34) .982 0.79 (0.41–1.52) .485 . CVDa . . T2Db . . COPDc . . Cancerd . . . HRe (95% CI) . P . HRe (95% CI) . P . HRe (95% CI) . p . HRe (95% CI) . p . Neither loneliness nor social isolation (reference group) – – – – – – – – Only loneliness 1.24 (1.02–1.51) .029 2.15 (1.54–3.00) <.001 1.26 (0.86–1.84) .227 1.18 (0.91–1.54) .207 Only social isolation 1.16 (0.93–1.46) .189 1.79 (1.21–2.65) .004 1.26 (0.78–2.05) .348 1.09 (0.79–1.49) .596 Loneliness × Social isolation 1.20 (0.80–1.80) .369 0.72 (0.37–1.39) .329 0.99 (0.42–2.34) .982 0.79 (0.41–1.52) .485 All estimates are weighted based on register data to represent the population of the Central Denmark Region, 2013. CDV cardiovascular disease; CI confidence interval; COPD chronic obstructive pulmonary disease; HR hazard ratio; SD standard deviation; T2D diabetes mellitus Type 2. aRespondents with CVD 3 years prior to the survey in 2013 were excluded from the analysis. bRespondents with T2D 3 years prior to the survey in 2013 were excluded from the analysis. cRespondents with COPD 3 years prior to the survey in 2013 were excluded from the analysis. dRespondents with cancer 3 years prior to the survey in 2013 were excluded from the analysis. eAdjusted for gender, age, and educational attainment. Open in new tab Fully adjusted Cox proportional hazard regression analysis demonstrated that neither loneliness nor SI predicted being diagnosed with CVD, T2D, COPD, or cancer when all sociodemographic and explanatory factors were added to the model (see Supplementary Material 3). Step 2: Multiple Mediation Analysis Neither loneliness nor SI had a direct effect on CVD when all explanatory factors were included (see Supplementary Material 4). When testing each explanatory factor individually, loneliness demonstrated a significant indirect effect on CVD through perceived stress, negative affect, daily smoking, physical inactivity, insufficient sleep duration, and obesity; SI demonstrated an indirect effect through perceived stress, negative affect, daily smoking, physical inactivity, and insufficient sleep duration (not in table). When testing the full multiple mediation model, loneliness had a significant indirect effect on CVD through perceived stress, daily smoking, insufficient sleep duration, and obesity, independent of gender, age, educational attainment, and SI (Table 4). Similarly, when testing the multiple mediation model, SI had a significant indirect effect on CVD through perceived stress, daily smoking, and insufficient sleep duration, independent of gender, age, educational attainment, and loneliness (Table 5). Perceived stress (43%) explained most of the variance in the association between loneliness and CVD (Table 4); perceived stress (11%) and daily smoking (17%) explained the most variance in the association between SI and CVD (Table 5). Table 4. The individual contribution of each explanatory factor on the association between loneliness and chronic disease . CVDa . . . . Loneliness . . . . Indirect% . Total% . p . Perceived stress 59.0 43.0 .014 Negative affect 2.5 1.9 .913 Daily smoking 20.9 16 <.001 Physical inactivity 4.4 3.4 .239 Insufficient sleep duration 10.4 8.0 .048 Obesity 5.7 4.4 .004 T2Db Loneliness Indirect% Total% p Perceived stress 21.3 12.0 .335 Negative affect 45.7 25.9 .035 Physical inactivity 2.3 1.3 .538 Insufficient sleep duration 14.1 8.0 .003 Obesity 16.7 9.5 <.001 . CVDa . . . . Loneliness . . . . Indirect% . Total% . p . Perceived stress 59.0 43.0 .014 Negative affect 2.5 1.9 .913 Daily smoking 20.9 16 <.001 Physical inactivity 4.4 3.4 .239 Insufficient sleep duration 10.4 8.0 .048 Obesity 5.7 4.4 .004 T2Db Loneliness Indirect% Total% p Perceived stress 21.3 12.0 .335 Negative affect 45.7 25.9 .035 Physical inactivity 2.3 1.3 .538 Insufficient sleep duration 14.1 8.0 .003 Obesity 16.7 9.5 <.001 All estimates are weighted based on register data to represent the population of the Central Denmark Region, 2013. All estimates are adjusted for gender, age, educational attainment, and social isolation. Indirect% = contribution (%) to the indirect effect. Total% = contribution (%) to the total effect. CDV cardiovascular disease; T2D diabetes mellitus Type 2. aRespondents with CVD 3 years prior to the survey in 2013 was excluded from the analysis. bRespondents with T2D 3 years prior to the survey in 2013 was excluded from the analysis. Open in new tab Table 4. The individual contribution of each explanatory factor on the association between loneliness and chronic disease . CVDa . . . . Loneliness . . . . Indirect% . Total% . p . Perceived stress 59.0 43.0 .014 Negative affect 2.5 1.9 .913 Daily smoking 20.9 16 <.001 Physical inactivity 4.4 3.4 .239 Insufficient sleep duration 10.4 8.0 .048 Obesity 5.7 4.4 .004 T2Db Loneliness Indirect% Total% p Perceived stress 21.3 12.0 .335 Negative affect 45.7 25.9 .035 Physical inactivity 2.3 1.3 .538 Insufficient sleep duration 14.1 8.0 .003 Obesity 16.7 9.5 <.001 . CVDa . . . . Loneliness . . . . Indirect% . Total% . p . Perceived stress 59.0 43.0 .014 Negative affect 2.5 1.9 .913 Daily smoking 20.9 16 <.001 Physical inactivity 4.4 3.4 .239 Insufficient sleep duration 10.4 8.0 .048 Obesity 5.7 4.4 .004 T2Db Loneliness Indirect% Total% p Perceived stress 21.3 12.0 .335 Negative affect 45.7 25.9 .035 Physical inactivity 2.3 1.3 .538 Insufficient sleep duration 14.1 8.0 .003 Obesity 16.7 9.5 <.001 All estimates are weighted based on register data to represent the population of the Central Denmark Region, 2013. All estimates are adjusted for gender, age, educational attainment, and social isolation. Indirect% = contribution (%) to the indirect effect. Total% = contribution (%) to the total effect. CDV cardiovascular disease; T2D diabetes mellitus Type 2. aRespondents with CVD 3 years prior to the survey in 2013 was excluded from the analysis. bRespondents with T2D 3 years prior to the survey in 2013 was excluded from the analysis. Open in new tab Table 5. The individual contribution of each explanatory factor on the association between social isolation and chronic disease . CVDa . . . . Social isolation . . . . Indirect% . Total% . p . Perceived stress 28.1 10.6 .010 Negative affect −2.1 −0.8 .832 Daily smoking 44.9 16.9 <.001 Physical inactivity 13.9 5.2 .124 Insufficient sleep duration 15.2 5.7 .036 T2Db Social isolation Indirect% Total% p Perceived stress 25.4 9.6 .101 Negative affect 21.1 7.8 .085 Physical inactivity 24.1 9.1 .043 Insufficient sleep duration 29.3 11.0 .002 . CVDa . . . . Social isolation . . . . Indirect% . Total% . p . Perceived stress 28.1 10.6 .010 Negative affect −2.1 −0.8 .832 Daily smoking 44.9 16.9 <.001 Physical inactivity 13.9 5.2 .124 Insufficient sleep duration 15.2 5.7 .036 T2Db Social isolation Indirect% Total% p Perceived stress 25.4 9.6 .101 Negative affect 21.1 7.8 .085 Physical inactivity 24.1 9.1 .043 Insufficient sleep duration 29.3 11.0 .002 All estimates are weighted based on register data to represent the population of the Central Denmark Region, 2013. All estimates are adjusted for gender, age, educational attainment, and loneliness. Indirect% = contribution (%) to the indirect effect. Total% = contribution (%) to the total effect. CDV cardiovascular disease; T2D diabetes mellitus Type 2. aRespondents with CVD 3 years prior to the survey in 2013 was excluded from the analysis. bRespondents with T2D 3 years prior to the survey in 2013 was excluded from the analysis. Open in new tab Table 5. The individual contribution of each explanatory factor on the association between social isolation and chronic disease . CVDa . . . . Social isolation . . . . Indirect% . Total% . p . Perceived stress 28.1 10.6 .010 Negative affect −2.1 −0.8 .832 Daily smoking 44.9 16.9 <.001 Physical inactivity 13.9 5.2 .124 Insufficient sleep duration 15.2 5.7 .036 T2Db Social isolation Indirect% Total% p Perceived stress 25.4 9.6 .101 Negative affect 21.1 7.8 .085 Physical inactivity 24.1 9.1 .043 Insufficient sleep duration 29.3 11.0 .002 . CVDa . . . . Social isolation . . . . Indirect% . Total% . p . Perceived stress 28.1 10.6 .010 Negative affect −2.1 −0.8 .832 Daily smoking 44.9 16.9 <.001 Physical inactivity 13.9 5.2 .124 Insufficient sleep duration 15.2 5.7 .036 T2Db Social isolation Indirect% Total% p Perceived stress 25.4 9.6 .101 Negative affect 21.1 7.8 .085 Physical inactivity 24.1 9.1 .043 Insufficient sleep duration 29.3 11.0 .002 All estimates are weighted based on register data to represent the population of the Central Denmark Region, 2013. All estimates are adjusted for gender, age, educational attainment, and loneliness. Indirect% = contribution (%) to the indirect effect. Total% = contribution (%) to the total effect. CDV cardiovascular disease; T2D diabetes mellitus Type 2. aRespondents with CVD 3 years prior to the survey in 2013 was excluded from the analysis. bRespondents with T2D 3 years prior to the survey in 2013 was excluded from the analysis. Open in new tab Neither loneliness nor SI had a direct effect on T2D when all explanatory factors were included (see Supplementary Material 4). When testing each explanatory factor individually, loneliness demonstrated a significant indirect effect on T2D via perceived stress, negative affect, physical inactivity, insufficient sleep duration, and obesity; SI demonstrated an indirect effect on T2D through perceived stress, negative affect, physical inactivity, and insufficient sleep duration (not in table). When testing the full multiple mediation model, loneliness had a significant indirect effect on T2D through negative affect, insufficient sleep duration, and obesity, independent of gender, age, educational attainment, and SI (Table 4), whereas SI had a significant indirect effect on T2D through physical inactivity and insufficient sleep duration, independent of gender, age, educational attainment, and loneliness (Table 5). Negative affect (26%) explained most of the variance in the association between loneliness and T2D (Table 4), whereas physical inactivity and insufficient sleep duration explained 9% and 11%, respectively, of the association between SI and T2D (Table 5). Sensitivity Analyses Treating loneliness and SI as continuous measures indicated associations similar to the main findings. Loneliness independently predicted CVD (AHR = 1.09, 95% CI [1.04, 1.15]) and T2D (AHR = 1.25, 95% CI [1.14, 1.37]); the associations of SI with CVD (AHR = 1.05, 95% CI [0.99, 1.12]) and T2D (AHR = 1.10, 95% CI [0.99, 1.24]) were borderline significant. This was contrary to the results of the main analyses. Likewise, SI independently predicted COPD (AHR = 1.20, 95% CI [1.08, 1.33]). Neither loneliness nor SI predicted cancer (loneliness: AHR = 1.02, 95% CI [0.94, 1.11]; SI: AHR = 1.06, 95% CI [0.97, 1.15]); see Supplementary Material 5. No significant gender differences were found when investigating the interaction of loneliness and gender (AHR = 1.21, 95% CI [0.86, 1.68]) and SI and gender (AHR = 1.13, 95% CI [0.77, 1.65]) in the association with CVD (not in table). Likewise, we found no significant gender differences when investigating the interaction of loneliness and gender (AHR = 1.39, 95% CI [0.78, 2.43]) or the interaction of SI and gender (AHR= 0.85, 95% CI [0.46, 1.58]) in the association with T2D, COPD (Loneliness × Gender: AHR = 1.73, 95% CI [0.87, 3.43]). SI × Gender: AHR = 1.61, 95% CI [0.73, 3.53]), and cancer (Loneliness × Gender: AHR = 0.83, 95% CI [0.51, 1.35]). SI × Gender: AHR = 0.81, 95% CI [0.46, 1.42]; not in table). Discussion Our findings contribute to the growing body of literature on loneliness and SI as risk factors for poor health. When examined simultaneously, loneliness and SI were associated with being diagnosed with CVD and T2D within a 5 year follow-up period. However, it is important to note the bidirectional associations between loneliness, SI, and poor health [60, 61]. As such, loneliness and SI may also arise from long-term illness or due to functional limitations in the prestages of undetected disease. Health issues may lead to less social interaction [62] or the course of illness might exacerbate feelings of alienation and abandonment, leaving the individual prone to isolation or feelings of loneliness. The present study suggests that the co-occurrence of loneliness and SI does not have a synergistic effect on the associations with CVD and T2D. As such, the combined effect of loneliness and SI did not exceed the additive effect of the two occurring together. This finding is contrary to a recent study on mortality [32]. Nevertheless, the additive effect of co-occurring loneliness and SI might still be of importance. Individuals experiencing co-occurring loneliness and SI are likely to be at greater risk of chronic diseases compared to individuals experiencing only one of the two. The present study replicates other recent studies showing that loneliness and SI increase the risk of CVD [14, 29, 30], with similar effects for CVD incurred by loneliness and SI. Previous studies have reported mixed findings, with one study indicating a stronger association between loneliness and CVD [29], and another reporting similar risk for incident AMI and stroke, but an increased risk of mortality following AMI and stroke only for SI [30]. Moreover, the present study does not indicate gender differences, contrary to previous findings [18, 41], suggesting that gender might not moderate the associations of loneliness and SI with chronic disease. However, further research is needed to replicate these novel findings. The associations of loneliness and SI with CVD were fully explained by both psychological and behavioral factors, which are in line with other recent studies [30, 63]. As such, results indicated that perceived stress was especially important in explaining the association between loneliness and CVD. Moreover, perceived stress was also an important explanatory factor in the association between SI and CVD. As such, lonely and socially isolated individuals might cope less adaptively to stress, leaving them more prone to the pathogenic influence of stress [31]. This might particularly apply to lonely individuals as the present results indicate that perceived stress is a unique factor explaining more than 40% of the association between loneliness and CVD. Daily smoking was especially important in explaining the association between SI and CVD. Likewise, health behaviors also explained some of the variance in the association between loneliness and CVD. These findings are in line with theoretical approaches, indicating that loneliness and SI operate through multiple explanatory pathways, for instance, via psychological factors, health behaviors, and/or biological systems [12, 14, 34]. To the best of our knowledge, the present study is the first to demonstrate a longitudinal association of loneliness and SI with T2D. Previous studies have documented a cross-sectional association between loneliness and SI with diabetes [17–19]. Interestingly, the strength of the association of loneliness and SI with T2D is apparently greater than the associations with CVD [14, 29, 30] and mortality [1], which might indicate that loneliness and SI coexist with risk factors of special relevance to T2D etiology. The association of loneliness and SI with T2D was fully explained by psychological and behavioral factors. Thus, the considerable increased risk of T2D among lonely and isolated individual may largely be explained by poor mental health and poor health behaviors. Negative affect was the most important factor in explaining the association between loneliness and T2D. As such, the results suggest that factors like depression and anxiety might play a key role in the association with T2D, perhaps reflecting the diabetes risk associated with continuous use of antidepressants [64]. Health behaviors explained much of the variance in the association between SI and T2D. Lack of social contact might hinder the maintenance of health promoting behaviors; for instance, marital partners, family members, or friends might be likely to encourage each other in an effort to influence health habits [65]. As such, the link between health behaviors and chronic disease might be especially important in explaining the adverse health effects of SI. Our findings regarding the explanatory factors cannot establish directionality. Findings, therefore, solely illustrate the potentially harmful coexistence of mental health issues and health compromising behaviors with loneliness and SI, increasing the risk of chronic disease. In the present study, we found no support for an association between loneliness and SI with either COPD or cancer. However, loneliness and SI independently increased the risk of COPD noticeably by 26%. Power issues might explain why these findings were not statistically significant. Together, this could indicate that loneliness and SI are associated with an increased risk of those noncommunicable diseases that have a strong linkage with metabolic functions (e.g., blood pressure, cholesterol, and insulin production), such as CVD and T2D. Likewise, a previous study has demonstrated a strong cross-sectional association between loneliness and the metabolic syndrome, known as a clustering of interrelated biological factors that have been associated with an increased risk of diabetes and CVD [66]. To the best of our knowledge, no previous studies have used prospective data to investigate the association of loneliness and SI with COPD and cancer and, as suggested elsewhere, the absence of null findings could be related to selective publishing of positive results [14, 30]. Thus, future studies are warranted to further elucidate the relationship between loneliness and SI in relation to incident COPD and cancer. Especially, it would be preferable to investigate the association of loneliness and SI with specific cancers using larger population-based samples. Limitations Despite the use of high-quality data from a representative survey combined with detailed register data, the possibility of reversed causality bias should be considered, as well as issues regarding data completeness. As such, data on diagnosis and treatment within general practice are not possible to collect. Moreover, register data from 2010 to 2013 were used to identify respondents with chronic disease prior to the survey in 2013. However, the Danish National Patient Registry is contact based (e.g., inpatient, outpatient, and emergency service contact); for each patient contact, one primary and optional secondary diagnosis are recorded according to the International Classification of Diseases [44]. Thus, diagnoses recorded prior to 2010 are not necessarily included in the present data. However, since most patients with chronic diseases, such as CVD, T2D, COPD, and cancer, attend regular outpatient checkups or receive ongoing ambulant treatment, we expect that the large majority of respondents were included. The associations of loneliness and SI with poor health is bidirectional [60] and the potential role of unmeasured confounding resulting from other existing health problems, especially among older people, must not be ignored. Thus, we cannot rule out the issue of reversed causality. Another noteworthy limitation is the use of the khb method to investigate multiple mediation as this analysis was based on a logistic regression model without the use of censuring. The method is not compatible with survival data and, as described elsewhere [67], no method to analyze multiple mediation using survival data are available in STATA. Taken together, the results of the multiple mediation analysis should be interpreted with caution. Moreover, the use of cross-sectional data in the investigation of explanatory factors is also a limitation. Theory suggests the presence of some directionality with loneliness and SI leading to poor mental health and an unhealthy lifestyle, thereby affecting morbidity and mortality [12, 33]. However, in the present study, loneliness and SI were measured together with the explanatory factors, making it impossible to determine directionality. However, the results emphasize the important coexistence of adverse psychological and behavioral factors, which may help explain the associations of loneliness and SI with chronic disease. As in most survey research, nonresponse bias cannot be ruled out. However, the use of calibrated weights is likely to limit any such effects [46]. Also, it must be noted that negative affect was based on a purpose-designed measure, which has not previously been validated. The poor operationalization of that measure might affect the results presented. The correlation between perceived stress and negative affect was high. Standardized measures of depression and anxiety with greater discriminant validity would have been greatly preferred. Finally, the SI index was a composite measure counting the objective number of social contacts. Therefore, treating the index as a continuous variable reflecting a latent variable is not advisable. As such, we used a dichotomized version of the T-ILS and the SI index comparable to most recent research in this field [14, 24, 27, 30]. However, collapsing scale data (e.g., the T-ILS) has some limitations, such as reduced variance. Nonetheless, sensitivity analysis treating loneliness and SI as continuous measures demonstrated overall similar results to those using dichotomized measures. The similarity of results based on continuous and dichotomized measures has been proposed to reflect variations in risk across the entire spectrum of loneliness and SI rather than risk limited to individuals who are extremely lonely or socially isolated [24]. Summarizing, we acknowledge several limitations in the study design, including the unmeasured confounding variables that might determine causality. This, however, is an issue affecting most observational research, including other recent studies investigating the longitudinal association of loneliness and SI with morbidity and mortality [29, 30]. Implications for Theory and Practice Our findings have important theoretical implications. First, up until now, knowledge about the combined effects of loneliness and SI on the risk of chronic disease has been missing [32, 33]. The present study suggests that the co-occurrence of loneliness and SI does not result in a synergistic effect on chronic disease. Even so, a purely additive effect of loneliness and SI might still suggest that individuals exposed to both loneliness and SI constitute a specific vulnerable group, who would benefit from targeted prevention and intervention. However, future studies are needed to further investigate such matters. Researchers would benefit from including measures of both loneliness and SI in future studies, allowing for the continuing investigation of the complex relationship between the two. The present study also has implications for our understanding of explanatory factors that link loneliness and SI with chronic disease. Especially, findings indicate that loneliness and SI coexist with multiple adverse psychological and behavioral factors, which together explain the association of loneliness and SI with chronic disease. Pending further research and replication of the current preliminary findings, our results suggest that various psychological and behavioral factors explain the associations of loneliness and SI with chronic disease. Creating effective holistic prevention and intervention initiatives that focus on coexisting poor mental health and lifestyle factors could potentially help prevent chronic diseases among individuals suffering from loneliness and SI. Acknowledgments The health and morbidity survey “How are you?” was funded by the Central Denmark Region. Moreover, we are grateful for the funding provided by the Mary Foundation, the Aase and Ejnar Danielsen Foundation, DEFACTUM - Public Health and Health Service Research,—Central Denmark Region, and the University of Southern Denmark. Compliance with Ethical Standards Authors’ Statement of Conflict of Interest and Adherence to Ethical Standards The authors declare that they have no conflict of interest. Authors’ Contributions J.C., R.L., P.Q., C.H.M., S.S.P., and M.L. conceptualized the study. J.C. and M.L. devised the analyses plan. J.C. conducted the analyses and wrote the first draft of the manuscript in close collaboration with M.L.. J.C., R.L., P.Q., C.H.M., S.S.P., and M.L. contributed to data interpretation and reviewed, edited and approved the final manuscript. Ethical Approval The present study was approved by the Danish Data Protection Agency (record number 1-16-02-888-17). Information about the survey was provided to potential participants in an invitation letter emphasizing that participation was voluntary. According to Danish law, no formal ethical approval of survey and register-based studies is required from an ethics committee or other research oversight and no such institution exist. Informed Consent The participants’ voluntary completion of the survey questionnaires constituted implied consent. References 1. Holt-Lunstad J , Smith TB, Baker M, Harris T, Stephenson D. Loneliness and social isolation as risk factors for mortality: A meta-analytic review . Perspect Psychol Sci. 2015 ; 10 : 227 – 237 . Google Scholar Crossref Search ADS PubMed WorldCat 2. Lasgaard M , Friis K, Shevlin M. Where are all the lonely people? A population-based study of high-risk groups across the life span . Soc Psych Psych Epid. 2016 ; 51 ( 10 ): 1373 – 1384 . Google Scholar Crossref Search ADS WorldCat 3. Leigh-Hunt N , Bagguley D, Bash K, et al. An overview of systematic reviews on the public health consequences of social isolation and loneliness . Public Health. 2017 ; 152 : 157 – 171 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Pedersen SS , Denollet J, van Gestel YR, Serruys PW, van Domburg RT. Clustering of psychosocial risk factors enhances the risk of depressive symptoms 12-months post percutaneous coronary intervention . Eur J Cardiovasc Prev Rehabil. 2008 ; 15 : 203 – 209 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Pedersen SS , van den Broek KC, Erdman RA, Jordaens L, Theuns DA. Pre-implantation implantable cardioverter defibrillator concerns and Type D personality increase the risk of mortality in patients with an implantable cardioverter defibrillator . Europace. 2010 ; 12 : 1446 – 1452 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Peplau LA , Perlman D. Perspectives on loneliness . In: Peplau AL, Perlman D, eds. Loneliness: A Sourcebook of Current Theory, Research and Practice. New York, NY : John Wiley ; 1982 : 1 – 17 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 7. De Jong-Gierveld J , van Tilburg TG, Dykstra PA. Loneliness and social isolation . In: Perlman D, Vangelisti A, eds. The Cambridge Handbook of Personal Relationships. Cambridge, UK : Cambridge University Press ; 2006 : 485 – 500 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 8. Coyle CE , Dugan E. Social isolation, loneliness and health among older adults . J Aging Health. 2012 ; 24 : 1346 – 1363 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Perissinotto CM , Covinsky KE. Living alone, socially isolated or lonely—What are we measuring? J Gen Intern Med. 2014 ; 29 : 1429 – 1431 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Bennett R . The concept and measurement of social isolation . In: Bennett R, ed. Aging, Isolation and Resocialization. New York, NY : Van Nostrand Reinhold Co. ; 1980 : 9 – 26 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 11. Townsend P , Tunstall S. Sociological explanations of the lonely . In: Townsend P, ed. The Social Minority. London, UK : Allen Lane ; 1973 :257–263. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 12. Hawkley LC , Cacioppo JT. Loneliness matters: A theoretical and empirical review of consequences and mechanisms . Ann Behav Med. 2010 ; 40 : 218 – 227 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Qualter P , Vanhalst J, Harris R, et al. Loneliness across the life span . Perspect Psychol Sci. 2015 ; 10 : 250 – 264 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Valtorta NK , Kanaan M, Gilbody S, Ronzi S, Hanratty B. Loneliness and social isolation as risk factors for coronary heart disease and stroke: Systematic review and meta-analysis of longitudinal observational studies . Heart. 2016 ; 102 : 1009 – 1016 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Xia N , Li H. Loneliness, social isolation, and cardiovascular health . Antioxid Redox Signal. 2018 ; 28 : 837 – 851 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Penninx BW , van Tilburg T, Kriegsman DM, Boeke AJ, Deeg DJ, van Eijk JT. Social network, social support, and loneliness in older persons with different chronic diseases . J Aging Health. 1999 ; 11 : 151 – 168 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Petitte T , Mallow J, Barnes E, Petrone A, Barr T, Theeke L. A systematic review of loneliness and common chronic physical conditions in adults . Open Psychol J. 2015 ; 8 : 113 – 132 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Christiansen J , Larsen FB, Lasgaard M. Do stress, health behavior, and sleep mediate the association between loneliness and adverse health conditions among older people? Soc Sci Med. 2016 ; 152 : 80 – 86 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Hempler NF , Ekholm O, Willaing I. Differences in social relations between persons with type 2 diabetes and the general population . Scand J Public Health. 2013 ; 41 : 340 – 343 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Orlewska K , Sliwczynski A, Orlewska E. An ecological study of the link between the risk of most frequent types of cancer in Poland and socioeconomic variables . Int J Public Health. 2018 ; 63 : 777 – 786 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Canli T , Yu X, et al. Loneliness 5 years ante-mortem is associated with disease-related differential gene expression in postmortem dorsolateral prefrontal cortex . Transl Psychiatry. 2018 ; 8 (1) : 2 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Canli T , Yu L, Canli T, et al. Loneliness 5-year ante-mortem is associated with disease-related differential gene expression in postmortem dorsolateral prefrontal cortex . Transl Psychiatry. 2018 ; 10 ( 8 ): 1 . Google Scholar OpenURL Placeholder Text WorldCat 23. Holwerda TJ , Beekman AT, Deeg DJ, et al. Increased risk of mortality associated with social isolation in older men: Only when feeling lonely? Results from the Amsterdam Study of the Elderly (AMSTEL) . Psychol Med. 2012 ; 42 : 843 – 853 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Steptoe A , Shankar A, Demakakos P, Wardle J. Social isolation, loneliness, and all-cause mortality in older men and women . Proc Natl Acad Sci USA. 2013 ; 110 : 5797 – 5801 . Google Scholar Crossref Search ADS WorldCat 25. Strodl E , Kenardy J. The 5-item mental health index predicts the initial diagnosis of nonfatal stroke in older women . J Womens Health (Larchmt). 2008 ; 17 : 979 – 986 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Strodl E , Kenardy J, Aroney C. Perceived stress as a predictor of the self-reported new diagnosis of symptomatic CHD in older women . Int J Behav Med. 2003 ; 10 : 205 – 220 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Elovainio M , Hakulinen C, Pulkki-Råback L, et al. Contribution of risk factors to excess mortality in isolated and lonely individuals: An analysis of data from the UK Biobank cohort study . Lancet Public Health. 2017 ; 2 : e260 – e266 . Google Scholar Crossref Search ADS PubMed WorldCat 28. Tanskanen J , Anttila T. A prospective study of social isolation, loneliness, and mortality in Finland . Am J Public Health. 2016 ; 106 : 2042 – 2048 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Valtorta NK , Kanaan M, Gilbody S, Hanratty B. Loneliness, social isolation and risk of cardiovascular disease in the English Longitudinal Study of Ageing . Eur J Prev Cardiol. 2018 ; 25 : 1387 – 1396 . Google Scholar Crossref Search ADS PubMed WorldCat 30. Hakulinen C , Pulkki-Råback L, Virtanen M, Jokela M, Kivimäki M, Elovainio M. Social isolation and loneliness as risk factors for myocardial infarction, stroke and mortality: UK Biobank cohort study of 479 054 men and women . Heart. 2018 ; 104 : 1536 – 1542 . Google Scholar Crossref Search ADS PubMed WorldCat 31. National Academies of Sciences, Engineering, and Medicine . Social Isolation and Loneliness in Older Adults: Opportunities for the Health Care System . Washington, DC : The National Academies Press ; 2020 . Google Scholar PubMed OpenURL Placeholder Text Google Preview WorldCat COPAC 32. Beller J , Wagner A. Loneliness, social isolation, their synergistic interaction, and mortality . Health Psychol. 2018 ; 37 : 808 – 813 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Courtin E , Knapp M. Social isolation, loneliness, and health in old age: A scoping review . Health Soc Care Comm. 2017 ; 25 ( 3 ): 799 – 812 . Google Scholar Crossref Search ADS WorldCat 34. Holt-Lunstad J , Smith TB. Loneliness and social isolation as risk factors for CVD: Implications for evidence-based patient care and scientific inquiry . Heart. 2016 ; 102 : 987 – 989 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Segrin C , Passalacqua SA. Functions of loneliness, social support, health behaviors, and stress in association with poor health . Health Commun. 2010 ; 25 : 312 – 322 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Stickley A , Koyanagi A. Physical multimorbidity and loneliness: A population-based study . PLoS One. 2018 ; 13 : e0191651 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Hawkley LC , Burleson MH, Berntson GG, Cacioppo JT. Loneliness in everyday life: Cardiovascular activity, psychosocial context, and health behaviors . J Pers Soc Psychol. 2003 ; 85 : 105 – 120 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Cacioppo JT , Cacioppo S. Social relationships and health: The toxic effects of perceived social isolation . Soc Personal Psychol Compass. 2014 ; 8 : 58 – 72 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Berkman LF , Glass T, Brissette I, Seeman TE. From social integration to health: Durkheim in the new millennium . Soc Sci Med. 2000 ; 51 : 843 – 857 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Uchino BN . Social support and health: A review of physiological processes potentially underlying links to disease outcomes . J Behav Med. 2006 ; 29 : 377 – 387 . Google Scholar Crossref Search ADS PubMed WorldCat 41. Henriksen J , Larsen ER, Mattisson C, Andersson NW. Loneliness, health, and mortality . Epidemiol Psychiatr Sci. 2017 ; 30 : 1 – 6 . Google Scholar OpenURL Placeholder Text WorldCat 42. Larsen FB , Friis K, Lasgaard M, et al. Bind 1 Hvordan har du det? 2013. Sundhedsprofil for regionen og kommunerne [How are you? 2013. Public Health Survey of the Central Denmark Region]. Central Denmark Region, Denmark : Public Health and Quality Improvement ; 2014 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 43. Statistics Denmark. Population at the first day of the quarter by municipality, sex, age, marital status, ancestry, country of origin and citizenship . 2015 . Available at http://statistikbanken.dk/FOLK1. Accessibility verified February 25, 2015. 44. Schmidt M , Schmidt SA, Sandegaard JL, Ehrenstein V, Pedersen L, Sørensen HT. The Danish National Patient Registry: A review of content, data quality, and research potential . Clin Epidemiol. 2015 ; 7 : 449 – 490 . Google Scholar Crossref Search ADS PubMed WorldCat 45. Särndal CE , Lundstrøm S. Estimation in Surveys with Nonresponse. Hoboken, NJ : Wiley ; 2005 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 46. Christensen AI , Ekholm O, Glümer C, et al. The Danish National Health Survey 2010. Study design and respondent characteristics . Scand J Public Health. 2012 ; 40 : 391 – 397 . Google Scholar Crossref Search ADS PubMed WorldCat 47. Hughes ME , Waite LJ, Hawkley LC, Cacioppo JT. A short scale for measuring loneliness in large surveys: Results from two population-based studies . Res Aging. 2004 ; 26 : 655 – 672 . Google Scholar Crossref Search ADS PubMed WorldCat 48. Musich S , Wang SS, Hawkins K, Yeh CS. The impact of loneliness on quality of life and patient satisfaction among older, sicker adults . Gerontol Geriatr Med. 2015 ; 1 : 2333721415582119 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 49. Berkman LF , Syme SL. Social networks, host resistance, and mortality: A nine-year follow-up study of Alameda County residents . Am J Epidemiol. 1979 ; 109 : 186 – 204 . Google Scholar Crossref Search ADS PubMed WorldCat 50. Cohen S , Williamson G. Perceived stress in a probability sample of the U.S . In: Spacapam S, Oskamp S, eds. The Social Psychology of Health: Claremont Symposium on Applied Social Psychology. Newbury Park, CA : Sage ; 1988:31–67 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 51. Stringer DM . Negative affect . In: Gellman MD, Turner JR, eds. Encyclopedia of Behavioral Medicine. New York, NY : Springer New York ; 2013 : 1303 – 1304 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 52. The Danish Health Authority. Sundhedsstyrelsens nye udmelding vedrørende alkohol. [New Statement from the Danish Health Authority on Alcohol]. Copenhagen, Denmark : The Danish Health Authority ; 2010 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 53. Toft U , Kristoffersen LH, Lau C, Borch-Johnsen K, Jørgensen T. The Dietary Quality Score: Validation and association with cardiovascular risk factors: The Inter99 study . Eur J Clin Nutr. 2007 ; 61 : 270 – 278 . Google Scholar Crossref Search ADS PubMed WorldCat 54. Cappuccio FP , D’Elia L, Strazzullo P, Miller MA. Sleep duration and all-cause mortality: A systematic review and meta-analysis of prospective studies . Sleep. 2010 ; 33 : 585 – 592 . Google Scholar Crossref Search ADS PubMed WorldCat 55. Tsai TC , Wu JS, Yang YC, Huang YH, Lu FH, Chang CJ. Long sleep duration associated with a higher risk of increased arterial stiffness in males . Sleep. 2014 ; 37 : 1315 – 1320 . Google Scholar Crossref Search ADS PubMed WorldCat 56. World Health Organization. Obesity: Preventing and Managing the Global Epidemic. Report of a WHO Consultation . World Health Organization Technical Report Series . Vol. 894 . Geneva, Switzerland: WHO; 2000 : 1 – 253 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 57. Prior A. Psychological Stress and Multimorbidity in General Practice. Aarhus, Denmark : HEALTH, Aarhus University ; 2016 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 58. Katzke VA , Kaaks R, Kühn T. Lifestyle and cancer risk . Cancer J. 2015 ; 21 : 104 – 110 . Google Scholar Crossref Search ADS PubMed WorldCat 59. Kohler U , Karlson KB, Holm A. Comparing coefficients of nested nonlinear probability models . Stata J. 2011 ; 11 : 420 – 438 . Google Scholar Crossref Search ADS WorldCat 60. Luo Y , Hawkley LC, Waite LJ, Cacioppo JT. Loneliness, health, and mortality in old age: A national longitudinal study . Soc Sci Med. 2012 ; 74 : 907 – 914 . Google Scholar Crossref Search ADS PubMed WorldCat 61. Cohen-Mansfield J , Shmotkin D, Goldberg S. Loneliness in old age: Longitudinal changes and their determinants in an Israeli sample . Int Psychogeriatr. 2009 ; 21 : 1160 – 1170 . Google Scholar Crossref Search ADS PubMed WorldCat 62. Barlow MA , Liu SY, Wrosch C. Chronic illness and loneliness in older adulthood: The role of self-protective control strategies . Health Psychol. 2015 ; 34 : 870 – 879 . Google Scholar Crossref Search ADS PubMed WorldCat 63. Das A . Loneliness does (not) have cardiometabolic effects: A longitudinal study of older adults in two countries . Soc Sci 2019; 223:104–112 . Google Scholar OpenURL Placeholder Text WorldCat 64. Rubin RR , Ma Y, Peyrot M, et al. ; Diabetes Prevention Program Research Group. Antidepressant medicine use and risk of developing diabetes during the diabetes prevention program and diabetes prevention program outcomes study . Diabetes Care. 2010 ; 33 : 2549 – 2551 . Google Scholar Crossref Search ADS PubMed WorldCat 65. Umberson D , Crosnoe R, Reczek C. Social relationships and health behavior across life course . Annu Rev Sociol. 2010 ; 36 : 139 – 157 . Google Scholar Crossref Search ADS PubMed WorldCat 66. Whisman MA . Loneliness and the metabolic syndrome in a population-based sample of middle-aged and older adults . Health Psychol. 2010 ; 29 : 550 – 554 . Google Scholar Crossref Search ADS PubMed WorldCat 67. Wang W , Albert JM. Mediation analysis for the cox proportional hazards model with a smooth baseline hazard estimator . J R Stat Soc Serie C. 2016 ; 66 ( 4 ): 741 – 775 . Google Scholar Crossref Search ADS WorldCat © Society of Behavioral Medicine 2020. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Loneliness, Social Isolation, and Chronic Disease Outcomes
JF - Annals of Behavioral Medicine
DO - 10.1093/abm/kaaa044
DA - 2020-08-31
UR - https://www.deepdyve.com/lp/oxford-university-press/loneliness-social-isolation-and-chronic-disease-outcomes-yznLdU6IAY
SP - 1
EP - 1
VL - Advance Article
IS - 
DP - DeepDyve
ER -