Abstract
Objectives
To examine the association between types of loneliness (transient, incident, and chronic) and the risk of functional disability.
Methods
Data were from the Health and Retirement Study 2006/2008–2016/2018. A total of 7,148 adults aged ≥50 was included. Functional status was measured by activities of daily living (ADL) and instrumental activities of daily living (IADL). Loneliness was assessed using the 3-item UCLA Loneliness Scale. We defined loneliness as no/transient/incident/chronic loneliness based on the pattern and duration of loneliness across 2006/2008 and 2010/2012. We applied multivariate Cox proportional hazard models with the new-onset ADL/IADL disability as outcome.
Results
Overall, 69.3% respondents showed no loneliness; while 10.3%, 8.9%, and 11.5% showed transient, incident, and chronic loneliness, respectively. A total of 1,298 (18.16%) and 1,260 (17.63%) functionally normal respondents developed ADL and IADL disability during 36,294 person-years of follow-up, respectively. After adjusting for socio-demographic, behavioral, and health factors, chronic loneliness was associated with higher risks of ADL (hazard ratio [HR]=1.37, 95% confidence interval [CI]=1.16–1.63, p<0.001, χ2=3.60, degree of freedom [df]=1) and IADL disability (HR=1.25, 95% CI=1.09–1.44, p=0.002, χ2=3.17, df=1) compared to no loneliness. By contrast, no significant associations between transient loneliness and ADL (HR=1.17, 95% CI=0.88–1.57, p=0.273, χ2=1.10, df=1) or IADL disability (HR=1.16, 95% CI=0.97–1.39, p=0.112, χ2=1.59, df=1) were found. Chronic loneliness was not associated with the risk of IADL disability in men (HR=1.13, 95% CI=0.91–1.40, p=0.263, χ2=1.12, df=1).
Conclusions
Chronic loneliness, rather than transient loneliness, is an independent risk factor for functional disability in middle-aged and older adults, especially for women.
Keywords: physical function, social isolation, mental health, gender disparity, cohort study
INTRODUCTION
Functional ability is an individual’s actual or potential capacity to perform the activities and tasks that can be normally expected. A prior national survey in 2012 found that about 33% of Medicare beneficiaries reported mild to severe functional limitations, and 4% reported complete limitations.1 Knowledge of modifiable risk factors of functional disability is crucial to health professionals and society as it leads to excessive mortality1 and elevated burdens of health care expenditures for society.2
Loneliness is a subjective experience defined by the mismatch between actual and desired social relationships.3 An estimated 40% of adults aged ≥60 reported feeling lonely at least some time.3 Considerable literature indicates the association between loneliness and adverse health outcomes, including dementia,4 depression,5 geriatric symptoms,6 and mortality.7 However, the association between loneliness and functional disability is inconclusive. Findings from previous studies that investigated the association between loneliness and physical health in older adults are inconsistent. Loneliness is associated with increased difficulties with activities of daily living8,9 and more rapid motor decline,10 while some studies found no association between loneliness and subsequent functional status deterioration.11,12 One explanation for this discrepancy may be that previous research only used one-time assessments of loneliness.5,13 Notably, older adults vary in their life experiences and abilities to cope with stressful life events such as bereavement, retirement, chronic conditions, and sensory impairment, which might affect inter-individual trajectories of loneliness.14–16
Loneliness is classified into transient and chronic according to its pattern and course.13,14,17 Transient loneliness was described as temporary and infrequent feelings, while chronic loneliness refers to persistent lonely feelings lasting longer than two years.17 Over the life course, trajectories of loneliness depend on individual’s circumstances and perceptions. Transient loneliness can occur due to stressful life events, reductions in social network, or increases in disabilities, all of which are situations that most people recover from after a period of time;16 while chronic loneliness refers to a long-enduring feeling resulting from the lack of satisfying social relationships throughout the years.14,15 The health effect of loneliness may differ depending on its chronicity. A 18-year cohort study using data from Framingham Heart Study found that chronic loneliness was associated with higher, and transient loneliness with lower, risk of dementia onset, compared to no loneliness.13 Another population-based study indicated that compared with no loneliness, transient loneliness was associated with decreases in depression and anxiety.5 However, less is known about the differences in physical health consequences associated with transient or chronic loneliness.
Sex differences in loneliness and its effect on health have been well documented. Men and women have distinct approaches to build social networks. Men, for example, have smaller social networks and fewer intimate relationships.18 Additionally, men are less likely to express their emotions than women due to cultural norms.19 Sex differences have also been observed in the association between loneliness and functional status and mortality. Associations of loneliness with physical function tend to be stronger in women as compared to men,12,20 whereas the opposite has been found in studies on mortality.7,21 Given the mixed results from previous studies, whether the mind-body link is particularly important for women or men is unknown.
Since the onset and progression of functional disability may span across a few decades,22 and loneliness trajectories can be influenced by individual’s contexts14–16 and sex,23 we aimed to investigate the associations between types of loneliness and the risk of functional disability and whether men and women differently respond to loneliness among adults aged ≥50 in the US.
METHODS
Respondents and procedure
Data were drawn from the Health and Retirement Study (HRS), a biennial cohort study on a nationally representative community-dwelling adults aged ≥50 in the US. The survey collects information on social, demographical, economic, behavioral, and health conditions.24 HRS respondents have been asked about psychosocial characteristics in the left-behind questionnaires every four years, in addition to the biennial surveys: a random half of respondents have been asked these questions since 2006 (i.e., 2006, 2010, 2014, 2018), while the other half has been asked since 2008 (i.e., 2008, 2012, 2016).
In this study, data were from six waves of HRS. Because the loneliness data was first collected in 2006/2008 and collected every four years afterwards, we assessed types of loneliness based on two waves: 2006/2008 (T1) and 2010/2012 (T2). To fully use the data, we assessed follow-up functional status using the following five waves of the survey: 2010, 2012, 2014, 2016, and 2018. The respondents were limited to those age ≥50 at the baseline (T1) (n = 13,767). Respondents who had missing values on loneliness measure or activities of daily living/instrumental activities of daily living (ADL/IADL) at T1 or T2 (the follow up wave) (n = 2,233) were excluded. Respondents who had reported any difficulties in ADL/IADL prior to T2 (n = 1,851) were also excluded. T2 was used as the baseline to adjust for covariates. A total of 9,683 respondents were eligible for analysis (Figure 1). Of these respondents, 2,535 were excluded because they reported functional disabilities at T2 or were not followed up. The excluded respondents were older than those included. No differences were detected in sex, race/ethnicity, education, or loneliness between those included and excluded. The analytical sample included 7,148 respondents who had complete measures at T1 and T2 and a follow-up period to the end of 2016/2018 (seven years maximum) with time-to-events being death or newonset ADL/IADL disabilities (sample selection flowchart in Supplementary Figure S1).
Figure 1.
Timeline of Health and Retirement Study (HRS) data used for analysis
Note:
The population-based sample HRS had the first loneliness measured on a random-half sample in 2006, the other half were interviewed in 2008. Another round of survey on loneliness was implemented at 2010/2012, on average 4 years later. T2 was used as baseline for all analyses to control for potential covariates (i.e., respondents who developed ADL/IADL disability prior to T2 were excluded). The new-onset of ADL/IADL disability between 2010/2012 and 2016/2018 was used as outcomes for the data analyses.
Measures
Loneliness
The 3-item UCLA Loneliness Scale was used to assess loneliness. Respondents were asked to rate how often they felt they were: 1) lacking companionship, 2) left out, and 3) isolated from others, on a 3-point scale from 1 = often to 3 = hardly ever or never. The 3-item version has a good internal consistency (alpha = 0.829) and adequate concurrent and discriminant validity.25 The final score, ranging from 3 to 9, was obtained by summing the three items. For illustrative propose, we followed the procedure of Steptoe et al.26 to classify respondents with top quintile scores (≥7) as feeling lonely. This approach has been employed in multiple studies conducted in Europe26 and US.8,27
Chronic/Incident/Transient loneliness classification
In accordance with previous studies,5,8,16 we defined “chronic loneliness” as the presence of loneliness in two consecutive waves (lonely at both T1 and T2), whereas “transient loneliness” expressed lonely at T1 but not lonely at T2, and “incident loneliness” is defined not lonely at T1 but lonely at T2. Additionally, respondents who did not report loneliness at both waves were classified as having “no loneliness”. Consequently, 4,956 (69.3%), 739 (10.3%), 633 (8.9%), and 820 (11.5%) respondents were classified as having no, transient, incident, and chronic loneliness, respectively.
Functional disability
Functional disability was assessed by ADL and IADL limitations. HRS asked respondents if they required assistance with any of the six ADLs: 1) dressing, 2) eating, 3) using the toilet, 4) bathing or showering, 5) getting into or out of bed, and 6) walking across a room, or the five IADLs: 1) preparing a hot meal, 2) shopping for groceries, 3) making telephone calls, 4) taking medicines, and 5) managing money). Following previous studies,11,12 we dichotomized ADL and IADL disability into “no limitation = 0” or “at least one limitation = 1”. All respondents were free of ADL/IADL disability at baseline. Older adults would be defined to have “new-onset ADL/IADL disability” if one or more limitations emerged at follow-up.
Covariates
We selected covariates based on the review of the literature.4,12,28 Covariates were obtained at baseline and included individuals’ socio-demographic characteristics (age, sex, race/ethnicity, education, income, marital status), health-risk behavior (smoking, drinking), and mental/physical health status (depressive symptoms, self-reported health, comorbidities), see Supplement Table S1 for details.
Statistical analysis
Characteristics of respondents with different types of loneliness were compared using ANOVA for continuous variables and the χ2 tests for categorical variables. Kaplan-Meier plots were used to demonstrate the cumulative incidence rates of ADL/IADL disability with log-rank test to compare the functional disability-free probabilities between types of loneliness.
Respondents were censored at the date of record of new-onset ADL/IADL disability, death, or 2016/2018 survey, whichever came first. We calculated hazard ratios (HRs) and 95% confidence intervals (95% CIs) using Cox proportional hazards models, and adjusted for covariates that include age, sex, race/ethnicity, education, income, marital status, smoking, drinking, depressive symptoms, self-rated health, and comorbidities (Model 1). We then added the product term to test the interaction between sex and loneliness (Model 2). If the interaction effect was significant, then stratified analyses were conducted to test whether loneliness was associated with functional disability by sex. P-values for all HRs were based on Wald χ2 tests with one degree of freedom (df). The proportional hazards assumption was checked for all models by looking at Schoenfeld residuals, and the assumption was met for all models. All analyses were conducted using Stata 15.1 (College Station, TX: StataCorp LLC.), with a two-tailed value of p < 0.05 was considered statistically significant. Because there were multiple comparisons, to avoid type 1 errors, we additionally presented a more conservative Bonferroni-corrected p < 0.0083 (0.05/6).
RESULTS
Respondents’ characteristics at baseline by loneliness subgroups
Table 1 presents characteristics of the whole sample and across loneliness subgroups. The average age of the 7,148 respondents was 66.2 (SD = 8.5) years. Of them, 69.3% respondents reported no loneliness, 10.3% reported transient loneliness, 8.9% reported incident loneliness, and 11.5% reported chronic loneliness. Regarding the incident loneliness group, we treated it as a separate group because we did not include follow-up data to determine whether respondents would fit into the chronic or transient loneliness group. Table 1 shows that compared to those who reported no loneliness, respondents who reported any loneliness were more likely to be women, racial/ethnic minority, be unmarried, have lower levels of education and income, be a current smoker, have poorer self-rated health, have a higher level of depressive symptom and prevalence of diabetes (all p < 0.001). The no loneliness group had a lower incidence of ADL disability (15.8%) and IADL disability (15.9%) than other loneliness subgroups (all p < 0.001).
Table 1.
Characteristics of the whole sample and sample characteristics by loneliness subgroups (n = 7,148)
| Loneliness (n = 7,148)a | |||||||
|---|---|---|---|---|---|---|---|
| Variables | Whole sample (n = 7,148) | Not (n = 4,956, 69.3%) | Transient (n = 739, 10.3%) | Incident (n = 633, 8.9%) | Chronic (n = 820, 11.5%) | Test statisticsb | p-valuec |
| Age (years), M ± SD | 66.2 ± 8.5 | 66.6 ± 8.4 | 65.2 ± 8.7 | 66.2 ± 8.7 | 64.5 ± 8.5 | F (3, 7144) = 17.4 | ) <0.001 |
| Women, n (%) | 4,266 (58.7) | 2,856 (57.6) | 462 (62.5) | 416 (65.7) | 532 (64.9) | χ2 (3) = 28.0 | <0.001 |
| Race/ethnicity, n1/n (%) | χ2 (9) = 59.6 | <0.001 | |||||
| Non-Hispanic | 5,858/7,148 | 4,159/5,858 | 555/5,858 | 512/5,858 (8.7) | 632/5,858 | ||
| White | (82.0) | (71.0) | (9.5) | (10.8) | |||
| Non-Hispanic | 686/7,148 | 419/686 | 91/686 | 67/686 (9.8) | 109/686 | ||
| Black | (9.6) | (61.1) | (13.3) | (15.9) | |||
| Hispanic | 454/7,148 (6.3) | 288/454 (63.4) | 72/454 (15.9) | 39/454 (8.6) | 55/454 (12.1) | ||
| Others | 150/7,148 (2.1) | 90/150 (60.0) | 21/150 (14.0) | 15/150 (10.0) | 24/150 (16.0) | ||
| Married, n (%) | 1,881 (26.3) | 1,084 (21.9) | 282 (38.2) | 182 (28.8) | 333 (40.6) | χ2 (3) = 194.8 | <0.001 |
| Education (school years), n 1/n (%) | χ2 (6) = 43.3 | <0.001 | |||||
| 0–11 | 1,031/7,148 (14.4) | 628/1,031 (60.9) | 139/1,031 (13.5) | 115/1,031(11.2) | 149/1,031 (14.5) | ||
| 12 | 2,512/7,148 (35.2) | 1,743/2,512 (69.4) | 254/2,512 (10.1) | 218/2,512 (8.7) | 297/2,512 (118) | ||
| 13+ | 3,605/7,148 (50.4) | 2,585/3,605 (71.7) | 346/3,605 (9.6) | 300/2,585 (8.3) | 374/3,605 (10.4) | ||
| Household income, n1/n (%) | χ2 (9) = 135.6 | <0.001 | |||||
| 1st quartile | 1,790/7,148 (25.0) | 1,070/1,790 (59.8) | 238/1,790 (13.3) | 184/1,790 (10.3) | 298/1,790 (16.6) | ||
| 2nd quartile | 1,784/7,148 (25.0) | 1,225/1,784 (68.7) | 194/1,784 (10.9) | 163/1,784 (9.11) | 202/1,784 (113) | ||
| 3rd quartile | 1,787/7,148 (25.0) | 1,300/1,787 (72.8) | 167/1,787 (9.3) | 148/1,787 (8.3) | 172/1,787 (9.6) | ||
| 4th quartile | 1,787/7,148 (25.0) | 1,361/1,787 (76.2) | 140/1,787 (7.8) | 138/1,787 (7.7) | 148/1,787 (8.3) | ||
| Current smoker, n (%) | 780 (10.9) | 482 (9.7) | 94 (12.7) | 82 (13.0) | 122 (14.9) | χ2 (3) = 26.6 | <0.001 |
| Current heavy drinker, n (%) | 1,055 (14.8) | 805 (16.3) | 97 (13.1) | 75 (11.9) | 78 (9.5) | χ2 (3) = 32.2 | <0.001 |
| CES-D (ranges 07), M (SD) | 0.9 ± 1.4 | 0.6 ± 1.1 | 1.4 ± 1.8 | 1.0 ± 1.5 | 1.9 ± 2.1 | F (3, 7144) = 251.4 | <0.001 |
| Hypertension, n (%) | 3,900 (54.7) | 2,669 (54.0) | 408 (55.5) | 338 (53.7) | 485 (59.4) | χ2 (3) = 8.0 | 0.046 |
| Diabetes, n (%) | 1,173 (16.4) | 739 (14.9) | 133 (18.0) | 122 (19.4) | 179 (21.9) | χ2 (3) = 30.1 | <0.001 |
| Heart diseases, n (%) | 1,429 (20.1) | 994 (20.1) | 134 (18.2) | 130 (20.6) | 171 (21.0) | χ2 (3) = 2.2 | 0.533 |
| Self-rated health (ranges 1–5), M (SD) | 2.5 ± 1.0 | 2.4 ± 0.9 | 2.7 ± 1.0 | 2.7 ± 1.0 | 2.9 ± 1.0 | F (3, 7144) = 82.2 | <0.001 |
| Follow-up characteristic | |||||||
| New-onset ADL disability at follow-up, n (%) | 1,298 (18.2) | 783 (15.8) | 161 (21.8) | 145 (23.1) | 208 (25.4) | χ2 (3) = 69.4 | <0.001 |
| New-onset IADL disability at follow-up, n (%) | 1,260 (17.6) | 788 (15.9) | 162 (21.9) | 138 (21.8) | 172 (21.0) | χ2 (3) = 34.1 | <0.001 |
Notes:
The four loneliness subgroups were defined as: no loneliness = respondents did not report loneliness at either T1 (2006/2008) or T2 (2010/2012, baseline); transient loneliness = respondents reported loneliness only T1; incident loneliness = respondents reported loneliness only at T2; chronic loneliness = respondents reported loneliness at both T1 and T2.
F (degree of freedom)/χ2 (degree of freedom).
P-values derived from ANOVA for continuous variables and Pearson’s χ2 tests for categorical variables. Abbreviations: M = Mean; SD = Standard Deviation; ADL = Activities of Daily Living; IADL = Instrumental Activities of Daily Living; CESD = Center for Epidemiologic Studies Depression Scale score without the loneliness item.
The associations between the loneliness subgroups and functional disability
Figure 2 shows the association between loneliness subgroups and the risk of ADL/IADL disability. Compared to the respondents who had no loneliness, those with transient or chronic loneliness had higher risks of ADL disability (15.8% vs. 21.8% and 25.4%, respectively, p < 0.001, χ2 = 69.4, df = 3) (Figure 2A) and IADL disability (15.9% vs. 21.9% and 21.0%, respectively, p < 0.001, χ2 = 34.1, df = 3) (Figure 2B). As illustrated in Kaplan-Meier plots (Figure 2C, 2D), respondents with chronic or transient loneliness were more likely to develop functional disabilities than those with no loneliness (log-rank tests, all p < 0.001).
Figure 2.
Accumulative incidence and Log-rank test analysis for ADL and IADL disabilities based on loneliness subgroups
Notes:
No loneliness = respondents did not report loneliness at either T1 (2006/2008) or T2 (2010/2012, baseline); transient loneliness = respondents reported loneliness only T1; incident loneliness = respondents reported loneliness only at T2; chronic loneliness = respondents reported loneliness at both T1 and T2.
The incident (A) ADL disability and (B) IADL disability between those with no/transient/incident/chronic loneliness were compared by using χ2 test.
Kaplan-Meier plot (C, D) was used to show the survival curves of up to 84-month of follow-up, and the Logrank test was used to compare the survival curves of loneliness with the no loneliness group as reference.
Bonferroni-corrected statistical significance is shown as *, p < 0.05; **, p< 0.01; ****, p < 0.001.
Abbreviations: ADL = Activities of Daily Living; IADL = Instrumental Activities of Daily Living.
In the multivariate Cox models, chronic loneliness was associated with higher risks of ADL/IADL disabilities compared with the no loneliness group when accounting for all covariates in the full model (Table 2, Model 1). Transient loneliness was not significantly associated with the risk of ADL (HR = 1.17, 95% CI = 0.88–1.57) and IADL (HR = 1.16, 95% CI = 0.97–1.39) disability. Other factors independently predicting ADL and IADL disability in the full model were age, education, smoking, depressive symptoms, self-reported health, and diabetes.
Table 2.
Multivariate Cox hazard models for the association between loneliness and the new-onset of ADL and IADL disabilitiesa
| Variables | Event = ADL disability | Event = IADL disability | ||||||
|---|---|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 1 | Model 2 | |||||
| HR (95% CI) | Wald χ2 (p-value)b | HR (95% CI) | Wald χ2 (p-value)b | HR (95% CI) | Wald χ2 (p-value)b | HR (95% CI) | Wald χ2 (p-value)b | |
| Loneliness (ref. no loneliness) | ||||||||
| Transient loneliness | 1.17 (0.88,1.57) | 1.10 (0.273) | 1.15 (0.87,1.51) | 0.96 (0.337) | 1.16 (0.97,1.39) | 1.59 (0.112) | 1.12 (0.99,1.26) | 1.71 (0.090) |
| Incident loneliness | 1.33 (1.12,1.59) | 3.22 (0.001) | 1.35 (1.01,1.80) | 3.18 (0.002) | 1.26 (1.04,1.52) | 2.40 (0.017) | 1.20 (1.01,1.43) | 2.09 (0.036) |
| Chronic loneliness | 1.37 (1.16,1.63) | 3.60 (<0.001) | 1.36 (1.14,1.62) | 3.54 (<0.001) | 1.25 (1.09,1.44) | 3.17 (0.002) | 1.26 (1.10,1.44) | 3.19 (0.002) |
| Age (years) | 1.07 (1.07,1.08) | 18.64 (<0.001) | 1.07 (1.07,1.08) | 18.60 (<0.001) | 1.09 (1.08,1.10) | 22.38 (<0.001) | 1.09 (1.08,1.10) | 22.34 (<0.001) |
| Women | 0.94 (0.84,1.06) | 0.07 (0.946) | 0.98 (0.84,1.13) | 0.06 (0.951) | 0.92 (0.81,1.04) | 1.92 (0.055) | 0.92 (0.79,1.07) | 1.51 (0.132) |
| Race/ethnicity (ref. Non- Hispanic White) | ||||||||
| Non-Hispanic Black | 1.09 (0.91,1.31) | 1.42 (0.155) | 1.09 (0.91,1.32) | 1.42 (0.156) | 1.05 (0.86,1.27) | 0.41 (0.684) | 1.05 (0.86,1.27) | 0.41 (0.683) |
| Hispanic | 0.92 (0.73,1.15) | 1.16 (0.245) | 0.92 (0.73,1.15) | 1.17 (0.241) | 0.95 (0.75,1.19) | 0.21 (0.832) | 0.95 (0.75,1.19) | 0.21 (0.834) |
| Others | 1.17 (0.80, 1.69) | 0.04 (0.967) | 1.16 (0.80,1.68) | 0.05 (0.958) | 0.93 (0.61,1.43) | 0.45 (0.655) | 0.93 (0.61,1.43) | 0.45 (0.652) |
| Education (ref. 0–11 school years) | ||||||||
| 12 years | 0.72 (0.61,0.84) | 3.23 (0.001) | 0.72 (0.61,0.84) | 3.24 (0.001) | 0.78 (0.67,0.92) | 3.32 (0.001) | 0.78 (0.67,0.92) | 3.32 (0.001) |
| 13+ years | 0.78 (0.67,0.92) | 3.23 (0.001) | 0.78 (0.67,0.92) | 3.24 (0.001) | 0.79 (0.67,0.93) | 3.20 (0.001) | 0.79 (0.67,0.93) | 3.21 (0.001) |
| Household income (ref. 1st quartile) | ||||||||
| 2nd quartile | 0.99 (0.85,1.15) | 0.77 (0.441) | 0.99 (0.85,1.15) | 0.76 (0.445) | 0.98 (0.84,1.14) | 0.80 (0.423) | 0.98 (0.84,1.14) | 0.80 (0.423) |
| 3rd quartile | 0.90 (0.76,1.08) | 0.74 (0.460) | 0.90 (0.76,1.08) | 0.74 (0.462) | 0.80 (0.67,0.96) | 2.70 (0.007) | 0.80 (0.67,0.96) | 2.70 (0.007) |
| 4th quartile | 0.87 (0.71,1.07) | 1.61 (0.108) | 0.87 (0.71,1.06) | 1.61 (0.107) | 0.86 (0.70,1.05) | 1.95 (0.052) | 0.86 (0.70,1.05) | 1.94 (0.052) |
| Married | 0.97 (0.84,1.12) | 0.17 (0.866) | 0.97 (0.84,1.12) | 0.18 (0.860) | 0.88 (0.76,1.02) | 1.56 (0.120) | 0.88 (0.76,1.02) | 1.56 (0.119) |
| Current smoker | 1.61 (1.36,1.91) | 4.67 (<0.001) | 1.61 (1.36,1.91) | 4.67 (<0.001) | 1.57 (1.31,1.88) | 5.51 (<0.001) | 1.57 (1.31,1.88) | 5.51 (<0.001) |
| Current heavy drinker | 0.89 (0.75,1.06) | 0.72 (0.472) | 0.89 (0.75,1.06) | 0.72 (0.470) | 0.88 (0.74,1.04) | 1.30 (0.193) | 0.88 (0.74,1.04) | 1.30 (0.194) |
| CESD | 1.12 (1.07,1.16) | 5.67 (<0.001) | 1.12 (1.07,1.16) | 5.67 (<0.001) | 1.07 (1.02,1.11) | 3.72 (<0.001) | 1.07 (1.02,1.11) | 3.71 (<0.001) |
| Self-reported health | 1.35 (1.26,1.44) | 9.16 (<0.001) | 1.35 (1.26,1.44) | 9.17 (<0.001) | 1.34 (1.25,1.44) | 8.67 (<0.001) | 1.34 (1.25,1.44) | 8.67 (<0.001) |
| Having hypertension | 1.07 (0.95,1.21) | 0.90 (0.370) | 1.07 (0.95,1.21) | 0.90 (0.366) | 1.12 (0.99,1.26) | 1.72 (0.086) | 1.12 (0.99,1.26) | 1.72 (0.086) |
| Having diabetes | 1.25 (1.09,1.44) | 3.17 (0.002) | 1.25 (1.09,1.44) | 3.16 (0.002) | 1.21 (1.05,1.39) | 2.72 (0.007) | 1.21 (1.05,1.39) | 2.71 (0.007) |
| Having heart diseases | 1.04 (0.91,1.18) | 1.19 (0.236) | 1.04 (0.91,1.18) | 1.16 (0.246) | 1.05 (0.92,1.20) | 0.79 (0.429) | 1.05 (0.92,1.20) | 0.79 (0.431) |
| Women × Transient loneliness | 1.08 (0.76,1.53) | 0.44 (0.661) | 1.07 (1.01, 1.14) | 2.71 (0.007) | ||||
| Women × Incident loneliness | 0.99 (0.69,1.42) | 0.06 (0.953) | 1.07 (0.68,1.47) | 0.70 (0.420) | ||||
| Women × Chronic loneliness | 1.00 (0.73,1.37) | 0.00 (0.999) | 1.11 (1.02,1.20) | 2.37 (0.019) | ||||
Note:
Cox proportional hazard regression models were applied to estimated hazard ratios (HRs) of ADL/IADL disability. All p-values for the overall test less than 0.001.
P-values were based on Wald χ2 tests with one degree of freedom.
Model 1: added loneliness subgroups and adjusted to age, sex, race/ethnicity, education, income, marital status, smoking, drinking, depression, self-rated health, and comorbidities. Model 2: Model 1 plus the interaction between sex and loneliness.
Abbreviations: CI = Confidence Interval; ADL = Activities of Daily Living; IADL = Instrumental Activities of Daily Living; CESD = Center for Epidemiologic Studies Depression Scale score without the loneliness item.
The associations between loneliness and functional disability stratified by sex
In the Cox models that tested the interaction effect of sex and loneliness, significant interactions between sex and loneliness were obtained for the risk of IADL disability (p = 0.007 for women × transient loneliness; p = 0.019 for women × chronic loneliness; Table 2, Model 2). We therefore stratified the respondents into men (n = 2,882) and women (n = 4,266). As shown in Supplementary Table S2, compared with men, women are more likely to experience transient, incident, and chronic loneliness (p < 0.001). No sex difference was observed in the risk of ADL (17.9% for men vs. 18.4% for women, p = 0.223) and IADL disability (17.9% for men vs. 17.4% for women, p = 0.450).
In the sex-stratified Cox models (Table 3), for men, the association between chronic loneliness and IADL disability did not reach statistical significance (HR = 1.13, 95% CI = 0.91–1.40); for women, compared to the no loneliness group, individuals with chronic loneliness had a higher risk of IADL disability (HR = 1.29, 95% CI = 1.16–1.44). Notably, different from chronic loneliness, transient loneliness was not associated with the risk of any functional disabilities in both men and women (men: HR = 1.17, 95% CI = 0.87–1.58; women: HR = 1.16, 95% CI = 0.92–1.46).
Table 3.
Multivariate Cox hazard models for the association between loneliness and the risk of IADL disability stratified by sexa
| Men (n = 2,882) | Women (n = 4,266) | |||
|---|---|---|---|---|
| HR (95% CI) | Wald χ2 (p-value)b | HR (95% CI) | Wald χ2 (p-value)b | |
| Loneliness (ref. no loneliness) | ||||
| Transient loneliness | 1.17 (0.87,1.58) | 1.05 (0.294) | 1.16 (0.92,1.46) | 1.28 (0.201) |
| Incident loneliness | 1.28 (0.94,1.73) | 1.57 (0.117) | 1.24 (1.08,1.42) | 3.10 (0.002) |
| Chronic loneliness | 1.13 (0.91,1.40) | 1.12 (0.263) | 1.29 (1.16,1.44) | 4.15 (<0.001) |
| Age (years) | 1.09 (1.08,1.10) | 14.26 (<0.001) | 1.09 (1.08,1.10) | 16.87 (<0.001) |
| Race/Ethnicity (ref. Non-Hispanic White) | ||||
| Non-Hispanic Black | 1.03 (0.75,1.40) | 0.17 (0.867) | 1.03 (0.80,1.31) | 0.21 (0.836) |
| Hispanic | 1.02 (0.73,1.44) | 0.14 (0.892) | 0.93 (0.68,1.27) | 0.46 (0.644) |
| Others | 1.17 (0.69,1.97) | 0.58 (0.561) | 0.56 (0.25,1.25) | 1.43 (0.154) |
| Education (ref. 0–11 school years) | ||||
| 12 years | 0.94 (0.73,1.21) | 0.47 (0.635) | 0.69 (0.56,0.84) | 3.58 (<0.001) |
| 13+ years | 0.82 (0.63,1.06) | 1.53 (0.125) | 0.75 (0.60,0.93) | 2.57 (0.010) |
| Household income (ref. 1st quartile) | ||||
| 2nd quartile | 0.72 (0.55,0.95) | 0.26 (0.797) | 0.84 (0.66,1.08) | 0.94 (0.350) |
| 3rd quartile | 0.72 (0.55,0.95) | 2.34 (0.019) | 0.84 (0.66,1.08) | 1.36 (0.173) |
| 4th quartile | 0.68 (0.51,0.93) | 2.45 (0.014) | 0.98 (0.74,1.29) | 0.17 (0.863) |
| Married | 0.76 (0.59,0.99) | 2.07 (0.038) | 0.97 (0.81,1.16) | 0.39 (0.700) |
| Current smoker | 1.79 (1.36,2.35) | 4.15 (<0.001) | 1.53 (1.21,1.94) | 3.55 (<0.001) |
| Current heavy drinker | 0.87 (0.68,1.10) | 1.20 (0.229) | 0.91 (0.70,1.18) | 0.74 (0.457) |
| CESD | 1.08 (1.01,1.15) | 2.25 (0.024) | 1.08 (1.03,1.13) | 2.98 (0.003) |
| Self-reported health | 1.37 (1.24,1.52) | 5.94 (<0.001) | 1.35 (1.23,1.48) | 6.44 (<0.001) |
| Having hypertension | 1.06 (0.88,1.28) | 0.82 (0.415) | 1.15 (0.97,1.35) | 1.50 (0.134) |
| Having diabetes | 1.13 (0.91,1.40) | 1.12 (0.264) | 1.29 (1.06,1.56) | 2.57 (0.010) |
| Having heart diseases | 0.93 (0.77,1.14) | 0.67 (0.503) | 1.15 (0.96,1.38) | 1.49 (0.136) |
Note:
Cox proportional hazard regression models were applied to estimated hazard ratios (HRs) of IADL disability. All p-values for the overall test less than 0.001.
P-values were based on Wald χ2 tests with 1 degree of freedom.
Abbreviations: CI = Confidence Interval; ADL = Activities of Daily Living; IADL = Instrumental Activities of Daily Living; CESD = Center for Epidemiologic Studies Depression Scale score without the loneliness item.
DISCUSSION
The present study examined the associations between types of loneliness and the risks of ADL/IADL disabilities among functionally-normal adults aged ≥50 over a 7-year follow-up. Compared with respondents who never experienced loneliness, those who experienced chronic loneliness were at higher risks of ADL/IADL disabilities. However, respondents who experienced transient loneliness showed no excess risk than those who experienced no loneliness. We also found women had a higher risk of IADL disability when they experienced chronic loneliness compared with men. In our study, 820 (11.5%) of adults aged ≥50 in the US reported chronic loneliness. Our estimates on the chronicity of loneliness are comparable to other national surveys. In the National Social Life, Health and Aging Project wave 1–2, 18% of adults aged 5785 years reported chronic loneliness.14 Data from Medicare Total Health Assessment (MTHA) 2013–2018 indicates 7.8% of older adults aged ≥65 were persistently lonely.5 In light of these population-level estimates of prevalence rates of chronic loneliness, and the estimated sizable magnitude for the effect of chronic loneliness on ADL/IADL disabilities (HR, 1.25–1.37, adjusted for a large set of important covariates),29 chronic loneliness may be a clinically important concern.
Loneliness is a well-recognized stressor that can cause increased cortisol levels and impair biological function.30 This can accelerate the physiological resilience decline in older adults, whose physical capacity to buffer against stressors is deteriorating due to aging. Loneliness may also relate to health-risk behaviors (e.g., smoking, alcohol abuse, and lack of physical activity)31 and psychological distress (e.g., depression, anxiety),32 both of which are known to be associated with poor physical function. Previous studies also indicate that stress and health-risk behaviors mediate the association between loneliness and adverse health conditions.33 More studies are needed to explore the pathways of the association between loneliness and physical health consequences.
Our findings distinguishing between chronic and transient loneliness clarified why prior studies may have inconsistent findings:9,28,32 cases of chronic loneliness represent a minority of those detected in a cross-sectional survey of the population, but account for most of the burden of functional disability associated with loneliness. It is worth noting that adults with incident loneliness are also at a higher risk for ADL/IADL disability. Loneliness could occur due to a variety of circumstances, including abrupt changes in living arrangements, life events, and/or perceived control. Incident loneliness may be associated with individuals’ increases in sensory impairment, poorer health status, and incapability in coping stressful life events (e.g., bereavement, retirement),16 all of which are detrimental to one’s physical functioning. These findings are further supported by Holt-Lunstad and Steptoe’s postulation that loneliness represents a weak social foundation leading to significant vulnerability.34
Unlike chronic loneliness, transient loneliness was not associated with the risks of functional disabilities in both sexes. These findings were expected given the distinct nature of transient and chronic loneliness. Transient loneliness could represent an adaptive response, but chronic loneliness is often a product of prolonged interpersonal difficulties and not a result of loss or other setbacks that can be overcome.17,35 Given the origins of loneliness and its potential adaptive function, Cacioppo et al.36 proposed the evolutionary theory of loneliness, arguing that transient loneliness may have been necessary for the evolution of the human species as a warning to reconnect with other individuals, increasing the chances of survival and achieving gene transmission to future generations. However, if it becomes chronic, it could be detrimental. The self-reinforcing loneliness loop may exist in the lonely individuals, making them more likely to see social contacts as threatening, anticipate negative social relationships, and recall more negative social information; therefore, they would prefer to avoid future social connections.36 Repeated exposure to social disconnection over prolonged periods may provide an obstacle to successfully modifying social connections in accordance with individuals’ abilities and then contribute to poorer health.5,13,36 In our study, compared to respondents who reported no loneliness, those with chronic loneliness were more likely to have poorer self-rated health, a higher level of depressive symptom, and higher prevalence of comorbidities (Table 1); all of which may increase the risk of functional disability. Inflammatory biomarkers may be linked to the trajectory of loneliness, particularly in a long-term exposure form. For example, chronic loneliness is associated with elevated inflammatory responses, which causes the Hypothalamic-Pituitary-Adrenal axis to become overactive and ineffective.30,34
In our study, the association between chronic loneliness and IADL disability was only significant in women rather than men. Loneliness may be experienced in different pattens by sex. Women, compared to men, are believed to be more sensitive to the interpersonal context, desiring greater interpersonal connectedness than interpersonal autonomy.37 In turn, this greater sensitivity appears to be associated with a higher risk of experiencing negative affective conditions (e.g., depression, stress), which are implicated in physical health.38 As argued in the social capital theory, lonelier older adults tend to have smaller social networks and lack functional and informational social resources that are necessary for maintaining good health.39 And it is well-recognized that men have smaller social networks and fewer intimate relationships compared with women.18 Empirical studies also indicate that men rely on spousal ties to preserve their physical health, whereas women rely on extended network members such as spouses, children, or others.40 Therefore, women who experienced loneliness had fewer chances to benefit from social connections and were more likely to develop functional disability.12
The sex difference in the effect of loneliness on the ADL and IADL disabilities can be further explained by their distinct nature. ADL scale evaluates the most basic processes involved in everyday function and focuses on at-home activities. Whereas the skills covered by the IADL are more cognitively influenced that require cognitive capacity.41 Older men are less vulnerable to the impact of social disconnection on cognitive function than older women,42 which indicates men has a better IADL than women when exposed to loneliness. A more rapid cognitive decline with age for women than men may further explain why risk factors such as loneliness more significantly influences IADL than ADL in women.41 It should be noted that we did not observe the significant associations between loneliness and the risk of IADL disability in men, probably because of the relatively short follow-up duration and/or the unbalanced distribution across sex (i.e., 40.3% were men and 59.7% were women). Studies with longer follow-up periods and more balanced distribution of sexes are needed to confirm these findings.
There are limitations to our research. First, the lack of more frequent data collection on loneliness between two data collection points restricts our ability to determine whether the feelings of loneliness were chronic or transient. We also do not know whether respondents who felt lonely only at the latter time point (“incident loneliness”) would be classified as transient or chronic if we added more follow-ups. Second, loneliness was assessed with only three direct questions regarding the perception of lacking companionship, being left out, and being isolated from others. Despite being widely used in the literature and strong correlations with several established scales8,25,27 this measure did not capture the multidimensional nature of loneliness. Third, sample in this study was limited to community-dwelling adults. Although loneliness and physical disability are more common in nursing home residents, the findings may not be generalizable to them.
Considering the growth of the aging population and the long-term impact COVID-19 pandemic, more people may have chronic feelings of loneliness. Future research will need to investigate the mechanisms that underpin loneliness and physical health consequences to investigate whether reducing loneliness could potentially protect from functional disability. In summary, our findings advance evidence in support of the evolutionary theory of loneliness from a national, long-term follow-up of older adults,30,36 and highlight the chronicity of loneliness as a critical dimension in determining impacts on healthy aging.
CONCLUSIONS
Chronic loneliness may be an independent, modifiable risk factor for functional disability, especially for women. Considering the population aging, more people may face increasingly stressful life events such as spousal loss, living alone, financial hardship, and natural disasters, all of which are likely to exacerbate loneliness. Our results suggest that focusing on loneliness interventions could reduce the risk of functional disability. The findings encourage tailored interventions to the right people at the right time to avoid loneliness chronicity and prevent functional disability.
Supplementary Material
Highlights.
1. What is the primary question addressed by this study?
Could the effect of loneliness on physical health consequences differ by its patterns and duration and whether there are sex differences in its effect?
2. What is the main finding of this study?
Chronic loneliness is an independent risk factor for new-onset ADL disability in both men and women, but transient loneliness is not.
Chronic loneliness was associated with the risk of IADL disability in women, not in men.
3. What is the meaning of the finding?
Tailored interventions are needed for the right people at the right time to avoid loneliness chronicity and prevent functional disability.
Acknowledgment
The Health and Retirement Study is sponsored by the National Institute on Aging (grant number U01AG009740) and is conducted by the University of Michigan. The authors thank all participants and staff for their contribution to this study.
Funding Statement
This study is supported by the National Institute on Aging (R01AG057800 and P30AG059304), National Institute on Minority Health and Health Disparities (P50MD017356), University Cancer Research Funds, Lineberger Cancer Center at University of North Carolina at Chapel Hill.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of Interest Statement
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Statement
The data has not been previously presented orally or by poster at scientific meetings.
Data Availability Statement
The Health and Retirement Study (HRS) datasets are publicly available at the University of Michigan Institute for Social Research. Researchers may obtain the datasets after sending a data user agreement to the HRS team (https://hrs.isr.umich.edu/data-products).
REFERENCES
- 1.Jia H, Lubetkin EI. Life expectancy and active life expectancy by disability status in older U.S. adults. Reddy H, ed. PLOS ONE. 2020;15(9):e0238890. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Fried TR, Bradley EH, Williams CS, Tinetti ME. Functional Disability and Health Care Expenditures for Older Persons. Arch Intern Med. 2001;161(21):2602. [DOI] [PubMed] [Google Scholar]
- 3.National Academies of Sciences, Engineering, and Medicine, ed. Social Isolation and Loneliness in Older Adults: Opportunitiies for the Health Care System. the National Academies Press; 2020. [PubMed] [Google Scholar]
- 4.Sutin AR, Stephan Y, Luchetti M, Terracciano A. Loneliness and Risk of Dementia. Martire L, ed. J Gerontol Ser B. 2020;75(7):1414–1422. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Steiner JF, Ross C, Stiefel M, et al. Association between changes in loneliness identified through screening and changes in depression or anxiety in older adults. J Am Geriatr Soc. 2022;70(12):3458–3468. [DOI] [PubMed] [Google Scholar]
- 6.Powell VD, Kumar N, Galecki AT, et al. Bad company: Loneliness longitudinally predicts the symptom cluster of pain, fatigue, and depression in older adults. J Am Geriatr Soc. Published online April 12, 2022:jgs.17796. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Rico-Uribe LA, Caballero FF, Martín-María N, Cabello M, Ayuso-Mateos JL, Miret M. Association of loneliness with all-cause mortality: A meta-analysis. Jacobs JM, ed. PLOS ONE. 2018;13(1):e0190033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Crowe CL, Domingue BW, Graf GH, Keyes KM, Kwon D, Belsky DW. Associations of Loneliness and Social Isolation With Health Span and Life Span in the U.S. Health and Retirement Study. Newman AB, ed. J Gerontol Ser A. 2021;76(11):1997–2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Perissinotto CM, Stijacic Cenzer I, Covinsky KE. Loneliness in Older Persons: A Predictor of Functional Decline and Death. Arch Intern Med. 2012;172(14). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Buchman AS, Boyle PA, Wilson RS, et al. Loneliness and the rate of motor decline in old age: the rush memory and aging project, a community-based cohort study. BMC Geriatr. 2010;10(1):77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Stessman J, Rottenberg Y, Shimshilashvili I, Ein-Mor E, Jacobs JM. Loneliness, Health, and Longevity. J Gerontol A Biol Sci Med Sci. 2014;69(6):744–750. [DOI] [PubMed] [Google Scholar]
- 12.Guo L, An L, Luo F, Yu B. Social isolation, loneliness and functional disability in Chinese older women and men: a longitudinal study. Age Ageing. 2021;50(4):1222–1228. [DOI] [PubMed] [Google Scholar]
- 13.Akhter-Khan SC, Tao Q, Ang TFA, et al. Associations of loneliness with risk of Alzheimer’s disease dementia in the Framingham Heart Study. Alzheimers Dement. Published online March 24, 2021:alz.12327. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hawkley LC, Kocherginsky M. Transitions in Loneliness Among Older Adults: A 5-Year Follow-Up in the National Social Life, Health, and Aging Project. Res Aging. 2018;40(4):365–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mund M, Freuding MM, Möbius K, Horn N, Neyer FJ. The Stability and Change of Loneliness Across the Life Span: A Meta-Analysis of Longitudinal Studies. Personal Soc Psychol Rev. 2020;24(1):24–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Newall NEG, Chipperfield JG, Bailis DS. Predicting stability and change in loneliness in later life. J Soc Pers Relatsh. 2014;31(3):335–351. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Peplau LA, Perlman D, eds. Loneliness: A Sourcebook of Current Theory, Research, and Therapy. Wiley; 1982. [Google Scholar]
- 18.Borys S, Perlman D. Gender Differences in Loneliness. Pers Soc Psychol Bull. 1985;11(1):63–74. [Google Scholar]
- 19.Tijhuis MA, De Jong-Gierveld J, Feskens EJ, Kromhout D. Changes in and factors related to loneliness in older men. The Zutphen Elderly Study. Age Ageing. 1999;28(5):491–495. [DOI] [PubMed] [Google Scholar]
- 20.Yu B, Steptoe A, Niu K, Jia X. Social Isolation and Loneliness as Risk Factors for Grip Strength Decline Among Older Women and Men in China. J Am Med Dir Assoc. 2020;21(12):1926–1930. [DOI] [PubMed] [Google Scholar]
- 21.Yang YC, McClintock MK, Kozloski M, Li T. Social Isolation and Adult Mortality: The Role of Chronic Inflammation and Sex Differences. J Health Soc Behav. 2013;54(2):183203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Yang Y, George LK. Functional Disability, Disability Transitions, and Depressive Symptoms in Late Life. J Aging Health. 2005;17(3):263–292. [DOI] [PubMed] [Google Scholar]
- 23.Dahlberg L, Andersson L, McKee KJ, Lennartsson C. Predictors of loneliness among older women and men in Sweden: A national longitudinal study. Aging Ment Health. 2015;19(5):409–417. [DOI] [PubMed] [Google Scholar]
- 24.Sonnega A, Faul JD, Ofstedal MB, Langa KM, Phillips JW, Weir DR. Cohort Profile: the Health and Retirement Study (HRS). Int J Epidemiol. 2014;43(2):576–585. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.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(6):655–672. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Steptoe A, Shankar A, Demakakos P, Wardle J. Social isolation, loneliness, and all-cause mortality in older men and women. Proc Natl Acad Sci. 2013;110(15):5797–5801. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Qi X, Pei Y, Malone SK, Wu B. Social Isolation, Sleep Disturbance, and Cognitive Functioning (HRS): A Longitudinal Mediation Study. J Gerontol Ser A. Published online January 6, 2023:glad004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Shankar A, McMunn A, Demakakos P, Hamer M, Steptoe A. Social isolation and loneliness: Prospective associations with functional status in older adults. Health Psychol. 2017;36(2):179–187. [DOI] [PubMed] [Google Scholar]
- 29.Selvin S Survival Analysis for Epidemiologic and Medical Research: A Practical Guide. Cambridge University Press; 2008. [Google Scholar]
- 30.Quadt L, Esposito G, Critchley HD, Garfinkel SN. Brain-body interactions underlying the association of loneliness with mental and physical health. Neurosci Biobehav Rev. 2020;116:283–300. [DOI] [PubMed] [Google Scholar]
- 31.Lund R, Laban J, Petersen GL, Dissing A, Hansen ÅM, Rod NH. Loneliness and objectively measured physical capability in middle-aged adults. Eur J Public Health. 2018;28(1):16–23. [DOI] [PubMed] [Google Scholar]
- 32.Domènech-Abella J, Mundó J, Switsers L, Tilburg T, Fernández D, Aznar-Lou I. Social network size, loneliness, physical functioning and depressive symptoms among older adults: Examining reciprocal associations in four waves of the Longitudinal Aging Study Amsterdam (LASA). Int J Geriatr Psychiatry. 2021;36(10):1541–1549. [DOI] [PubMed] [Google Scholar]
- 33.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. [DOI] [PubMed] [Google Scholar]
- 34.Hawkley LC, Cacioppo JT. Loneliness matters: a theoretical and empirical review of consequences and mechanisms. Ann Behav Med Publ Soc Behav Med. 2010;40(2):218–227. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Cacioppo JT, Ernst JM, Burleson MH, et al. Lonely traits and concomitant physiological processes: the MacArthur social neuroscience studies. Int J Psychophysiol. 2000;35(23):143–154. [DOI] [PubMed] [Google Scholar]
- 36.Cacioppo JT, Cacioppo S, Boomsma DI. Evolutionary mechanisms for loneliness. Cogn Emot. 2014;28(1):3–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Cross SE, Madson L. Models of the self: Self-construals and gender. Psychol Bull. 1997;122(1):5–37. [DOI] [PubMed] [Google Scholar]
- 38.Lenze EJ, Rogers JC, Martire LM, et al. The Association of Late-Life Depression and Anxiety With Physical Disability: A Review of the Literature and Prospectus for Future Research. Am J Geriatr Psychiatry. 2001;9(2):113–135. [PubMed] [Google Scholar]
- 39.Lin N, Cook KS, Burt RS, eds. Social Capital: Theory and Research. Aldine de Gruyter; 2001. [Google Scholar]
- 40.Saito T, Murata C, Aida J, Kondo K. Cohort study on living arrangements of older men and women and risk for basic activities of daily living disability: findings from the AGES project. BMC Geriatr. 2017;17(1):183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Scheel-Hincke LL, Möller S, Lindahl-Jacobsen R, Jeune B, Ahrenfeldt LJ. Cross-national comparison of sex differences in ADL and IADL in Europe: findings from SHARE. Eur J Ageing. 2020;17(1):69–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Joyce J, Ryan J, Owen A, et al. Social isolation, social support, and loneliness and their relationship with cognitive health and dementia. Int J Geriatr Psychiatry. 2022;37(1):gps.5644. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The Health and Retirement Study (HRS) datasets are publicly available at the University of Michigan Institute for Social Research. Researchers may obtain the datasets after sending a data user agreement to the HRS team (https://hrs.isr.umich.edu/data-products).