Ella Cohn-Schwartz, Michal Levinsky, and Howard Litwin have written an important paper on the relationship between social network type and cognitive health. Their study underlines the importance of social networks in later life and yields important results in view of the increasing aging population in West European countries. Especially as cognitive decline is one of the major factors that determine whether older persons can live independently in later life.

Drawing on two waves of the Ageing, Health and Retirement in Europe (SHARE) survey, Cohn-Schwartz et al. (Reference Cohn-Schwartz, Levinsky and Litwin2020) offer a systematic analysis of this relationship based on five network types retrieved by means of latent class analysis. The SHARE survey offers two compelling advantages. First, the multiple waves enabled the authors to include a longitudinal dimension, regressing follow-up data on cognitive function on the network types. Second, the survey uses a “smart” approach to measuring functioning networks, asking in the first instance to name “confidants” with whom they discussed important matters. Only in the second instance were respondents asked to indicate whether these confidants are family, wider relatives, or friends. Cohn-Schwartz et al. (Reference Cohn-Schwartz, Levinsky and Litwin2020) distinguish a “friend-enhanced” network, with relatively high shares of friends, and a lower chance to be married or have children. In addition, there are “close family” (consisting mostly of spouses and children), “family poor” (consisting mostly of wider relatives), and “family rich” (spouses, children and wider relatives) networks. Finally, “multi-tie” networks consist of ties with both friends and family.

Almost 50% of the respondents belong to the “close family” network type. The two network types where friends play a meaningful role (“friend-enhanced” and “multi-tie”) are also characterized by a higher likelihood of performing volunteer work. Both of these networks appear to nourish cognition rather well, especially the “multi-tie”, diverse network. This is because they challenge older people to continue to navigate a greater diversity of social stimuli. The lowest cognition was found in respondents with “family poor” networks. Having such a network correlates with older age, low education, and poor health. This correlation is not elaborated in detail in the paper. We can imagine that older age may imply a higher likelihood of having lost your spouse, and that older age, low education, and poor health together may lower the ability to improve the quality of contact with wider family. In hindsight, the term “family poor” may be a bit unfortunate in the sense that low cognition seems to be related mostly to the lack of having close family, rather than having only wider family.

In the remainder of this commentary, we would like to reflect on the implications of these two important findings: the role of friends and the low cognition found in “family-poor” networks.

To start with the role of friends, this finding makes for an interesting comparison with trend research on demographic change and informal care. In the Netherlands, all but the most vulnerable older people are encouraged to “age in place” in cities and local communities by mobilizing informal care networks. With growing numbers of people aging in place, also the numbers of informal carers will need to grow (De Boer et al., Reference Boer de, Plaisier and De Klerk2019). Yet, prognoses show a decrease in potential informal caregivers in the future (Kooiker et al., Reference Kooiker, De jonge, Verbeek-Oudijk and De Boer2019). In addition, it is also predicted that more people will live further away from family (Kooiker et al., Reference Kooiker, De jonge, Verbeek-Oudijk and De Boer2019) and/or will live part of their adult life outside of traditional family structures as singles, single parents, partners without children, or combined families with children from both partners (e.g. Van Duin et al., Reference Van Duin, Te Riele and Stoeldraijer2018).

The question then is whether, and to what extent, this informal care can be provided by friends from existing social networks. A tentative answer to this question is provided in the literature concerning lesbian, gay, bisexual, and transgender (LGBT) aging. The research by Cohn-Schwartz et al. (Reference Cohn-Schwartz, Levinsky and Litwin2020) concentrates on current networks in Western societies in which friends are seen as an “extra” next to traditional family relationships. Interestingly, the literature on networks of older LGBT people gives insight into networks that are mostly built on acquaintances and friends instead of traditional family networks, partly because LGBT persons less often have a partner and children (Hawthorne et al., Reference Hawthorne, Camic and Rimes2020). Based on sets of questions in cross-sectional needs assessment surveys (e.g. Croghan et al., Reference Croghan, Moone and Olson2014; Orel, Reference Orel2014), this literature shows that friends appear to take on less intensive forms of informal care, although this has not systematically been investigated. The (scarce) evidence on older LGBT people with early dementia even suggests that friends may actually drop out from social networks altogether with increased cognitive decline (Fredriksen-Goldsen et al., Reference Fredriksen-Goldsen, Jen, Bryan and Goldsen2018).

In sum, although friend-oriented networks are related to better cognitive health, there is a need for future research to further explore and more systematically analyze how the benefits of social networks composed partly or mainly of friends can be maintained when the care needs grow stronger, especially insituations where cognitive decline occurs. In this respect, the authors rightly mention there can also be a reverse relationship between cognitive functioning and the functioning of social networks (Cohn-Schwartz et al., Reference Cohn-Schwartz, Levinsky and Litwin2020). This area for future research would thus take up the challenge to understand informal care provided by friends in relation to the composition of people’s social networks, and the trajectories through which people become (more) engaged in informal care (Lapierre and Keating, Reference Lapierre and Keating2013).

Another trend distinguished in the aforementioned demographic trend research in the Netherlands is that in the future, more people will be part of so-called “alternative household forms” (e.g. Jansen et al., Reference Jansen, Stavenuiter, Dijkhuis, Van Dongen and Van Tricht2008). This term points to the growing interesting in communal housing and co-housing projects, including intergenerational living arrangements. It is interesting to note that, in the Northern European context of the Netherlands, intergenerational living is seen as an “alternative household form”, whereas in Southern Europe, this is a common way of living together (Cohn-Schwartz et al., Reference Cohn-Schwartz, Levinsky and Litwin2020). We very much support the authors in their proposal to take on board household composition in future research on social networks and cognitive functioning. Aranda (Reference Aranda2015) investigated the impact of young adults moving back in with their parents due to economic insecurity on the mental health of the parents. Interestingly, this study found the older generation in countries with a Catholic tradition showed less signs of depression, while the mental health of the older generation in countries with a Protestant tradition was not affected. Research on household composition should thus also focus on other resources that may impact social networks in households, including, for instance, health and income Girardian et al. Reference Girardian, Widmer, Connidis, Castrén, Gouveia and Masotti(2020).

To support older people with “family poor” networks, the authors recommend mental health professionals to try to increase social involvement (Cohn-Schwartz et al., Reference Cohn-Schwartz, Levinsky and Litwin2020). While we fully concur with this recommendation, we also suggest a realistic approach: for many of older persons with these networks, moving toward another network type will be extremely difficult. Therefore, we suggest to start encouraging social involvement already in “younger” cohorts of older people. At the same time, the scope of support to frail older people could be broadened by also focusing on the direct living environment. In human geography and sociology, the relationship between social–environmental embeddedness and feelings of belonging and inclusion has been widely studied, e.g. through the role of daily routines and rhythms in the neighborhood (e.g. Lager et al., Reference Lager, Van Hoven and Huigen2016; Van Eck and Pijpers, Reference Van Eck and Pijpers2017), and the meaning of “therapeutic landscapes” (Gesler, Reference Gesler1992), such as communal gardens (Milligan et al., Reference Milligan, Gatrell and Bingley2004). Opportunities to meet people and use multipurpose sites foster contacts and intergenerational connections (Phillipson, Reference Phillipson, Angel and Settersten2011). Therefore, it may be argued that improving local surroundings may also influence the quality of social networks for these specific older persons. This is in line with the ideas behind the development of dementia-friendly communities (Lin and Lewis, Reference Lin and Lewis2015), and age-friendly cities (Van Hoof et al., Reference Van Hoof, Kazak, Perek-Białas and Peek2018).

These understandings of cognitive health, social networks, and the use of the local environment would be an important addition to the finding of Cohn-Schwarz, Levinsky, and Litwin. Nevertheless, they provide highly needed scientific and policy-relevant insights into the social dimension of aging in Western Europe.