Prevalence of Loneliness in Older Adults: A Scoping Review

,


Introduction
Loneliness-the discrepancy between a person's desired and actual level of social contact-can impact people of all ages [1].A study among the general population in the US showed that a high level of loneliness is observed among young adults (<30 years), middle age (50-60 years), and very old age (>80 years) [2].In the literature, a U-shaped association between age and loneliness is often described where the highest level of loneliness is observed in adolescence/ emerging adulthood and the oldest old [3,4].Moreover, loneliness is particularly marked in more vulnerable populations, such as those with chronic illnesses, impaired mobility, and declining economic resources (all of which are associated with advanced age) [5].Te impact of the COVID-19 pandemic and associated precautionary measures, such as the restriction of contact, emphasized the importance of addressing loneliness in the specifc population of older people.
We can distinguish two kinds of loneliness: (1) relational (or social) loneliness, which is associated with a small social network, and (2) emotional loneliness, resulting from the lack of an intimate relationship [1].Tere is also a distinction to make between loneliness and social isolation: loneliness is a subjective feeling related to the unpleasant lack of relationships whereas social isolation is objectifed by contact frequency and could be defned by an objective state of having minimal social contact with other individuals [1].
Terefore, loneliness can appear without social isolation: for example, "I have three professionals every day at my home and my daughter comes several times per week but despite that I feel lonely," and social isolation does not lead automatically to loneliness: for example, "I'm alone all the week but my daughter comes the week end and this one visit is enough for me not to feel alone" [6].
Loneliness is linked with multiple negative consequences.Longitudinal studies show that loneliness is associated with depressive symptoms, sleep fragmentation, higher blood pressure, dysregulation of the autonomic nervous system (hypothalamic-pituitary-adrenocortical axis activity), cognitive decline, and even mortality [7,8].In a meta-analytic review, it was found that in lonely people across all age groups, the likelihood of death was 26% higher than that in people who did not feel lonely [9].Tis was also observed in older adults specifcally [1,10].Tese deleterious consequences could be explained by the observation that people who feel lonely are more likely to engage in detrimental health behaviors such as more smoking or less physical activity [1].
In view of its prevalence and consequences, loneliness is considered a public health concern [9].Te prevalence estimates vary with the year of measurement, the tool which is used to measure loneliness, the country or region where it has been measured, and also the population [1].Tis raises the necessity to comprehensively map the prevalence of loneliness while taking these diferent elements into account, including the impact of COVID-19. .Tis area stretches out from Leuven in the west to the borders of Cologne in the east and runs from Eindhoven in the north all the way down to the border of Luxemburg.Over 5.5 million people live in this cross-border region, where the best of three countries merge into a truly European culture.With the investment of EU funds in Interreg projects, the European Union directly invests in the economic development, innovation, territorial development, social inclusion, and education of this region) concerning the loneliness of older people.Te EMR is a border region covering parts of Belgium, Germany, and the Netherlands, which contains 150 municipalities.A municipality refers to a town or a district that has a local government.It is estimated that nearly 20% of the inhabitants of the EMR are older than 65 years [11], and this number is expected to increase in the upcoming years.Tis project ran from September 2021 until August 2023.It focuses on preventing and combating loneliness in older people living in the EMR and increasing awareness of the efects of loneliness.Terefore, this scoping review aimed at examining the prevalence of loneliness.Both scientifc literature and grey literature were included in order to have a global idea of loneliness through the world and public data sources to analyze more specifcally loneliness in the EMR (due to the place of our Interreg project, inside this region).(4) data had to be collected in or after 2012.Exclusion criteria were synthesized articles or systematic reviews.

Study Selection and Data Extraction.
Abstracts were screened by two authors independently (S.Schroyen and S. Adam).In case of discrepancies, these were discussed, and a consensus was reached between the authors.Full texts were examined, and we extracted data such as type of population, age of the sample, number of participants, place of the study, year of data collection, measure of loneliness, and prevalence estimates.If some information was missing, we wrote to the authors of the article, and if no answers were obtained, we excluded the article (n � 4).

Search Strategy for the Grey Literature (Focused on Belgium, Te Netherlands, and Germany
). Te research on the grey literature was performed in French, German, and Health & Social Care in the Community Dutch given that these are the main languages in the Euregio Meuse-Rhine, and the data are used in the context of the euPrevent PROFILE project.For the grey literature search, we only included results from Belgium, the Netherlands, and Germany whereas results from all over the world have been included in the white literature search.Te same keywords and inclusion criteria were applied.Grey literature was searched by using Google turning the incognito mode on and examining the frst fve result pages.

Results
A total of 391 articles were identifed.After screening the abstracts, 164 articles remain.After the screening of full texts, 37 articles were included (see Figure 1).Of the 126 articles excluded, six were excluded because of the languages, 97 were excluded because the sample was too young, 16 were excluded because data were collected before 2012, three were excluded from a hetero-assessment of loneliness, and 4 from a lack of information (e.g., the year of data collection or the number of participants).Of the 37 articles remaining, 19 refer to the general population before COVID-19 and well described in Table 1, 13 to specifc populations before COVID-19 (Table 2), and 5 to the general population after COVID-19 (Table 3).

White Literature.
As shown in Figure 2 and Table 1, before COVID-19, prevalence estimates of loneliness appeared to be the highest in China and Nigeria (35.3% and 46.0%, respectively) [12,13].A high level of loneliness (47.1%) was also observed in Sweden in a sample of the oldest old (85+ years old, based on one article) [14].Te lowest prevalence of loneliness was observed in Australia (5%), Korea (5.4%), and New Zealand (9.3%) [15][16][17].As we can observe in Figure 2, in Europe, the lowest prevalence of loneliness is in the north.
In the following table, the prevalence of loneliness before COVID-19, among the general population, is described.For clearer visibility, data are sorted in ascending order of prevalence.
Concerning specifc populations, fve categories were formed post hoc (see Table 2): (1) specifc communities (e.g., retirement village), (2) impairment/diseases, (3) nursing homes, (4) minorities, and (5) rural areas.Overall, prevalence estimates show to be higher as compared to the general population.In specifc communities (retirement village in New Zealand, community in Nepal, and independent living sector of a senior housing community in the United States), the prevalence of loneliness varies from 37.4% to 85% [18][19][20].
Among residents of nursing homes, prevalence measured in Singapore, Norway, and China, ranged from 27.6% to 59.6% [26][27][28].Te ethnic minorities are only studied in one research (Chinese in Chicago) and show a prevalence of loneliness of 25.8% [29].Only one study was conducted in a rural area in China and found a loneliness prevalence of 25% among older people.
In the following table, prevalence of loneliness before COVID-19, among specifc population, is described.For clearer visibility, data are sorted in ascending order of prevalence.
Tree studies reported prevalence data for all age categories [16,17,21].Two of them include the general population (New Zealand and Korea, see Table 1): the frst one, in New Zealand, indicates a lower prevalence estimate of loneliness with higher age [17].Prevalence of loneliness was 22.6% for 18-30 years old, 14.7% for 31-45 years old, 13.8% for 46-60 years old 8.4% for 61-75 years old, and 10.2% for those who are 75 years and older.In comparison, the study in Korea showed a slightly higher prevalence of loneliness when younger ones are compared to older ones (4% for 15-29 years old, 2.6% for 30-44 years old, 3.4% for 45-59 years old, and 5.4% for 60-74 years old) [16].Te third study analyzed loneliness among a specifc population (patients receiving mental healthcare, see Table 2) and shows a lower prevalence of loneliness when younger and older are compared (52.2% in 18-35 years old, 58% in 36-50 years old, 42% in 51-65 years old, and 43.4% in 66 years old and more) [21].
Te prevalence of loneliness for older people during COVID-19 seems to be higher than before the pandemic (see Table 3): it varies from 22.0% to 59.3% according to studies [30][31][32][33][34].One study shows a comparison during/after the pandemic, and an increase in loneliness is also observed [31].
In the following table, the prevalence of loneliness during/after COVID-19, among the general population, is described.For clearer visibility, data are sorted in ascending order of prevalence.

Grey Literature.
A grey literature search was conducted to identify the prevalence of loneliness inside the EMR.Terefore, the comparison between Belgium, Netherlands, and Germany (see Table 4) shows that loneliness percentages were the lowest in Germany (between 7.5% and 7.9% for people aged 65 and older before COVID-19) [35,36].In Belgium, the prevalence of loneliness was between 12% and 15% in seniors over 65 years old [37], and in the Netherlands, the numbers showed to be even higher (between 32% and 41%) [38,39].During the pandemic, the prevalence of loneliness has been rising in each country (between 8.7% and 22.1% in Germany, between 20% and 22% in Belgium, and between 44% and 65% in the Netherlands) [36][37][38].
In the following table, the prevalence of loneliness in the Euregio (grey literature) among the general population is described.For clearer visibility, data are sorted in ascending order of prevalence.

Discussion
Te objective of this scoping review was to provide an overview of the prevalence of loneliness during/after COVID-19 among older people, including specifc populations, globally (white literature) and with a focus on inside the EMR (grey literature).In the context of the grey literature, it is important to note that the median age of the population living in Europe is the highest in the world (WHO, Strategy and Action plan 2012-2020).
Te prevalence of loneliness among older people from the general population before COVID-19 is the highest in China and Nigeria (respectively 35.3% and 46%) [12,13] and the lowest in Australia (5%) [15], Korea (5.4%) [16], and New Zealand (9.3%) [17].In Europe, the lowest prevalence of loneliness is in the north: this observation is confrmed in a recent meta-analysis across 113 countries: more precisely, they found that for all adult age groups, including older adults, the prevalence of loneliness was the lowest in northern Europe and the highest in eastern Europe [40].Across countries, levels of loneliness were globally higher when looking at specifc populations, than looking at the general population, namely, varied between 25% and 85% depending on the subpopulation [20,41].One important fnding is that the prevalence data of many other countries are lacking in the literature (mainly countries in South America and Africa).Te heterogeneity observed between studies could be explained by the tools used to assess loneliness and the criteria, for example, the assessment of loneliness as a unidimensional or multidimensional construct, or a cutof score depending on the scale or depending on the mean score of the sample.It could also be explained by the country, the population selected, the mean age, and the period of data recollection.All these diferences make a comparison between studies hazardous.In order to explain diferences between countries concerning the level of loneliness, several explanations can be found as for example, the social-cultural and historical-political characteristics [42]: indeed, higher levels of loneliness are observed in countries where there is a social disengagement leading to a low level of trust in other people (observed in posttotalitarian countries).Other explanations could be demographic composition as gender diferences in life expectancy [43] or cultural diferences in relationship expectations: when living alone is unexpected (e.g., collectivist culture), people living alone are more likely to feel lonely (in comparison to people who live in a country where living alone is expected) [44].
During the COVID-19 pandemic, a lot of countries have taken lockdown measures; therefore, loneliness for older adults was a central preoccupation [45].We observed that levels of loneliness were higher in studies who measured the prevalence of loneliness during the COVID-19 pandemic although still very heterogeneous between studies (variation from 23.4% to 59.28%) [30,33].Te increase in loneliness is confrmed across age groups in general meta-analysis as well as the great heterogeneity between studies [46].For older people more specifcally, a greater vulnerability is observed [47].Concerning the pandemic, one study showed that people who sufered from loneliness before the pandemic were also at higher risk of developing depression: in other words, loneliness was a risk factor for a deterioration of mental health during the period of COVID-19 [48].Terefore, people who feel already lonely may be vulnerable to further detrimental consequences during a pandemic, such as COVID-19.
When analyzing the grey literature of countries that belong to the Euregio, the prevalence of loneliness is the lowest in Germany (between 7.5% and 7.9%) [35,36] and the highest in the Netherlands (between 32% and 41%) [38,39].Concerning Belgium, we found a mean prevalence of 13.5% in the grey literature: nevertheless, it is important to note that another study observed that 22% of people in Belgium felt lonely in 2017 [49].Tis last study was not included in our research as it is written in English (we only include research in French, Dutch, or German concerning the grey  Health & Social Care in the Community  "sometimes," or "always" (vs "never") literature).However, even considering these higher numbers, the prevalence of loneliness in Belgium is still higher than the ones observed in Germany and lower in comparison to the Netherlands.Another study (not included as the number of participants is not indicated) nuanced our results as they observed that loneliness in the Netherlands increased after 2011 but was lower than in Belgium (in 2013, data indicated a level of loneliness of 22% in the Netherlands, 25% in Belgium, and 16.8% in Germany) [50].In order to explain the diferences between these three countries, we can look at the tools used to assess loneliness: in the Netherlands and Germany, the same scale is used (DjG scale) but not the same cutof score (2.5 in Germany, whereas a cutof score of 2 is used in the Netherlands); therefore, a higher estimate of loneliness in the Netherlands is logic (as a lower score, in comparison to Germany, is needed to be categorized as lonely).In Belgium, a single-question scale is used.Another methodological issue is that response rates vary across countries and that could be problematic if nonresponse is associated with loneliness [51].Other hypotheses could be raised: individual characteristics of participants (as civil status, socio-economic level, institutionalization. ..), historical, cultural, and social characteristics of countries (such as ageism, solidarity between generations or social security systems, level of trust in other people, and rates of mobility and migration) [42,49,51].8 Health & Social Care in the Community  >10 000 (no exact number) (1) 8.7 (1) 80-84 (2) 14.5 (2) 85-89 (3) 22.1 (3) 90+ Health & Social Care in the Community Considering the relatively high prevalence of loneliness in some countries, it is necessary to think about potential interventions.Broadly, interventions can emerge at four levels: individual, relationships, community, and societal [52].At the individual level, meta-analysis identifes two successful strategies [53]: (1) improving social skills (e.g., improving conversation skills) and (2) addressing maladaptive social cognition through cognitive behavioral therapy (e.g., teaching people to identify automatic negative thoughts and regard them as hypotheses rather than facts).At the relationship/ community level, we found, for example, some initiatives with the aim to bring social connections to people (for example weekly phone calls and home visits to lonely older adults) [52].On the societal level, an example could be educational public awareness campaigns to increase awareness of loneliness and promote positive social behaviors (for instance, there is a week of loneliness/solidarity in Belgium and the Netherlands to sensitize the general population to this topic) [52].From a global perspective, many interventions to reduce loneliness among older people have been developed, but we cannot develop a standardized approach suitable for everyone: interventions need to be individualized or adapted to specifc groups, depending on the context, the population, or the degree of loneliness and individuals' needs [54,55].Te suboptimal success of current approaches lies not in the interventions per se, but in the lack of integration and adjusting particular interventions to "the right person, at the right time" [54].
A strength of this scoping review is to include the older general population as well as older specifc populations.Moreover, data before and after COVID-19 were observed.Nevertheless, there are missing data from a lot of countries.As we limit our research to papers published in English, French, Dutch, or German, it is possible that some relevant studies were omitted.Concerning the grey literature, we limit our search to countries inside the EMR.

Conclusion
Large diferences in prevalence estimates are observed between countries and populations studied.Furthermore, data for South America and Africa are lacking.When focusing on the Euregio Meuse-Rhine, the lowest level of loneliness is observed in Germany (in comparison to Belgium and the Netherlands).Without surprise, the pandemic (and measures associated) increases the level of loneliness all around the world.In order to better compare countries, the use of a standardized measure would be a frst step, preferably by using a validated instrument such as the Jong Gierveld Scale and also allowing to make a further distinction between social and emotional loneliness.Using standardized measurements and implementing routine monitoring will in turn enhance cross-country comparison, examine trends over time, and being able to better inform public policy on addressing loneliness (e.g., policy on household help to decrease loneliness, sensibilization of general population about loneliness, and focusing on individual factors linked to loneliness such as poverty).Also, a recommendation for future research would be to investigate cross-country diferences (as social-cultural or historical-political characteristics) in order to be able to better understand diferences in reported loneliness.With a better comparison, we could be able to learn about each other more efciently and to reduce the prevalence of loneliness.Tese analyses are needed to better understand the heterogeneity of loneliness that we can observe in our review and avoid erroneous and/ or too hasty interpretations.

2. 1 .
Study Setting.Tis study took place within the euPrevent PROFILE project realized within the INTERREG Euregio Meuse-Rhine (EMR) program (Te Interreg V-A Euregio Meuse-Rhine (EMR) program invests almost EUR 100 million in the development of the Interreg-region until 2020

1 Figure 1 :
Figure 1: Flowchart of the review process.

Figure 2 :
Figure 2: World map showing loneliness estimates including data between 2010 and 2019.
frequently did you feel lonely in the past week?":"most of the time or "always, almost always" (vs."never or almost never" or "sometimes")

Table 4 :
Prevalence of loneliness in the Euregio (grey literature).