Abstract
Objectives
Later life is often categorized by higher-than-average levels of loneliness, but individual differences are vast and not well understood. Emerging evidence indicates that broad-based contextual factors such as the coronavirus disease 2019 (COVID-19) pandemic—and the use of the internet throughout—are differentially associated with the experience of loneliness. We, therefore, target internet usage and loneliness among middle-aged and older adults during the pandemic and examine the moderating role of age, gender, and limiting illness therein.
Methods
We applied hierarchical regression models to data from the COVID-19 substudy Wave 1 (June/July 2020) of the English Longitudinal Study of Ageing (N = 4,790; Mage = 70.2; standard deviation = 9.0; range: 50–90; 43.5% male).
Results
Infrequent internet use was associated with less loneliness compared with very frequent users—an association that strengthened with age. Conversely, the purpose of internet use was associated with more loneliness, with higher levels exhibited by those searching for health-related information—an effect stronger among those with a limiting illness.
Discussion
Findings imply that infrequent internet use may reduce loneliness, while health-related internet searches may increase loneliness among older adults with different physical capacities. Findings are contrary to prepandemic reports, underscoring the importance of broad-based contextual factors for understanding loneliness across adulthood and old age.
Keywords: Contextual development, Information and Communication Technology (ICT), Social isolation
The experience of loneliness reflects the social contexts in which we are living (Elder, 1995). Throughout 2020, the first wave(s) of the coronavirus disease (COVID-19) caused governments to implement restrictive measures, resulting in elevated isolation and psychological distress in middle-aged and older populations (Nelson & Bergeman, 2021; Peng & Roth, 2022). Importantly, evidence suggests that pandemic-specific isolation represents a new and emerging risk factor for older adults (Pai & Vella, 2021). Although evidence for pandemic-related loneliness in this demographic has been mixed (Ernst et al., 2022), researchers have begun identifying factors influencing different experiences of loneliness during this time (e.g., Choi et al., 2022; Van Tilburg, 2022). For example, Peng and Roth (2022) reported that inexperience with online socializing likely led to elevated isolation among older adults. It is, therefore, necessary to refine our understanding of the role of internet usage, especially given its growing relevance for intervention and policy. Conceptual accounts, such as the Differential Investment of Resources model (DIRe; Huxhold et al., 2022) and the Social Relationship Expectations Framework (Akhter-Khan et al., 2023), indicate that loneliness and investment in social contact might be moderated by personal characteristics (e.g., age, sex) and resources (e.g., physical capacity). This study, therefore, examines associations among internet usage and loneliness during the first wave(s) of the pandemic, and tests the moderating role of age, gender, and limiting illness.
Internet Use and Loneliness Among Middle and Older-Aged Adults During COVID-19
In recent years, internet use among middle-aged and older populations has increased, representing a promising tool for alleviating loneliness. Indeed, the Internet provides opportunities to develop and maintain a sense of community and support during isolation (review: Khosravi et al., 2016). Despite this, prepandemic reports on the effectiveness of digital technology for loneliness in this demographic have been mixed, including a recent study reporting no associations among the participants used in the present study (Stockwell et al., 2021).
However, dramatic increases in social isolation during the pandemic may have created a testing-the-limits situation that exacerbated the need for social connectedness via digital technology. Such increases in isolation, together with widespread misinformation of health information, may have alternatively created a hostile environment leading to greater feelings of isolation. Indeed, Wallinheimo and Evans (2022) reported that more frequent internet use, especially for e-mail communication but not for (health) informational searches, during the pandemic was associated with lower levels of loneliness among middle and older-aged adults in the United Kingdom (see also Kung & Steptoe, 2023). However, emerging evidence for internet use and loneliness during the pandemic has been mixed (for review, see Ernst et al., 2022), potentially reflecting disparate country-level lockdown measures, small samples, or no consideration of prepandemic trait-report loneliness.
The Role of Age, Gender, and Limiting Illness
Age
Increasing age was associated with elevated mortality throughout the pandemic. Correspondingly, older adults in the United Kingdom were more likely to comply with government-imposed restrictions (e.g., Ganslmeier et al., 2022), implying greater isolation. Conceptual accounts such as the Strength and Vulnerability Integration model (Charles, 2010) suggest that while increasing age is associated with elevated resilience to short-term stressors, older adults may be more psychologically vulnerable to long-term stressor exposure (such as the pandemic), implying greater need for social connection via digital technology. However, older adults typically prefer face-to-face communication (e.g., Stockwell et al., 2021). Indeed, the lack of digital confidence and accessibility typically reported in advanced old age may work to exacerbate feelings of isolation and loneliness during usage.
Gender
Initial evidence indicates that females experienced more loneliness during the pandemic (Wickens et al., 2021), implying a greater need and benefit from digital communication. Females were also more likely to use the internet to search for COVID-19-related information (Wallinheimo & Evans, 2022), underscoring the need to investigate whether such differences in purpose of use modulate experiences of loneliness.
Limiting illness
Those with preexisting life-limiting health conditions were more vulnerable to the consequences of COVID-19 and were more likely to continue to shield even after restrictive measures were lifted in June/July 2020. Emerging evidence indicates that poorer prepandemic health may sensitize people to health threats (Potter et al., 2023), potentially resulting in more cautious socialization and increased loneliness. On the contrary, conceptual accounts of habituation and coping (e.g., Rankin et al., 2009) suggest that those with long-standing life-limiting illnesses may be more familiar with and able to adjust to isolation during the pandemic.
Present Study
The aim of this study was to (1) substantiate initial evidence of associations among internet usage and loneliness in middle-aged and older adults during the first wave(s) of the pandemic and (2) extend such research by testing if these differ by age, gender, and illness. Drawing from previous reports, we expected more frequent internet use to be associated with lower levels of loneliness. Likewise, it was expected that the purpose of internet use would differentially predict levels of loneliness, with lower levels expected for communication purposes and higher levels expected for information-gathering. We also expected that age, gender, and illness would moderate associations between frequency/purpose of internet use and loneliness, but no direction of association was specified due to conflicting theoretical and empirical work.
Method
Covariates were drawn from Wave 9 (2018–2019) of the English Longitudinal Study of Ageing (ELSA) and key study variables from the associated COVID-19 substudy Wave 1 (June/July 2020; see Addario et al., 2020).
Participants
Seven hundred and three participants were excluded for not living in a private household (i.e., hospitalized) or for missing data, leaving a total of 4,790 (Mage = 70.2; standard deviation = 9.0; range: 50–90; 43.5% male). See Supplementary Material for selectivity statistics.
Measures
Internet usage
Frequency was assessed with the following question: ‘Since the coronavirus outbreak, on average, how often did you use the internet or email?’ (1 = more than once a day; 2 = daily; 3 = weekly/monthly; 4 = less than monthly/never). Purpose of internet use was assessed in participants who reported using the internet at least once a month by asking which of the following activities they used the internet for in the past 3 months: (1) emailing, (2) video/voice calling, (3) finding health-related information, (4) finances, (5) shopping/services, (6) social networking, (7) reading newspaper/blog/websites, (8) streaming TV/videos/radio, listening to music, playing games, (9) information about Government services, and (10) none of the above (participants were required to answer yes/no for each purpose). Both frequency and purpose of internet use were treated as categorical variables in analyses to preserve the original measurement scale (i.e., as intervals between response choices are unequal) and to remain consistent with and directly build upon previous studies examining internet use and loneliness with ELSA data (Stockwell et al., 2021; Wallinheimo & Evans, 2022).
Loneliness
Loneliness was assessed with the short form of the Revised University of California Los Angeles (UCLA) Loneliness Scale (Russell, 1996): ‘How often do you feel … (1) you lack companionship (2) left out (3) isolated from others’ answered on a 3-point scale (1 = hardly ever/never; 2 = some of the time; 3 = often). Responses were summed into a total score (range: 3–9) with a higher score indicating greater loneliness.
Covariates
Analyses included: age, gender, ethnicity, education (no qualification; some education; degree level), household composition (partnered; have children; live with others), employment status (active/inactive), household non-pension wealth quintile, and urban/rural living. Following Stockwell et al. (2021), we included limiting illness and depression. Limiting illness was operationalized as the presence of a long-standing limiting illness, disability, or infirmity that limits activity (yes/no), while depression was measured with the eight-item Centre for Epidemiologic Studies Depression Scale (CED-S) which asks participants about the occurrence of depressive symptoms (yes/no) over the past week. Responses were summed into a total score (range: 0–7; see Author Note 1) with participants classified as depressed if they scored ≥3, as per Steptoe et al. (2013). To ensure results do not reflect prepandemic levels, preexisting loneliness was included from Wave 9 ELSA, measured with the Revised UCLA Loneliness Scale.
Statistical Analysis
Data were weighted to account for nonresponse to the COVID-19 substudy, contingent on responding to Wave 9. Pearson’s chi-square analysis was conducted to test for raw differences in key variables by frequency/purpose of internet use. To investigate our research questions, multiple regression analyses were conducted on Stata 16.0, adjusted hierarchically for (1) socio-demographics, (2) physical and mental health (depression and limiting illness), and (3) preexisting loneliness. To test moderation, the interaction term was included in the final adjusted model (Model 4). Following common practice, standard p values were used to determine significance of moderation in models examining purpose of internet use (as each purpose was a binary variable), whereas likelihood ratio tests were used to determine significance of moderation in models examining frequency of internet use (as this variable had more than 2 categories).
Results
Sample characteristics by frequency of internet use are reported in Table 1. As can be seen, a sizable number were classified as lonely–an increase from prepandemic levels (19.4% to 32.4%; see Stockwell et al., 2021). Similar to before the pandemic, the majority were classified as very frequent internet users (more than once a day; 51.2%), while the rest were either frequent users (every/almost every day; 23.8%), moderate users (weekly/monthly; 9.1%), or infrequent users (less than monthly/never; 15.9%). As can be seen in Supplementary Table 2, the most common purpose of use was emailing (89.9%) and shopping (73.9%), consistent with prepandemic reports, whereas video calling increased from before to during the pandemic (26.3% to 62.3%; see Stockwell et al., 2021). Intercorrelations can be found in Supplementary Table 1.
Table 1.
Description of Sample Characteristics Stratified by Frequency of Internet Use
| Characteristics | Frequency of internet use | p Value | ||||
|---|---|---|---|---|---|---|
| V. Frequent | Frequent | Moderate | Infrequent | All | ||
| N (%) | 2,452 (51.2) | 1,141 (23.8) | 435 (9.1) | 762 (15.9) | 4,790 | |
| Loneliness, mean (SD) | 4.1 (1.5) | 4.2 (1.5) | 4.4 (1.6) | 4.1 (1.5) | 4.2 (1.5) | .030 |
| Loneliness, n (%) | ||||||
| Not lonely | 1,701 (69.4) | 759 (66.5) | 260 (59.8) | 519 (68.1) | 3,239 (67.6) | .001 |
| Lonely | 751 (30.6) | 382 (33.5) | 175 (40.2) | 243 (31.9) | 1,551 (32.4) | |
| Loneliness (wave 9), mean (SD) | 3.9 (1.4) | 4.0 (1.4) | 4.3 (1.6) | 4.3 (1.5) | 4.0 (1.4) | <.001 |
| Age in years, mean (SD) | 67.6 (8.4) | 70.6 (8.7) | 72.6 (8.7) | 77.1 (7.8) | 70.2 (9.0) | <.001 |
| Gender, n (%) | ||||||
| Men | 1,147 (46.8) | 479 (42.0) | 160 (36.8) | 299 (39.2) | 2,085 (43.5) | <.001 |
| Women | 1,305 (53.2) | 662 (58.0) | 275 (63.2) | 463 (60.8) | 2,705 (56.5) | |
| Education, n (%) | ||||||
| No qualification | 321 (13.1) | 285 (25.0) | 125 (28.7) | 428 (56.2) | 1,159 (24.2) | <.001 |
| Some education | 811 (33.1) | 428 (37.5) | 147 (40.0) | 204 (26.8) | 1,617 (33.8) | |
| Degree level | 1,320 (53.8) | 428 (37.5) | 136 (31.3) | 130 (17.1) | 2,014 (42.1) | |
| Living arrangement, n (%) | ||||||
| Living alone | 497 (20.3) | 309 (27.1) | 137 (31.5) | 311 (40.8) | 1,254 (26.2) | <.001 |
| Living with others | 1,955 (79.7) | 832 (72.9) | 298 (68.5) | 451 (59.2) | 3,536 (73.8) | |
| Employment status, n (%) | ||||||
| Inactive | 1,492 (60.9) | 801 (70.2) | 338 (77.7) | 690 (90.6) | 3,321 (69.3) | <.001 |
| Active | 960 (39.2) | 340 (29.8) | 97 (22.3) | 72 (9.5) | 1,469 (30.7) | |
| Wealth quintile, n (%) | ||||||
| 1 (Lowest) | 194 (7.9) | 161 (14.1) | 69 (15.9) | 179 (23.5) | 603 (12.6) | <.001 |
| 2 | 319 (13.0) | 207 (18.1) | 91 (20.9) | 180 (23.6) | 797 (16.6) | |
| 3 | 481 (19.6) | 270 (23.7) | 111 (25.5) | 201 (26.4) | 1,063 (22.2) | |
| 4 | 630 (25.7) | 297 (26.0) | 107 (24.6) | 134 (17.6) | 1,168 (24.4) | |
| 5 (Highest) | 828 (33.8) | 206 (18.1) | 57 (13.1) | 68 (8.9) | 1,159 (24.2) | |
| Area, n (%) | ||||||
| Urban | 1,745 (71.2) | 837 (73.4) | 334 (76.8) | 573 (75.2) | 3,489 (72.8) | .027 |
| Rural | 707 (28.8) | 304 (26.6) | 101 (23.2) | 189 (24.8) | 1,301 (27.2) | |
| Ethnicity, n (%) | ||||||
| White | 2,387 (97.4) | 1,093 (95.8) | 414 (95.2) | 742 (97.4) | 4,636 (96.8) | .014 |
| Non-White | 65 (2.7) | 48 (4.2) | 21 (4.8) | 20 (2.6) | 154 (3.2) | |
| Having children, n (%) | ||||||
| No child | 442 (18.0) | 165 (14.5) | 84 (19.3) | 16 (15.2) | 807 (16.9) | .016 |
| Have children | 2,010 (82.0) | 976 (85.5) | 351 (80.7) | 646 (84.8) | 3,982 (83.2) | |
| Marital status, n (%) | ||||||
| Unmarried | 568 (23.2) | 334 (29.3) | 154 (35.4) | 343 (45.0) | 1,399 (29.2) | <.001 |
| Married/living as married | 1,884 (76.8) | 807 (70.7) | 281 (64.6) | 419 (55.0) | 3,391 (70.8) | |
| Depression, n (%) | ||||||
| Not depressed | 1,208 (49.3) | 532 (46.6) | 178 (40.9) | 312 (40.9) | 2,230 (46.6) | <.001 |
| Depressed | 1,244 (50.7) | 609 (53.4) | 257 (59.1) | 450 (59.1) | 2,560 (53.4) | |
| Limiting long-term illness, n (%) | ||||||
| No | 1,850 (75.5) | 751 (65.8) | 269 (61.8) | 421 (55.3) | 3,291 (68.7) | <.001 |
| Yes | 602 (24.6) | 390 (34.2) | 166 (38.2) | 341 (44.8) | 1,499 (31.3) | |
Notes. N = 4,790. V. Frequent = very frequent internet use (more than once a day); frequent internet use (daily); moderate internet use (weekly/monthly); infrequent (less than monthly/never); some education = higher education but no degree; active employment = employed/self-employed/currently working; inactive employment = retired, self-employed but not currently working, unemployed, permanently sick/disabled, looking after home/family. For the purpose of this table, participants were classified as lonely if they scored ≥6 UCLA points, as per Steptoe et al. (2013). Differences in key study variables by frequency of use were analyzed using Pearson’s chi-square analysis.
SD, standard deviation.
Association Among Internet Usage and Loneliness
Results are reported in Table 2. For details of the full multivariate analyses, see Supplementary Table 3 (for frequency of use) and Supplementary Table 4 (for purpose of use).
Table 2.
Hierarchical Regression Analyses for Associations Between Internet Usage (Frequency and Purpose) and Loneliness
| Variable | Lonelinessa | |||
|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 4 | |
| β (95% CI) | β (95% CI) | β (95% CI) | β (95% CI) | |
| Frequency (n = 4,790) | ||||
| Very frequent (more than once a day) | Reference | Reference | Reference | Reference |
| Frequent (daily) | 0.18 (0.03 to 0.33)* | 0.09 (−0.06 to 0.24) | 0.06 (−0.07 to 0.2) | 0.01 (−0.11 to 0.11) |
| Moderate (weekly/monthly) | 0.49 (0.25 to 0.73)*** | 0.30 (0.09 to 0.51)** | 0.23 (0.03 to 0.43)* | 0.01 (−0.16 to 0.17) |
| Infrequent (less than monthly/never) | 0.12 (−0.03 to 0.28) | −0.12 (−0.3 to 0.08) | −0.17 (−0.35 to −0.01) | −0.31 (−0.47 to −0.16)*** |
| Purpose (n = 4,029) | ||||
| Sending/receiving e-mail | −0.20 (−0.43 to 0.03) | −0.12 (−0.33 to 0.09) | −0.05 (−0.25 to 0.15) | 0.09 (−0.08 to 0.26) |
| Making video call/voice calls | −0.16 (−0.3 to −0.01)* | −0.12 (−0.25 to 0.02) | −0.09 (−0.22 to 0.04) | 0.01 (−0.10 to 0.11) |
| Finding information on health-related issues | 0.42 (0.29 to 0.55)*** | 0.33 (0.2 to 0.45)*** | 0.2 (0.08 to 0.32)** | 0.12 (0.02 to 0.23)* |
| Managing finances | −0.01 (−0.15 to 0.12) | −0.01 (−0.14 to 0.11) | 0.01 (−0.1 to 0.13) | 0.05 (−0.05 to 0.15) |
| Shopping/buying goods or services | −0.08 (−0.24 to 0.07) | 0.04 (−0.11 to 0.19) | 0.03 (−0.11 to 0.17) | 0.03 (−0.09 to 0.14) |
| Using social networking sites | 0.11 (−0.02 to 0.24) | 0.08 (−0.04 to 0.20) | 0.07 (−0.04 to 0.19) | 0.10 (−0.01 to 0.20) |
| Reading news/newspaper/blog website | −0.16 (−0.29 to −0.03)* | −0.09 (−0.21 to 0.03) | −0.07 (−0.18 to 0.05) | −0.05 (−0.15 to 0.05) |
| Streaming, music, games, reading | −0.04 (−0.18 to 0.09) | −0.07 (−0.2 to 0.06) | −0.06 (−0.18 to 0.06) | −0.07 (−0.17 to 0.04) |
| Getting information about Government services | −0.08 (−0.21 to 0.05) | −0.05 (−0.17 to 0.08) | −0.08 (−0.20 to 0.04) | −0.03 (−0.13 to 0.07) |
| Others | −0.54 (−1.27 to 0.18) | −0.28 (−0.84 to 0.29) | −0.15 (−0.65 to 0.36) | 0.03 (−0.36 to 0.41) |
Notes. β = unstandardized beta coefficient; 95% confidence intervals are presented in brackets (lower bound, upper bound). The sample size did not differ for different purposes of internet use as participants were required to respond yes/no to every purpose included in the questionnaire. As such, all categories of internet purpose were included in each model. Model 1 = unadjusted; Model 2 = adjusted for covariates (age, gender, education, living arrangement, employment status, SES, area, ethnicity, having children, marital status); Model 3 = further adjusted for health variables (depressive symptoms, limiting long-standing illness); Model 4 = further adjusted for prior levels of loneliness.
aHigher scores represent increased loneliness.
* p < .05.
** p < .01.
*** p < .001.
Frequency of internet use
The unadjusted model (Model 1) indicated that frequent (β = 0.18, p = .016) and moderate internet users (β = 0.49, p < .001) were significantly lonelier than very frequent users. After adjusting for covariates (Model 2), only the association between moderate internet use and loneliness remained significant (β = 0.30, p = .005). This also remained after adjusting for physical and mental health in Model 3 (β = 0.23, p = .026), but not when adjusting for prepandemic loneliness (Model 4). The association between infrequent internet use and loneliness became significant in the final model (Model 4: β = −0.31, p < .001).
Purpose of Internet Use
Although video calling (β = −0.16, p = .038), finding health-related information (β = 0.42, p < .001), and reading the news (β = −0.16, p = .014) were significantly associated with loneliness in the unadjusted model (Model 1), most of these dropped in significance when adjusting for covariates in Model 2. Only searching for health-related information remained significant in Model 3 (β = 0.20, p = .001) and the final model (Model 4; β = 0.12, p = .020).
Association Among Internet Usage and Loneliness: Moderators
Results are reported in Supplementary Table 5 (for frequency of use) and Supplementary Table 6 (for purpose of use).
Age
The association between frequency of internet use and loneliness was significantly moderated by age (likelihood ratio for interaction term: p < .01). As illustrated in Figure 1, loneliness remained relatively stable for very frequent and moderate internet users across adulthood and old age. On the other hand, loneliness increased slightly with age for frequent users. Conversely, loneliness scores dropped with increasing age for infrequent users. No moderation effects were observed for purpose of internet use.
Figure 1.
Association between frequency of internet use and loneliness by age. Notes: As can be seen, reduction in loneliness scores varied by frequency of internet use. While loneliness scores remained relatively stable for very frequent (A) and moderate internet use (C) across adulthood and old age, loneliness scores increased for frequent internet use with increasing age (B). Conversely, loneliness scores dropped dramatically for infrequent internet use with increasing age (D).
Gender
No gender moderation effects were detected for associations between frequency/purpose of internet use and loneliness.
Limiting illness
Limiting illness was not associated with frequency of internet use but did moderate associations between using the internet for health-related information and loneliness (p < .01). As seen in Supplementary Figure 1, using the internet to search for health-related information had a more pronounced negative impact for those with limiting illnesses, denoted by higher loneliness scores compared with nonusers with limiting illnesses.
Discussion
This study targeted internet usage and loneliness among middle-aged and older adults during the first wave(s) of the COVID-19 pandemic in the United Kingdom (June/July 2020) and examined the moderating role of age, gender, and limiting illness. Unsurprisingly, a sizable number were classified as lonely (32.4%) and reported dramatic increases in internet use, especially for making video calls, compared with prepandemic. This implies that digital interventions for social isolation have potential in this demographic. This is also consistent with the interpretation that dramatic increases in pandemic-related isolation exacerbated loneliness and the need for social connectedness via digital technology. However, results suggest that those who used the internet infrequently were less lonely compared with very frequent users–an association that strengthened with age. On the other hand, purpose of internet use was associated with more loneliness: those who used the internet to search for health-related information were lonelier, an association that was stronger among those with limiting illnesses. No evidence arose for the moderating role of gender.
Association Among Internet Usage and Loneliness During COVID-19
Infrequent internet use during the pandemic was unexpectedly associated with lower loneliness, contrary to prepandemic reports (Stockwell et al., 2021) and recent evidence of lower loneliness among frequent internet users in U.K. participants during the pandemic (Wallinheimo & Evans, 2022). Given that the former was conducted prepandemic and the latter did not account for prepandemic loneliness, these results may reflect a unique pandemic effect. That is, results imply that preexisting loneliness may influence the role of internet usage for loneliness during times of sudden or extreme isolation. For example, individuals with lower prepandemic loneliness might already have strong support networks, irrespective of internet use. The added frequency of online interactions during the pandemic will not then have contributed to their overall sense of loneliness. Moreover, infrequent internet users may simply have relied more on offline social connections that were fulfilling and supportive, thereby reducing reliance on online interactions. To substantiate and extend findings, future research should examine post-pandemic associations.
Consistent with prior findings, using the internet to find health-related information was associated with increased loneliness (Wallinheimo & Evans, 2022). It is possible that searching for COVID-related health information heightened perceived pandemic severity, prompting stricter self-isolation (Farooq et al., 2020), resulting in a decline in mental health and well-being (Bu et al., 2021). Findings should be substantiated at later stages of the pandemic when the spread of health misinformation declined.
Association Among Internet Usage and Loneliness: Moderators
Increasing age strengthened the association between infrequent internet use and lower loneliness, possibly reflecting an age-related shift in priorities and the digital divide. Indeed, older adults reported stopping using the internet to communicate because of a lack of perceived emotional feedback compared with in-person socializing (see Age UK, 2020). Moreover, descriptive statistics indicated that while there was a dramatic increase in internet use from before the pandemic, this attenuated with age. In light of considerations from the DIRe model (Huxhold et al., 2022), it might then be the case that older adults invest less time/energy in maintaining online social relationships and instead focus on meaningful emotional exchanges from close ties, potentially resulting in an increased reliance on (pre)existing offline support systems.
Having a limiting illness strengthened associations between searching for health-related information and higher loneliness, consistent with theories of sensitization and previous reports that poorer prepandemic health makes people more psychologically vulnerable during the pandemic (Potter et al., 2023). This finding also implies that those searching for health-related information might have had poorer health and were thus potentially more likely to adhere to isolation mandates and experience loneliness. To explore this further, studies should examine illnesses that denote more vulnerability to COVID-19 (e.g., respiratory disease).
Limitations
Our measure of illness did not capture type, severity, or life domain affected. However, more severe or restrictive illnesses may differentially prompt sensitization or habituation to isolation, making it important to examine more nuanced illnesses. In addition, findings from this cross-sectional design should be interpreted with caution due to possible bidirectionality.
Conclusion
Infrequent internet use was associated with lower levels of loneliness, an association strengthened by age, whereas using the internet to search for health-related information was associated with higher levels, an effect that was stronger among those with a limiting illness. These results indicate that socio-contextual factors modulate the experience of loneliness across adulthood and old age, thereby underscoring the importance of taking the wider social context into consideration. Moreover, results imply that individual differences in the utilization of the internet should be taken into consideration when addressing loneliness and isolation in older population and may thus be relevant to policymakers.
Supplementary Material
Contributor Information
Angelica Vania Hosea, School of Psychology and Vision Sciences, University of Leicester, Leicester, UK.
Claryn S J Kung, Department of Behavioural Science and Health, University College London, London, UK.
Sophie Potter, Department of Psychology, School of Social Science, Heriot-Watt University, Edinburgh, UK.
Andrew Steptoe, Department of Behavioural Science and Health, University College London, London, UK.
Author Note
1. As with previous research (Steptoe et al., 2013), one item was omitted because it overlapped with the loneliness measure (“Much of the time during the past week, you feel lonely”).
Funding
This article reports data from the COVID-19 substudy, which was funded by ESRC via UK Research and Innovation COVID-19 Rapid Response call (ES/V003941/1). ELSA is funded by the National Institute of Aging (R01AG017644), and the UK Government coordinated by the National Institute for Health and Care Research.
Conflict of Interest
None.
Data Availability
Materials are available on request. This study was not pre-registered. The data sets analyzed are available in the UK Data Service repository (Study Numbers 5050; 8688), https://doi.org/10.5255/UKDA-SN-5050-24; https://doi.org/10.5255/UKDA-SN-8688-3.
Author Contributions
A. Hosea conducted the statistical analysis, wrote the manuscript, and revised the manuscript. C. Kung supervised the data analysis and reviewed the manuscript. S. Potter wrote the manuscript and contributed to revising the manuscript. A. Steptoe conceptualized the study, provided access to the data set, and reviewed the final manuscript.
References
- Addario, G., Dangerfield, P., Hussey, D., Pacchioti, B., & Wood, M. (2020). Adapting fieldwork during the COVID-19 outbreak: A methodological overview of the ELSA COVID-19 Substudy (Wave 1). NatCen. https://11a183d6-a312-4f71-829a-79ff4e6fc618.filesusr.com/ugd/540eba_b8627be63ccb4ecd9bdc41b1fd1b7d52.pdf [Google Scholar]
- Age UK. (2020, March). Not like riding a bike: Why some older people stop using the internet. https://www.ageuk.org.uk/globalassets/age-uk/documents/reports-and-publications/lapsed_users_report_march-2020.pdf [Google Scholar]
- Akhter-Khan, S. C., Prina, M., Wong, G. H., Mayston, R., & Li, L. (2023). Understanding and addressing older adults’ loneliness: The social relationship expectations framework. Perspectives on Psychological Science, 18(4), 762–777. 10.1177/17456916221127218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bu, F., Steptoe, A., Mak, H. W., & Fancourt, D. (2021). Time use and mental health in UK adults during an 11-week COVID-19 lockdown: A panel analysis. British Journal of Psychiatry, 219(4), 551–556. 10.1192/bjp.2021.44 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Charles, S. T. (2010). Strength and vulnerability integration: A model of emotional well-being across adulthood. Psychological Bulletin, 136(6), 1068–1091. 10.1037/a0021232 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Choi, E. Y., Farina, M. P., Wu, Q., & Ailshire, J. (2022). COVID-19 social distancing measures and loneliness among older adults. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 77(7), e167–e178. 10.1093/geronb/gbab009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Elder, G. H., Jr. (1995). The life course paradigm: Social change and individual development. In Moen P., G. H.Elder, Jr., & Lüscher K. (Eds.), Examining lives in context: Perspectives on the ecology of human development (pp. 101–139). American Psychological Association. 10.1037/10176-003 [DOI] [Google Scholar]
- Ernst, M., Niederer, D., Werner, A. M., Czaja, S. J., Mikton, C., Ong, A. D., Rosen, T., Brähler, E., & Beutel, M. E. (2022). Loneliness before and during the COVID-19 pandemic: A systematic review with meta-analysis. American Psychologist, 77(5), 660–677. 10.1037/amp0001005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Farooq, A., Laato, S., & Islam, A. (2020). Impact of online information on self-isolation intention during the COVID-19 pandemic: Cross-sectional study. Journal of Medical Internet Research, 22(5), e19128. 10.2196/19128 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ganslmeier, M., Van Parys, J., & Vlandas, T. (2022). Compliance with the first UK Covid-19 lockdown and the compounding effects of weather. Scientific Reports, 12(1), 3821. 10.1038/s41598-022-07857-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Huxhold, O., Fiori, K. L., & Windsor, T. (2022). Rethinking social relationships in adulthood: The differential investment of resources model. Personality and Social Psychology Review, 26(1), 57–82. 10.1177/10888683211067035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Khosravi, P., Rezvani, A., & Wiewiora, A. (2016). The impact of technology on older adults’ social isolation. Computers in Human Behavior, 63, 594–603. 10.1016/j.chb.2016.05.092 [DOI] [Google Scholar]
- Kung, C. S. J., & Steptoe, A. (2023). Changes in Internet use patterns among older adults in England from before to after the outbreak of the COVID-19 pandemic. Scientific Reports, 13(1), 3932. 10.1038/s41598-023-30882-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nelson, N. A., & Bergeman, C. S. (2021). Daily stress processes in a pandemic: The effects of worry, age, and affect. Gerontologist, 61(2), 196–204. 10.1093/geront/gnaa187 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pai, N., & Vella, S. L. (2021). COVID-19 and loneliness: A rapid systematic review. Australian and New Zealand Journal of Psychiatry, 55(12), 1144–1156. 10.1177/00048674211031489 [DOI] [PubMed] [Google Scholar]
- Peng, S., & Roth, A. R. (2022). Social isolation and loneliness before and during the COVID-19 pandemic: A longitudinal study of US adults older than 50. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 77(7), e185–e190. 10.1093/geronb/gbab068 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Potter, S., Duezel, S., Demuth, I., Gerstorf, D., & Drewelies, J. (2023). Context matters: Health sensitivity in the daily lives of older adults living through the COVID-19 pandemic. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 78(6), 1018–1024. 10.1093/geronb/gbad006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rankin, C. H., Abrams, T., Barry, R. J., Bhatnagar, S., Clayton, D. F., Colombo, J., Coppola, G., Geyer, M. A., Glanzman, D. L., Marsland, S., McSweeney, F. K., Wilson, D. A., Wu, C. F., & Thompson, R. F. (2009). Habituation revisited: An updated and revised description of the behavioral characteristics of habituation. Neurobiology of Learning and Memory, 92(2), 135–138. 10.1016/j.nlm.2008.09.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Russell, D. W. (1996). UCLA Loneliness Scale (Version 3): Reliability, validity, and factor structure. Journal of Personality Assessment, 66(1), 20–40. 10.1207/s15327752jpa6601_2 [DOI] [PubMed] [Google Scholar]
- Steptoe, A., Shankar, A., Demakakos, P., & Wardle, J. (2013). Social isolation, loneliness, and all-cause mortality in older men and women. Proceedings of the National Academy of Sciences of the United States of America, 110(15), 5797–5801. 10.1073/pnas.1219686110 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stockwell, S., Stubbs, B., Jackson, S. E., Fisher, A., Yang, L., & Smith, L. (2021). Internet use, social isolation and loneliness in older adults. Ageing & Society, 41(12), 2723–2746. 10.1017/S0144686X20000550 [DOI] [Google Scholar]
- Van Tilburg, T. G. (2022). Emotional, social, and existential loneliness before and during the COVID-19 pandemic: Prevalence and risk factors among Dutch older adults. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 77(7), e179–e184. 10.1093/geronb/gbab101 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wallinheimo, A. S., & Evans, S. L. (2022). Patterns of internet use, and associations with loneliness, amongst middle-aged and older adults during the COVID-19 pandemic. Healthcare (Basel, Switzerland), 10(7), 1179. 10.3390/healthcare10071179 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wickens, C. M., McDonald, A. J., Elton-Marshall, T., Wells, S., Nigatu, Y. T., Jankowicz, D., & Hamilton, H. A. (2021). Loneliness in the COVID-19 pandemic: Associations with age, gender and their interaction. Journal of Psychiatric Research, 136, 103–108. 10.1016/j.jpsychires.2021.01.047 [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
Materials are available on request. This study was not pre-registered. The data sets analyzed are available in the UK Data Service repository (Study Numbers 5050; 8688), https://doi.org/10.5255/UKDA-SN-5050-24; https://doi.org/10.5255/UKDA-SN-8688-3.