This work draws from high-quality data that extend previous research on changes in loneliness from before to during the COVID-19 pandemic. In response to the social distancing measures, one would initially expect an increase in loneliness during the pandemic, when individuals were forced to stay apart. Although there was only a negligible increase in the level of loneliness overall, we observed differences associated with some sociodemographic, psychosocial, and health-related variables. Our study also showed that changes in self-perceptions of social relationships and the presence of depression during the outbreak were strongly associated with increases in loneliness. On the other hand, higher levels of neuroticism were significantly associated with a decrease in loneliness.
The study of Baarck et al. [42] has recently compared the prevalence of frequent loneliness in several European countries during the first months following the COVID-19 outbreak with the one in 2016. The authors, who defined loneliness through a binary indicator, showed that in 2016 around 12% of EU citizens reported feeling frequently lonely whereas, during the pandemic, this share increased to 25%. However, it is worth noting that Spaniards experienced a slightly lower increase as compared with other European countries. Ernst et al. [9], who conducted a meta-analysis of population-based longitudinal studies with pre-COVID-19 data, reported a small increase in loneliness since the start of the pandemic worldwide. The studies included varied in terms of study design, sample, loneliness measurement, and the stringency of the measures taken by governments in reducing the spread of the virus during the COVID-19 assessment, which may account for the heterogeneity of effects. Indeed, the evidence is heterogeneous when it comes to studies like ours. To illustrate, Entringer & Gosling [43] observed a large increase in loneliness during the COVID-19 outbreak as compared to pre-pandemic levels, according to a follow-up cohort study of 3,599 adults from Germany. Bartrés-Faz et al. [13] pointed out that loneliness decreased significantly during the first wave of the pandemic among 1,604 participants aged 40–65 years that were demographically representative of a Spanish cohort. The works of Luchetti et al. [21] and Peng & Roth [44] found no significant changes in loneliness after the issuance of the lockdown measures in two nationwide samples of American adults.
Several explanations may account for these disparate results. Our study focused on the first wave of the COVID-19 pandemic where, despite the physical distancing measures, there might have been a general feeling of closeness and social belonging (i.e., “everyone else is going through the same situation; we must stand together”). For example, to show their gratitude, Spaniards applauded frontline healthcare workers every evening from their windows and balconies [45]. During these weeks of strict lockdown restrictions, most people also found diverse ways to stay connected with friends and family more frequently than before the pandemic, even if it was through virtual contact. Within this context, feelings of loneliness may have appeared as a transient experience triggered by the associated physical distancing measures, that may not have evolved into a problematic level (i.e., become chronic). Differences may also be linked to the societal and cultural context: In collectivistic countries such as Spain, social relationships are thought to be more valuable than in individualistic countries, where people are also more likely to live apart from their families [46]. In addition, our study did not include adolescents or institutionalized individuals, and there was a small representation of the oldest old population (e.g., individuals aged 80 + years old).
That being said, our results showing a maintenance of the level of loneliness at a population level can be difficult to transfer to the individual’s personal context and experiences during the pandemic. It is when we focus on specific groups or individual characteristics that we can observe diverse trends. Therefore, we also aimed to identify at-risk groups of loneliness that may potentially benefit from prevention and intervention. The decreasing trend of loneliness with age during the lockdown accords with previous studies showing that loneliness was significantly more prevalent among younger adults [9]. The fact that many of these individuals are often transitioning through a stage where important decisions are made regarding their social relationships or professional paths, together with the large reduction of opportunities they had to socialise with their peers during the COVID-19 outbreak, probably contributed to the observed trend. In addition, individuals who had never been married showed higher scores of loneliness as compared with those in a romantic relationship as well as an increasing trend of loneliness from before to during the first wave of the COVID-19 outbreak. The same applies to those living alone. Having never married or living alone do not necessarily lead to feelings of loneliness. However, the severe disruption of their social lives may have taken a higher toll on their mental health, as they were uniquely isolated. It can be argued, however, that these sociodemographic variables may have been partially explained by the worsening in social relationships, as observed in multivariate analyses.
Some of the factors that were associated with changes in loneliness do not seem to be specifically linked to the COVID-19 outbreak. That was the case of depression: Research on loneliness and depression has indicated that these variables often co-occur [47]. Other studies have observed that loneliness precedes depression [48] or have pointed out the significance of loneliness as a modifiable risk factor for depression recurrence [49]. Several, but not all, studies have also suggested a bidirectional relationship between both constructs [50]. This long-standing relationship also holds true among COVID-19-related studies, e.g. [17, 18, 22, 51, 52]. Our research group has previously found that higher loneliness scores significantly increased the risk of developing depression during the pandemic, especially among the younger population [53, 54]. Negative social expectations, unhealthy lifestyle behaviours, emotional dysregulation, compromised attentional processes and over-expression of inflammation markers, among other factors, have been proposed as explanations for this association [50, 55, 56]. It is worth noting that we did not find a significant association between changes in loneliness and depression before the pandemic, but we did find it with the presence of depression during the pandemic.
The evidence regarding the association between neuroticism and loneliness is yet scarce. Personality is one of the factors explaining individual differences when facing stress and stressors. Overall, prior research has shown a positive association between neuroticism and loneliness. Neurotic individuals are more likely to react to challenging situations using maladaptive coping strategies, be less satisfied with their social relationships, behave dysfunctionally, be more sensitive to social stimuli (i.e. hypervigilance), and be emotionally unstable [57]. All of these in turn can be associated with feeling of loneliness. It is possible that, during the pandemic, the social situations individuals with high neuroticism use to perceive as threatening (e.g., cues of social rejection) no longer occurred, or occurred differently. Therefore, the negative emotions and consequences associated with social interactions (e.g., avoidance, rumination) could have been reduced as compared with pre-pandemic levels. That being said, our results suggest that these individuals present higher vulnerability to loneliness as compared to people with low neuroticism. In this regard, our findings go in line with those of Ikizer et al. [28] reporting that loneliness during the COVID-19 pandemic was positively related to higher levels of neuroticism, and Entringer & Gosling [58] who found greater increases in loneliness among neurotic individuals using longitudinal data from a nationwide sample of German adults.
4.1 Strengths and limitations
This work presents important strengths. Participants were recruited via a probabilistic sampling method and, therefore, our findings may have excellent external validity, and thus provide highly valuable and actionable knowledge about the changes in loneliness in this period. In addition, this is the first study of its kind among Spaniards and one of the few longitudinal population-based studies including data just prior to the COVID-19 outbreak, and a broad range of variables were mostly based on well-validated questionnaires. Our results should also be interpreted in the light of its limitations. First, asking for the frequency with which loneliness is experienced may not be the most precise way of assessing loneliness severity [59]. In addition, aspects such as intensity and length of time individuals feel lonely could provide more fine-grained information to better capture the changes in loneliness. Notwithstanding the above, the UCLA Loneliness Scale [30] is the most widely used instrument to assess loneliness in epidemiological surveys. Second, a significant number of participants were lost to follow-up, which falls within the commonly reported attrition rate for longitudinal studies [60]. It is then possible that individuals who reported higher loneliness at the pre-COVID-19 assessment and were not included at the follow-up, would have exhibited an increased risk of loneliness during the confinement. Thus, the absolute levels of loneliness may have been underestimated, as van Breen et al. [61] also underlined. Third, as with all COVID-19-related research, we switched from face-to-face interviews at the baseline to telephone interviews during the lockdown period, which may have also affected our results. Forth, recall bias may be present as this study relied on self-report data. Fifth, the time period we covered limited the chances to explore the dynamics of loneliness along the pandemic. Sixth, despite the observational design of this study, causality cannot be inferred. Finally, a word of caution should be made about the changes in social relationships variable, for which different interpretations may have a place (e.g., evaluations in terms of quantity, of frequency, or of quality of the interactions).
4.2 Conclusion
This is a high-quality study assessing changes of loneliness, and the risk and protective factors of these changes from before to during the first wave of the COVID-19 pandemic in Spain. The results of this work did not show changes in the levels of loneliness due to the COVID-19 outbreak overall. However, a more refined analysis highlighted that changes of loneliness were linked to the presence of depression, worsened social relationships, and neuroticism.
Future studies should continue providing evidence for understanding the mid- and longer-term impact of the pandemic on loneliness and mental health. Efforts should be focused on the identification of those who are target in interventions to reduce loneliness. It is indeed everyone's responsibility.