Friendly visiting by a volunteer for reducing loneliness or social isolation in older adults: A systematic review

Research evidence on the effectiveness of face-to-face visiting by a volunteer for improving social isolation and loneliness of older adults is very uncertain

We are unsure if friendly face-to-face visits by a volunteer can improve loneliness, social isolation, depressive feelings, life satisfaction and mental health-related outcomes in older adults. Decisionmakers who are considering friendly face-to-face visiting as a way to alleviate loneliness or social isolation in older adults should take this research uncertainty into account.

What is this review about?

Loneliness and social isolation are reaching epidemic proportions in both children and adults, despite the increasing connectedness in our 21st century world. Given their devastating impact on physical and mental health, it is important to identify and invest in feasible and sustainable options to decrease social isolation and feelings of loneliness.

Friendly face-to-face visiting, where people are matched to someone who visits them in-person on a regular basis, seems to be a realistic and sustainable option for providing social support.

What is the aim of this review?

We wanted to find out if friendly face-to-face visiting by a volunteer is effective at reducing loneliness or social isolation, or both, in adults aged 60 or older. We also wanted to find out if visits can improve depressive symptoms, life satisfaction and mental health outcomes.

What studies are included?

We found 13 relevant studies comparing friendly visiting by a volunteer to no friendly visiting, involving 470 older adults in total. Ten of these studies were conducted in the USA.

Friendly face-to-face visiting programmes ranged from six to 12 weeks in duration and mostly involved weekly visits by undergraduate students. Visits consisted mainly of casual conversation, but playing games and watching TV were also mentioned.

What are the main findings of this review?

The evidence is very uncertain about the effect of friendly face-to-face visiting by a volunteer on improving loneliness, social isolation, depressive symptom experiencing, life satisfaction and mental health-related outcomes in older adults.

None of the studies reported on the long-term effects (more than six months after the friendly visiting programme has ended) on loneliness or social isolation. None of the studies reported on the medium-term effects (1-6 months after the programme has ended) on mental health.

What do the findings of this review mean?

We have very little confidence in the evidence, because the studies were very small, used methods likely to lead to errors in their results, and often did not transparently report all data. Given the limitations of the available evidence, further research is very likely to change the results of our review.

How up to date is this review?

The authors searched for studies up to August 2021.

Description of the condition

The concepts of ‘loneliness’ and ‘social isolation’ have been debated and contested extensively, resulting in myriad definitions. In addition, these terms are often used interchangeably, although they are distinct (though related) concepts. Therefore, defining these concepts and highlighting the distinctions between them is of the essence.

In this systematic review, loneliness is defined as ‘a subjective, unwelcome feeling of lack or loss of companionship. It happens when we have a mismatch between the quantity and quality of social relationships that we have, and those that we want’ (the cognitive deficit model of Perlman, 1981). It is therefore a deeply personal and subjective negative experience.

In contrast, social isolation is an objective state, defined in terms of the quantity of social relationships and contacts. It reflects a reduction in social network size and paucity of social contact, which can be triggered by factors such as mobility impairments, unemployment, or deteriorating health (Steptoe, 2013).

Feeling lonely is therefore different from being socially isolated. In fact, a person may feel lonely even in the presence of other people. Similarly, an individual may live alone without feeling lonely.

Although we live in an increasingly connected world, millions of children and adults suffer from loneliness or social isolation, or both. The Joint Research Centre of the European Commission reported in a 2018 policy brief that 7% of adults in Europe (roughly 30 million people) frequently feel lonely, and 18% (around 75 million people) are socially isolated (i.e., meet socially with friends, relatives or work colleagues at most once a month) (d'Hombres, 2018). A cross-country survey of adults in the United States, the United Kingdom and Japan, performed by the Kaiser Family Foundation in partnership with The Economist, revealed that prevalence rates of loneliness or social isolation lie as high as 22% (US), 23% (UK) and 9% (Japan) (DiJulio, 2018).

Although these estimated prevalence rates themselves have remained stable over the past decades, the burden of loneliness and social isolation is expected to increase even further during the next couple of decades. Population ageing is one of the key contributors: as people grow older, they are at increased risk of living by themselves and of becoming disabled, which in its turn constitutes a barrier to social interaction. In its 2015 evidence review, Age UK stated that 6%–10% of older people say they always or often feel lonely, and that nearly half of the people over 65 say that television or pets are their main form of company (Davidson, 2015).

An increasing number of studies show that loneliness and social isolation can have a detrimental impact on physical and mental health. For instance, they reportedly have the same harmful effect as smoking 15 cigarettes a day (Holt-Lunstad, 2010), and put individuals at greater risk of developing clinical dementia (Holwerda, 2014). In addition, loneliness has been associated with negative psychological effects such as depressed mood, low levels of life satisfaction and happiness (Prince, 1997; Schultz, 1984). These findings highlight the need for effective interventions to tackle loneliness and social isolation.

A growing range of interventions are being developed to alleviate social isolation and loneliness. These include social facilitation interventions (e.g., friendship clubs, shared interest topic groups), psychological therapies (whereby recognized therapeutic approaches are delivered by trained professionals, e.g., mindfulness, reminiscence therapy), health and social care provision (whereby health and/or social care professionals are involved and participants are enroled in a formal care programme, either in a nursing home or in the community setting), animal interventions (e.g., animal-assisted therapy, robotic pets), befriending interventions (a form of social facilitation with the aim of formulating new friendships) and leisure/skill development interventions (e.g., gardening programmes, voluntary work, computer training courses) (reviewed by Gardiner, 2018).

Among the different existing interventions, friendly visiting, a befriending intervention whereby older persons are matched with someone who visits them on a regular basis, seems to be a realistic and sustainable option for providing social support. However, until this day, it remains unclear if friendly visiting by a volunteer is effective at reducing loneliness or social isolation, or both.

Description of the intervention

The intervention of interest for this review is any frequency and any duration of friendly visiting by a volunteer (of any age) to a community-dwelling or institutionalized older adult. We define a volunteer as a person who does something, especially helping other people, willingly and without being forced or paid to do it; either in a formal or informal setting. During these visits, the volunteer engages in friendly talking, playing games and/or reminiscing, with the sole purpose of reducing loneliness, social isolation, depressive symptoms, and/or improving life satisfaction and/or mental health in the older adult.

How the intervention might work

The Model of Depression and Loneliness (MODEL) theoretical framework may offer some insight in how friendly visiting might decrease social isolation and loneliness (Cohen-Mansfield, 2007). Rooted in a cognitive-behavioural theory, MODEL describes the influence of environmental resources, health, stressful life events and psychological factors on loneliness and depression in older adults. According to the framework, older adults experience less opportunities to meet people, may live alone more often and may face limitations in financial resources, making it harder to create new and maintain existing social contacts. Besides these environmental factors that reflect social isolation, health issues and difficulties with mobility represent additional barriers to developing meaningful social ties. Stressful life events such as retirement, deaths of friends and family, and relocation can cause people to lose long-standing social networks, thereby contributing to loneliness. Finally, long-standing reliance on established contacts, little need to initiate new contacts, and decreased social skills may affect their ability to engage in meaningful social relationships.

The MODEL framework was shown to explain 42% of the variance in loneliness and 47% of the variance in depressed affect among low-income older adults (Cohen-Mansfield, 2007), highlighting the importance of developing interventions targeting the modifiable factors that contribute to loneliness, including new contacts, mobility and financial resources.

Friendly visiting programmes are capable of overcoming most of these barriers, as they allow the older adults to meet and maintain social contact with volunteers, regardless of whether they experience any mobility or financial issues. As loneliness has been associated with depressed mood, low levels of life satisfaction and happiness (Prince, 1997; Schultz, 1984), friendly visiting may exert its beneficial effect on these outcomes through that on loneliness.

Why it is important to do this review

Several existing systematic reviews have looked at the effectiveness of interventions aiming to reduce loneliness or social isolation, but either applied a very broad or a rather narrow scope.

In 2017, a systematic review and meta-analysis was published investigating the effectiveness of befriending interventions targeting individuals with distressing physical and mental conditions (Siette, 2017). This review included a wide range of befriending interventions (social support delivery through face-to-face encounters at home, in support groups, or via telephone contact) in a very diverse population of interest (adults of any age with any type of physical or mental condition).

Similarly, another systematic review on the effectiveness of health promotion interventions that target social isolation and loneliness among older people, used a broad scope for its interventions of interest (Cattan, 2005). Studies were categorized as ‘group’, ‘one-to-one’, ‘service provision’ and ‘community development’ interventions. The ‘one-to-one’ category included a wide range of interventions, including home visits by professionals providing health assessments or services, telephone support-therapy by social services, friendly telephone calls by peers, and social support visits by volunteers.

During the development of the 2015 evidence-based guideline ‘Older people: independence and mental wellbeing’ by the National Institute for Health and Care Excellence (NICE, 2015), another very broadly scoped systematic review was developed to investigate the effectiveness of interventions to improve or protect the mental wellbeing and/or independence of older people in the United Kingdom (McDaid, 2015).

Similarly, a recently published integrative review included a wide range of interventions to reduce social isolation and loneliness among older people (Gardiner, 2018).

Also in 2018, the What Works Centre for Wellbeing published an overview of 14 systematic reviews of controlled studies published between 2008 and 2018 looking into the effectiveness of interventions aimed to alleviate loneliness (Victor, 2018). Again, the included studies investigated an extremely diverse range of interventions, delivered either in the community setting or in care homes and residential facilities.

Despite their broad scopes, none of these existing (overviews of) systematic reviews have allowed to make clear statements on the effectiveness of friendly face-to-face visiting by a volunteer to the generalizable older population, that is, older adults that do not suffer from any serious physical or mental illness.

Several other systematic reviews have narrowed the scope of their studied population to adults suffering from chronic non-cancer pain (Cooper, 2014) or older adults who experienced a fall (Tricco, 2022), only looked at interventions delivered by health or social care professionals (Grant, 2014; Montgomery, 2008; Sims-Gould, 2017), or did not investigate the effect of friendly visiting (Franck, 2016; Snowden, 2015). Another recent systematic review focused on the effectiveness of 20 interventions used to combat social isolation, but not loneliness, in older adults (Manjunath, 2021).

In conclusion, the existing systematic reviews highlight the need for a systematic collection, extraction and analysis of studies looking specifically at the effectiveness of friendly visiting by a volunteer on reducing loneliness or social isolation in older, otherwise healthy, adults. In addition, in their overview of reviews, Victor et al. highlighted the need for better reporting of numerical data and a focus on effect sizes and precision rather than using p values as a surrogate for effectiveness, in both future trials and reviews (Victor, 2018).

Loneliness and social isolation are proving to be among the most challenging social issues to our 21st century ageing society. Given their detrimental impact on physical and mental health (and vice versa), policy-makers should invest in effective interventions to reduce loneliness or social isolation, or both. In January 2018, British Prime Minister Theresa May has set the example, by appointing Tracey Crouch as the country's first Minister for Loneliness. Reviews that study the effects of feasible and sustainable interventions, such as friendly visiting by a volunteer, on loneliness, social isolation and wellbeing, may provide useful information to the minister and other governments and organizations that are preparing to face the challenge.

Criteria for considering studies for this review

Search methods for identification of studies

A comprehensive search for eligible published and unpublished studies and reports was performed to reduce the risk of publication bias and identify the best available evidence. No date, location or language restrictions were placed on the searches or included studies.

Data collection and analysis

Description of studies

Results of the search

We identified 1390 references via database searching, 1738 additional records via grey literature and hand-searching, and 293 records through searching other systematic reviews, reference lists and ‘Related articles’ features. After duplicate removal, titles and abstracts of the remaining 3128 records were screened. After full-text screening and resolving disagreements, 18 records on 13 unique studies were included. In addition, we identified one ongoing study (Ninesling, 2018) and eight studies awaiting classification (Al-Khazraji, 1974; Cattan, 2002; Cattan, 2003; ChiCTR1800017915; CTRI/2018/01/011466; NCT03405675; NCT03695133; NCT04224038). Figure 1 illustrates the PRISMA study selection flowchart, including reasons for article exclusion.

OPEN IN VIEWER

Figure 1

PRISMA study selection flowchart. SR: systematic review.

Risk of bias in included studies

For the randomized controlled trials, risk of bias was assessed using the Cochrane Risk of Bias tool (Higgins, 2011). The results are presented in Figures 2 and 3 and are summarized narratively in the sections below. Detailed judgements by domain can be found in the Characteristics of included studies. For the two studies that did not provide any extractable data on any of the primary/secondary outcomes (Keller, 1988; Schulz, 1976–1978), no risk of bias assessment was performed.

OPEN IN VIEWER

Figure 2

Risk of bias graph for the seven randomized controlled trials that provided extractable data on any of the primary/secondary outcomes.

OPEN IN VIEWER

Figure 3

Risk of bias summary for the seven randomized controlled trials that provided extractable data on any of the primary/secondary outcomes.

For the four non-randomized controlled trials (Arthur, 1973; Bogat, 1983; Kahlbaugh, 2011; Mulligan, 1978), the Risk of Bias tool assessment was not applicable and boxes were left blank in Figure 3. For these trials, a ROBINS-I assessment was done instead. The results are presented in a traffic light plot in Figure 4. Detailed judgements, together with the overall risk of bias judgements, can be found in Supporting Information: Appendix 3. In the paragraphs below, we provide a narrative synthesis for the seven domains. For the study of Mulligan (1978), judgements were different across the different outcomes, which explains why there are multiple lines for this study in the traffic light plot in Figure 4.

OPEN IN VIEWER

Figure 4

Traffic light plot ROBINS-I assessment.

Three of the four studies (Bogat, 1983; Kahlbaugh, 2011; Mulligan, 1978) were found to be at serious risk of bias due to confounding due to the overall or partial lack of randomization. The fourth study (Arthur, 1973) did not provide sufficient information to make a proper judgement.

Bias due to selection of participants, bias in classification of interventions, bias due to deviations from intended interventions and bias due to missing data were all judged to be low for three of the four trials (Arthur, 1973; Bogat, 1983; Kahlbaugh, 2011).

Mulligan (1978) was at low risk of bias in classification of interventions and bias due to deviations from intended interventions, but at serious risk of bias due to missing data for the outcome of long-term social isolation. This serious risk of bias resulted from substantial drop-out between the last visit to the 1-year follow-up time point, which was also imbalanced between the intervention and control group. The study provided insufficient information to make a proper judgement on bias due to selection of participants.

All four studies were at serious risk of bias in measurement of outcomes. As friendly visiting was part of the intervention, it was impossible to blind the older adults, who were also often the outcome assessors themselves. Because most of the data were self-reported and subjective, this lack of blinding may have affected the results due to social desirability bias. In addition, in the study by Bogat (1983), outcomes were collected in different manners in the intervention and control group.

Risk of bias in the selection of the reported result was serious for Bogat (1983), and for the outcomes of short-term and long-term social isolation in the study by Mulligan (1978). In Bogat (1983), some of the data were reported as post-test means adjusted for pre-test scores, whereas others were shown as mean changes between post- and pre-test scores. In the Methods section, Mulligan (1978) mentioned the use of two scales for the measurement of social isolation, that is, the Adulthood Isolation Index and the Past Month Isolation Index. However, only one of these measures was reported in the Results section. The risk of bias in the selection of the reported result was low in Arthur (1973), Kahlbaugh (2011) and for the outcome of presence of functional mental disorders in the study by Mulligan (1978).

As all four studies were judged to be at serious risk of bias in at least one domain, but not at critical risk of bias in any domain, a judgement of serious risk of bias was assigned to these non-randomized controlled trials.

Effects of interventions

See also: Summary of findings Table 1 for the short-term impact of friendly visiting by a volunteer; Table 3 for the extracted data; and Table 4 for the synthesis by vote counting based on direction of effect.

Summary of main results

To the best of our knowledge, this is the first systematic review that has collected and synthesized the available data on the effectiveness of volunteers providing friendly face-to-face visits to alleviate loneliness or social isolation, or both, in older adults. We have identified nine randomized and four non-randomized controlled trials that were relevant to elucidate this research question.

At the moment, the evidence is very uncertain about the effect of friendly face-to-face visiting by a volunteer on improving loneliness, social isolation and wellbeing in older adults. Overall, the identified evidence is scarce and of very low certainty, which precludes any conclusions about the added value of friendly face-to-face visiting by a volunteer.

Overall completeness and applicability of evidence

In total, we identified 13 studies (nine randomized and four non-randomized controlled trials) that investigated the impact of friendly face-to-face visiting by a volunteer on our outcomes of interest in older adults. For each of these outcomes, with the exception of life satisfaction (seven studies), the number of studies was limited to one to three. The vast majority of the studies were conducted in the United States (10 studies), included community-dwelling older adults (nine studies) and concerned intergenerational visiting by volunteering undergraduate students (nine studies). Due to the scarcity of evidence for each outcome, we were not able to perform subgroup analyses with regard to geopolicital region, culture, race/ethnicity or housing situation (i.e., community-dwelling vs. institutionalized). Also, we were not able to investigate the possible influences of diversity across gender, the frequency and duration of the visits, or the activities engaged in during the friendly visits (i.e., using interactive materials vs. social interaction only). Regarding the latter, subgroup analyses would have been very difficult nonetheless, due to the lack of clear descriptions of the interventions in the currently available studies. In the Implications for research, we discuss what type of future research is needed, and how it can best be designed and executed, to increase the overall completeness and applicability of the evidence.

Quality of the evidence

The vast majority (5/7) of the randomized controlled trials with usable data were at high risk of performance and detection bias. In addition, six of the seven trials were at unclear risk of selection bias. Therefore, the certainty of the evidence provided by the randomized controlled trials was downgraded by two levels due to risk of bias. Subsequently, the certainty of the evidence was downgraded by one levels for imprecision because of the limited sample sizes, wide confidence intervals and/or lack of data. We did not downgrade further for indirectness, inconsistency or publication bias, resulting in very low-certainty evidence.

With regard to the four non-randomized controlled trials, for each of the outcomes, the certainty of evidence was downgraded by two levels due to serious risk of bias, and by an additional level because of imprecision. We did not downgrade for indirectness, inconsistency or publication bias, resulting in very low-certainty evidence.

Although we acknowledge that non-randomized controlled trials are at larger risk of selection bias (see also Implications for research), we feel that in this case, the problem of inadequate blinding overshadows the lack of randomization. Therefore, we judged that the randomized and the non-randomized controlled trials provide evidence of equal certainty.

Potential biases in the review process

By pre-defining and documenting the review objectives and study eligibility criteria a priori in the published review protocol (Laermans, 2020), we minimized the potential for bias in the review process. In addition, both review authors (JL, HS) were methodologists, and not content experts, which further decreases the risk of bias. However, making decisions to in- or exclude studies on the basis of the aim of the intervention (‘with the sole purpose of reducing loneliness, social isolation, depressive symptoms, and/or improving life satisfaction and/or mental health’) required some judgement on their behalf, as many of the studies did not report this explicitly. It is therefore conceivable that some studies who did not (clearly) report on their specific aim were inappropriately in- or excluded. Nevertheless, we feel that by contacting the study authors in case of missing information, we have done everything possible to avoid this. Moreover, when in serious doubt and attempts to obtain additional information from the authors were unsuccessful, we classified the studies as ongoing.

Although we used a comprehensive search strategy, we might have missed studies because we included outcome-specific search terms (e.g., loneliness, social isolation) in our search strings. Also, whilst we also searched for grey literature, we may have missed unpublished program reports that were not made publicly available by government agencies or non-profit organizations.

Other potential bias in our review might have arisen from the poor reporting in many of the included studies. Often, data were missing, and a fair number of our attempts to contact the authors were unsuccessful because contact details were not available, authors did not respond or data were no longer available. This may have affected the completeness of our data as well as our risk of bias assessment, which may for some studies be harsher than necessary. In addition, the incomplete reporting only allowed us to synthesize the data through vote counting based on the direction of effect. Although this synthesis method is considered acceptable and may be considered superior to a narrative, it is less powerful than methods that combine p-values of studies. The method does not provide information on the magnitude of the effects, and does not account for differences in the relative sizes of the different studies (McKenzie, 2022).

Agreements and disagreements with other studies or reviews

As mentioned in Why it is important to do this review, several existing systematic reviews have looked at the effectiveness of a wide range of interventions aiming to reduce loneliness or social isolation.

In the systematic review of Cattan (2005), ‘one-to-one’ interventions included home visits by professionals providing health assessments or services, telephone support-therapy by social services, friendly telephone calls by peers and social support visits by volunteers. For the majority of these interventions, the reviewers were unable to demonstrate a significant effect in reducing social isolation and loneliness. Although the review only included one study on friendly visiting by a volunteer (Mulligan, 1978), we have reached the same conclusion that the effectiveness of home visiting and befriending schemes remains unclear.

In the McDaid (2015) review, five studies provided moderate evidence that friendship programmes can enhance various aspects of older peoples’ mental wellbeing and address issues of loneliness and isolation. Just one of these five studies fit our eligibility criteria and was included in our review (Lawlor, 2014). The other four were excluded from this review on the basis of study design (Pope, 2013) or intervention (Butler, 2006; Martina, 2006; Stevens, 2006 see Characteristics of excluded studies for the first two). Therefore, it is hard to compare the review results against ours. It shows however, that depending on the choices made on how to cluster interventions, systematic reviewers may reach different conclusions entirely.

Siette (2017) investigated the effectiveness of a wide range of befriending interventions (social support delivery through face-to-face encounters at home, in support groups, or via telephone contact) in a very diverse population of interest (adults of any age with any type of physical or mental condition). The review authors were unable to provide firm conclusions on the effect of befriending on loneliness, depression and quality of life. They concluded that there was moderate-certainty evidence from seven of the 14 studies that showed small improvements in combined patient-reported primary outcomes. Since just two of the 14 included studies fit our eligibility criteria (MacIntyre, 1999; McNeil, 1991–1995), we are not able to point out agreements and disagreements with this review.

In their integrative review, Gardiner (2018) included a wide range of interventions to reduce social isolation and loneliness among older people. The majority of interventions reported some success in reducing social isolation and loneliness, but the quality of evidence was generally weak. None of the four included befriending studies fitted our eligibility criteria.

None of these systematic reviews allowed to make clear statements on the effectiveness of friendly face-to-face visiting by a volunteer to the generalizable older population, that is, older adults that do not suffer from any serious physical or mental illness. Therefore, with our systematic review, we aimed for a systematic collection, extraction and analysis of studies looking specifically at the effectiveness of friendly visiting by a volunteer is effective in reducing loneliness or social isolation, or both, in older, otherwise healthy, adults. Unfortunately, the currently available data did not allow us to make clear statements either. At the root of the problem lie several issues that were already highlighted in an overview of 14 systematic reviews by Victor (2018), that is, the limited number of studies, the low sample sizes and the lack of clear (numerical) reporting, and that are confirmed by our systematic review. In the Implications for research, we discuss how these issues may be tackled in the future.

It will be interesting to see if the ongoing systematic review of Landeiro (2017), on the effectiveness of health promotion interventions on social isolation or loneliness in older people, reaches similar conclusions. In addition, the ongoing systematic review and network meta-analysis by Lee (2021), that aims to determine the comparative efficacy of interventions to alleviate social isolation and loneliness of community-dwelling older adults by comparing direct and indirect interventions, may shed further light on the effectiveness of friendly visiting interventions, and the individual components of these complex interventions, by a volunteer.

Implications for practice

Due to the very low-certainty evidence, we are unsure about the effectiveness of face-to-face friendly visiting by a volunteer with regard to improving loneliness, social isolation, depressive symptom experiencing, life satisfaction and mental health outcomes in older adults. Decision-makers that consider implementing face-to-face friendly visiting as a way to alleviate loneliness or social isolation, or both, in older adults should take into account this uncertainty.

Implications for research

First of all, this review highlights the need for additional randomized controlled trials. Interestingly, a previous meta-analysis investigated whether the success of certain loneliness reduction interventions could be attributed to study design, rather than to the quality of the intervention (Masi, 2011). This revealed that uncontrolled before-after studies and non-randomized controlled trials studies yielded larger mean effect sizes as compared to randomized controlled trials. However, this finding might originate from a combination of regression towards the mean and selection bias, which uncontrolled before-after studies and non-randomized controlled trials are prone to. Therefore, randomized controlled trials, that minimize selection bias and the effect of regression towards the mean, remain the preferred study design.

These randomized controlled trials should include more participants and should be better designed and executed. Special attention should be given to minimizing the risk of social desirability bias, which arises from the fact that blinding the older adults from the visiting intervention is impossible, and that loneliness and life satisfaction are very subjective outcomes. One option is to use third-party, non-involved interviewers. With regard to more objectifiable outcomes, such as social isolation and mental health, researchers should focus on using scales that can be administered by (non-involved) clinicians or professionals to measure these outcomes. Additionally, future research should include the cost-effectiveness of these interventions, and also consider studying any potentially harmful effects of withdrawal from the intervention.

In addition, to increase the applicability of the evidence, more studies should be conducted outside of the United States. It would be interesting to see if studies can detect differences between the effectiveness of the interventions between community-dwelling and institutionalized older adult, as well as across culture, race/ethnicity and gender. Also, additional studies may wish to use different frequencies and durations of the visits, or have volunteers engage in different types of activities during the friendly visits.

Finally, future studies should provide clear information about every aspect of the friendly visiting interventions according to the items in the TIDieR checklist, including the specific aims and planned and actual intervention adherence, to allow reliable implementation and replication of research findings. Also, they should make sure to report all raw data transparently.

On another note, during our review process, we encountered several studies that investigated the effectiveness of providing social support in non-face-to-face ways (e.g., through frequent telephone or video call contact) during the global COVID-19 pandemic. Future systematic review teams may wish to consider collecting evidence on the effectiveness of this type of contact to stay socially connected. This evidence may provide useful information to policy-makers, for example, in dealing with future pandemic outbreaks that require social distancing.