People with intellectual disabilities living in care facilities engaging in virtual social contact: A systematic review of the feasibility and effects on well‐being

Abstract Background During the initial phase of the COVID‐19 pandemic, many people with disabilities living in home care facilities could not receive visitors. The use of virtual social contact has been recommended by health authorities. This systematic review examined the scientific evidence of the use and feasibility of information and communication technology (ICT) for social contact by people with intellectual disabilities living in care facilities, and potential effects on well‐being. Methods Five databases were searched using traditional systematic screening and machine‐learning supported screening. Findings are presented in a narrative synthesis using thematic analysis. Results Nine studies were included. We described three themes: means of ICT used for social contact; effects on well‐being; and benefits, barriers, and preconditions. Conclusions Engaging in virtual social contact may be feasible for people with severe to mild intellectual disabilities, but there is little concrete evidence that this can be used as an alternative for in‐person contact.


| INTRODUCTION
To limit the spread of the SARS-CoV-2 virus and COVID-19, national governments took drastic measures over the course of 2020, such as restriction of travel, closure of schools, and the advice to quarantine in case of exposure or infection. Besides measures for the general population, actions were taken internationally to protect vulnerable people from getting infected. Among other measures, restrictions were placed on visits to long-term care facilities, such as those for people with disabilities (Salcher-Konrad et al., 2020;World Health Organization, 2020). Many people with disabilities living in home care facilities could therefore not receive visitors or visit anyone from outside their facility. But especially during this time of anxiety and disrupted routines, residents' need for contact with loved ones would likely have been heightened Schuengel et al., 2013). Adults with mild intellectual disability indicated that they missed having in-person contact with family and friends during the first COVID-19 lockdown period in the Netherlands . In addition, family and friends often play a significant role in the emotional and practical support of people with intellectual disabilities (Giesbers et al., 2019;Van Asselt-Goverts et al., 2013), also during the pandemic (Redquest et al., 2020). Given the potential impact of the pandemic measures on people with intellectual disabilities living in care facilities (Schuengel et al., 2020), it may be important for their well-being to maintain contact with friends and family.
The use of virtual social contact, such as telephoning and video conferencing, has been widely recommended as an alternative way to stay in contact during the pandemic (e.g., World Health Organization, 2020). Previous studies have discussed the use of information and communication technology (ICT) by people with intellectual disabilities (e.g., Chadwick et al., 2013), but research focusing on ICT for social contact is limited (see Barlott et al., 2020). People with intellectual disabilities increasingly use the internet, but may experience more barriers with using it than people without intellectual disabilities (Chadwick et al., 2013). On the other hand, the use of ICT may positively impact the well-being of people with intellectual disabilities (e.g., Caton & Chapman, 2016;Den Brok & Sterkenburg, 2015;Dyzel et al., 2020;Oudshoorn et al., 2020). The aim of this systematic review was to gain insight into the available scientific evidence on the use and feasibility of virtual social contact by people with intellectual disabilities living in home care facilities, and the potential to mitigate the impact of restricted in-person visiting policies. The following questions oriented our selection and extraction of the literature: • Which ICT-mediated means for social contact are available to people with disabilities living in home care facilities or supported living arrangements, and are being used by people with disabilities for social contact?
• What are the effects of using ICT for social contact on wellbeing, quality of life, and quality of interpersonal relationships in the context of restricted in-person visiting policies?
• How do people with disabilities, family and other social network members and professionals perceive the benefits, barriers, and preconditions of using ICT for social contact?
• Are use, effects, and perceived benefits, barriers and preconditions dif-

| Eligibility criteria
Inclusion criteria were specified using the PICO format (i.e., population, intervention, comparison, outcome; Liberati et al., 2009): (a) the population included children, adolescents, and adults with intellectual disabilities (borderline, mild, moderate, severe, profound), motor disability, autism, visual impairment, auditory impairment, deaf-blindness, acquired brain injury, and multiple disability, living in care facilities; broad population criteria were used because we expected few relevant studies and because the findings were expected to be relevant for all people with (intellectual) disabilities living in care facilities; (b) interventions included any form of ICT-mediated social contact between a person with disability living in a care facility with another person who lives outside the care facility, occurring through different modalities, such as video, audio, photo sharing, and text messaging; (c) ICT-mediated contact was used as an alternative for in-person contact, while in-person contact was restricted by (emergency) policy of the government or care facility; (d) outcomes included wellbeing, quality of interpersonal relationships, quality of life, and perceived benefits, barriers, and preconditions of using ICT. During the screening, we dropped part of the criterion under (c) that 'in-person contact is restricted by (emergency) policy of the government or care facility', because it applied to none of the otherwise eligible studies.
Studies were included if targeted at people living in home care facilities. We used a broad operationalisation of care facilities. This included home care facilities with 24-h support and people living in small-group supported living arrangements. We also included studies in which part of the sample lived in a care facility.
The following exclusion criteria applied: children, adolescents, and adults with disabilities living with their parents or other family members; people with disabilities only attending day care facilities; and elderly with neurodegenerative diseases.

| Search strategy
We searched MEDLINE, Web of Science, Scopus, LENS.org, and IEEE Xplore for original studies published until May 2020, in English, Dutch, or German. Both published and unpublished studies were sought, and no limitations were imposed on the study designs eligible for inclusion. The last search took place on 15 January 2021.
The search protocol can be found on Open Science Framework: https://osf.io/byxh8/ and as a Data S1.

| Study selection
The study selection was conducted using a novel two-phase approach: a first, manual selection of studies and a second, broader selection of studies using active systematic review software (ASReview; ASReview Core Development Team, 2019). ASReview uses machine learning combined with active learning to predict which studies are most likely to be included, based on prior decisions by the author of whether presented studies were considered relevant or irrelevant. This made it feasible to screen a large set of additional literature, decreasing the risk of false negatives. The search strategy in the second selection was broadened by dropping the C (comparator) terms because these related to residential care and restrictive policies, which were not essential for addressing the research questions. This yielded a larger selection of additional candidate studies.
The first selection was conducted as follows ( Figure 1a): (a) databases were searched using the PICO terms, references were imported into Zotero referencing software and duplicates were removed; (b) all titles and abstracts were manually screened independently by two reviewers using the Rayyan application (Ouzzani et al., 2016); (c) full-texts were assessed for eligibility by one reviewer; (d) a preliminary selection was made through discussion by two reviewers. For the second selection of studies, the following steps were used (Figure 1a): (a) databases were searched using the PIO terms (C terms dropped); (b) titles and abstracts were screened independently by two reviewers using ASReview (screening time: 3 h); (c) full-texts were assessed for eligibility by one reviewer; F I G U R E 1 PRISMA flow diagram. The left side of the figure (a) describes the first, manual selection of studies. The right side (b) describes the second, broader selection of studies (using ASReview software) (d) a preliminary selection was made through discussion by two reviewers.
The final selection was made by these two reviewers, with consensus reached through discussion.

| Data extraction
Two reviewers independently extracted the data using Google Forms, with discrepancies solved through discussion. We extracted general information about the studies, that is, bibliographic information, design, method (quantitative, qualitative, mixed/multiple methods), and method(s) of assessment. To answer the research questions, we extracted information about the target population, sample characteristics, reasons for restricted in-person contact, means of ICT used for social contact, support from the care facility with using ICT, independent and dependent variables (quantitative studies), definition of outcomes (qualitative studies), informants, and information about the nature/goal of the social contact, impact of restricted contact on wellbeing, the level of 'off-line' engagement between interactive partners, and perceived benefits, barriers and preconditions of using ICT for social contact by people with disabilities and their social network members. We also recorded any additional outcomes that were measured, such as the use of ICT for other purposes than communication.
Criteria were additionally reviewed for necessity and completeness by another reviewer before extraction.

| Data synthesis
The findings are presented in a narrative synthesis, based on a deductive thematic analysis following the steps described by Braun and Clarke (2006). The themes were defined prior to analysis, based on the review questions. The initial coding and analysis was done by two reviewers, who discussed the content of the themes and subthemes in two meetings, reaching consensus after the second meeting.
Because only four out of nine studies presented quantitative findings, on different outcomes, insufficient data were available for a quantitative synthesis.

| RESULTS
During the first selection ( Figure 1a; PICO terms), two reviewers independently reviewed the titles and abstracts of all 891 potentially eligible studies. Using the Rayyan application, articles were rated as 'included', 'maybe', or 'excluded'. For the inter-rater reliability, articles rated as maybe were re-coded as included, yielding an agreement of 55% (κ = 0.547). One reviewer reviewed the full-texts of the agreements and mismatches (k = 58), after which a selection of four eligible studies was made through discussion with a second reviewer. During the additional, broader selection ( Figure 1b; PIO terms), two independent reviewers screened k = 328 (first reviewer) and k = 610 studies (second reviewer) using ASReview (rated as 'relevant' or 'irrelevant').
One reviewer reviewed the full texts of the agreements and mismatches (k = 87), and a selection of five additional studies was made after discussion with a second reviewer. In total, nine studies were eligible for inclusion.
The characteristics of the included studies are presented in Table 1.
Studies were published between 2006 and 2019. Methods were qualitative or quantitative descriptive. Three studies were described as singlecase experiments (Lancioni et al., 2015;Lancioni, Singh, O'Reilly, Sigafoos, Oliva, Campodonico, D'Amico, et al., 2013), but were considered quantitative descriptive because results for the different conditions were described without statistical tests. One study (Shpigelman & Gill, 2014) described their method as a mixed method approach, but as the qualitative and the quantitative parts addressed different questions instead of using multiple methods to answer the same questions and aggregating the results, the study was considered as partly quantitative descriptive and partly qualitative.
Two independent reviewers assessed the quality of the included studies using the mixed methods appraisal tool (MMAT; Hong et al., 2018; see Table 1). The screening criteria of the MMAT were met by all studies, except for the study by Lancioni et al. (2015). In this study, the research questions were not clearly formulated. We still decided to include it, because it is an empirical study and the findings were relevant to our review questions. Three studies (Patterson & Potter, 2009;Ramsten et al., 2020;Shpigelman, 2018) met all designspecific criteria. One study (Ramsten et al., 2019) met 80% of criteria, but showed discrepancies between the data analysis and interpretation of results. Another study (Shpigelman & Gill, 2014) met 80% of the criteria for qualitative studies and 20% of the criteria for quantitative descriptive studies. The remaining studies had several methodological problems, such as risk of bias in the data collection methods and a lack of information about the derivation and interpretation of results (Parsons et al., 2006), and risk of nonresponse bias, lack of information regarding the sampling strategy, representativeness of the samples, outcome variables, and inappropriate statistical tests (Lancioni et al., 2015;Lancioni, Singh, O'Reilly, Sigafoos, Oliva, Campodonico, D'Amico, et al., 2013). The complete quality assessment of the included studies is provided in the Appendix.
All studies addressed adults (>18 years) with intellectual or multiple disabilities. One study included adults and a child with multiple disabilities including intellectual disability . Three studies included people with severe multiple disabilities (Lancioni et al., 2015;Lancioni, Singh, O'Reilly, Sigafoos, Oliva, Campodonico, D'Amico, et al., 2013) and four studies included people with mild to moderate intellectual disabilities (Patterson & Potter, 2009;Ramsten et al., 2019Ramsten et al., , 2020Shpigelman, 2018). In one study (Shpigelman & Gill, 2014), the level of intellectual disability of the target population was not specified.
However, given that the study was targeted at adults who were 'capable of using the internet', we estimated the target population to be people with mild-to-moderate intellectual disabilities. One study  (Parsons et al., 2006)

Text messaging
Young adults with mild-to-moderate intellectual disabilities in the study by Ramsten et al. (2020) received text messages from family members and staff, which was perceived as a 'positive way of receiving information'. (p. 8). Some participants reported problems with understanding how to text message and responded by voice calling instead. Support staff perceived that the use of text messaging to contact family and friends decreased loneliness in this population (Ramsten et al., 2019).

Social media
In the study by Ramsten et al. (2020), the majority of young adults with intellectual disabilities owned a smartphone and a computer.
They followed friends and family on Facebook and Instagram and T A B L E 2 Identified themes and subthemes across the included studies connected with them by chatting and liking and responding to photos and timeline messages. Facebook was also used to initiate new relationships with unknown people, based on similar interests. More than half of adults with intellectual disabilities in the survey study by Shpigelman and Gill (2014) visited Facebook at least once per week, and mostly used it for contacting people they also met offline, such as friends and family. These uses of Facebook were also described by Shpigelman (2018), who conducted observations and interviews with adult Facebook users with intellectual disabilities.
Email and other software (pre-social media) Parsons et al. (2006) conducted their study before social media were widely adopted. In this context, the internet and email were among the most frequently used types of ICT used for contact with family and friends. In addition, word processing software was used to write letters to loved ones, and the digital camera was used to share photos.

| Theme 2: Effects on well-being
The initial focus on well-being in the context of restricted visiting policies had to be abandoned because it was not studied. Instead, we focused on the effects of using ICT for social contact on the well-being of users. F I G U R E 2 Included studies, target populations, and means of information and communication technology used for social contact. Studies are presented in ascending order of publication year Support staff perceived that independent use of ICT for social contact may decrease loneliness in young adults with mild-tomoderate intellectual disability (Ramsten et al., 2019). Adults with intellectual disabilities in the study by Shpigelman and Gill (2014) reported both positive and negative feelings and experiences with using Facebook. The majority (82%) reported feeling more comfortable talking to people on Facebook, instead of face-to-face, and

| Subtheme 2.1: Emotional well-being and quality of life
responses to the open-ended questions indicated positive effects on their emotional well-being. According to Shpigelman (2018), the use of Facebook contributed to the 'general psychological well-being' of adults with intellectual disabilities. The participants reported feelings of 'mood elevation, happiness, laughter, and pleasure' after using Facebook (p. e84). The use of Facebook was also described as having positive effects on participants' sense of being part of the community.

Negative effects
Adults with intellectual disabilities responded that Facebook had negative effects on their well-being (e.g., 74% of participants reported not feeling like everyone else when using Facebook, and 88% did not enjoy using Facebook; Shpigelman & Gill, 2014). Respondents in another study (Shpigelman, 2018)

| Subtheme 2.2: Quality of interpersonal relationships
The case study by Patterson and Potter (2009) described examples of closing sections of telephone calls between a young adult with an intellectual disability and autism living in a care facility and her family members. According to the authors, the way in which a telephone call is ended may protect and improve the quality of the relationship between a person with intellectual disability and a person without intellectual disability. An example of a telephone closing that protects the quality of interpersonal relationships is as follows: the speaker announces that they intend to close the call, but that this is not something that they desire. They might tell their loved one to leave the call because they need to do something else. This shows reluctance to close the call, which indicates responsiveness and care for the other person (Patterson & Potter, 2009

| Theme 3: Benefits, barriers, and preconditions
This theme with subthemes describes how people with intellectual disabilities, family, and professionals perceived the benefits, barriers, and preconditions of using ICT for social contact.

Consumer technology
The use of consumer technology for social contact has been reported by most of the included studies. Social communication technology has seen a large number of innovations over the past roughly 15 years. Before services such as social media were widely adopted, Parsons et al. (2006) pointed out that mainstream programmes or devices were the most frequently reported uses of ICT by people with disabilities attending day care facilities or living in residential care. The authors suggested that this may reflect organisational decisions to provide users with certain computer packages and that 'mainstream programmes may be sufficient for the provision of varied and interesting ICT use and may be achieved without necessarily having to find additional funds for learning disability specific software', but also recognised that specialised software 'may enhance ICT provision by catering to more specific needs'. (p. 40). However, no sample characteristics were reported in this study. Fourteen years after the study by Parsons et al. (2006), Ramsten et al. (2020) concluded that the use of text messaging and social media by young adults with mild-to-moderate intellectual disability seemed 'mainstream', but that it was more limited than in the general population, 'because cognitive impairment increases the need for support ' (p. 15). In this population, ICT was also used for other activities than social contact, such as playing games and listening to music (Ramsten et al., 2019).

Practical difficulties
All adults with intellectual disabilities in the study by Shpigelman (2018) used Facebook independently, but also reported barriers. These included practical issues, such as problems with creating an account and understanding Facebook jargon, and literacy difficulties, such as using functions that require typing and reading long posts of other users. Similarly, some young adults in the study by Ramsten et al. (2020) reported problems with understanding how to text message, but seemed to solve this by responding to text messages with voice calling. Difficulties with spelling did not seem to hinder their use of text messaging. Other barriers related to the use of a mobile phone included problems with the device or service provider.
Although participants were often able to solve 'simple technical problems' with their phone (e.g., replacing the battery), technical issues sometimes led to problems with initiating contact with friends.

Perceived risks
Support staff perceived the concern that ICT, especially social media, is a risky environment for young adults with mild-to-moderate intellectual disability. These individuals were believed to be more socially vulnerable than the general population due to difficulties with social skills and social responsibility. Potential risks perceived by support staff included harassment, victimisation by fraud, and circulation of personal information (Ramsten et al., 2019).

| Subtheme 3.3: Preconditions
Implementation Studies by Ramsten et al. (2019Ramsten et al. ( , 2020 (2013) suggested that for people with severe multiple disabilities, social contact through technology is possible and appears pleasurable if they are provided with a person-specific adapted device as well as intensive training with this device. With help from research assistants, people with severe multiple disabilities learned how to use the microswitch telephone technology and were able to successfully use it during the intervention sessions and even after the interventions.
However, data from the post-intervention sessions were not reported. Ramsten et al. (2019) noted that social care organisations for people with disabilities in Sweden did not have comprehensive plans for ICT support. Therefore, support for the use of ICT might depend on the interests of individual support staff members. However, support from family and staff with the use of ICT by young adults with mild-tomoderate intellectual disability seemed to be lacking in these studies. Ramsten et al. (2020) reported that some young adults did not know who to ask for help when experiencing difficulties with the use of social media. According to the authors, this may indicate that support was not offered by either family or support staff. In addition, participants sometimes received an evasive answer when asking for help. Support staff mentioned having 'only superficial insight into the young adults' ICT use' (Ramsten et al., 2019, p. 173), because they believed that the young adults preferred to ask friends and family for ICT support, instead of asking staff for help. In addition, the young adults' use of ICT was perceived as being dependent on the interests and attitudes of (older) parents. The decision to use ICT for social contact was therefore a choice made by parents, rather than a decision of the young adults themselves. if provided with intensive training and adaptive technology (Lancioni et al., 2015;Lancioni, Singh, O'Reilly, Sigafoos, Oliva, Campodonico, D'Amico, et al., 2013). The use of video conferencing by people with mild-to-moderate intellectual disabilities was briefly mentioned by one study (Ramsten et al., 2019), but no further details were reported. The modes of communication described in the included studies are also used by the general population (e.g., Van den Berg et al., 2012) and by long-distance families who do not have regular in-person contact (Abel et al., 2020;Wilding, 2006).
People with mild-to-moderate intellectual disabilities used consumer technology for social contact. This may have benefits, such as social inclusion, and may be easier for social care organisations to implement than specialised technology. Reported barriers were practical (e.g., difficulties using a device) or cognitive in nature (e.g., literacy problems, understanding jargon). Further, the included studies showed mixed evidence with regard to well-being and quality of life, indicating potentially positive (e.g., positive moods, decreasing loneliness) and negative effects (e.g., frustration caused by practical difficulties). However, findings from the included studies were descriptive and there was considerable risk of bias in most of the studies (see Section 2 and the Appendix). Opportunities and barriers described in the studies have been mentioned by previous reviews on the use of ICT by people with intellectual disabilities (Borgström et al., 2019;Caton & Chapman, 2016;Chadwick et al., 2013). In the general population, the use of ICT for social contact with close others during the COVID-19 pandemic has been associated with more positive affect, but also more negative affect and stress (Tibbetts et al., 2021). This may also be the case for people with intellectual disabilities. More empirical research is needed to make generalisable statements about the effects of using ICT for social contact on the well-being of people with disabilities living in care facilities or supported living.
People with intellectual disabilities may need support with the use of ICT for social contact. This is important during the COVID-19 pandemic, so that relationships with loved ones can be maintained (Tromans et al., 2020). However, most of the included studies indicated a lack of support by staff and family. People with intellectual disabilities have less access to ICT than the general population ('digital divide', Chadwick et al., 2013) and support is vital for learning how to use these technologies (e.g., Bryen et al., 2007;Carey et al., 2005;Li-Tsang et al., 2005). Also, support may depend on organisational guidelines and the interests and attitudes from individual support staff and family members (Ramsten et al., 2019). Parsons et al. (2008) described that the implementation of ICT within service organisations for people with intellectual disabilities may not only depend on practical factors such as time, staff training, budget, and appropriate equipment, but also on the attitudes of support staff about the purpose, usefulness, and importance of ICT for the service users. For example, staff from organisations with little to no use of ICT did not consider the use of ICT appropriate for older users or those with more complex needs, whereas organisations with regular and enthusiastic use of ICT believed that it was a valuable tool that could flexibly be used with anyone (Parsons et al., 2008). Such beliefs, also from family members, are important to address when implementing ICT for social contact within care facilities. Ramsten et al. (2019) emphasised the importance of support with the use of social media by people with intellectual disabilities because of exposure to risks such as victimisation and exploitation. Research of these risks is limited, but previous studies mentioned cyberbullying and financial and sexual exploitation (Buijs et al., 2016;Holmes & O'Loughlin, 2014). A recent study (Lee et al., 2020) showed that adolescents without intellectual disabilities who received fewer 'likes' on social media experienced stronger feelings of rejection and negative affect, especially those who were already victimised. People with intellectual disabilities may be vulnerable to these experiences. However, use of social media can be an important way of maintaining social relationships and forming new ones during times in which inperson social activities are limited. A recent study (Sun et al., 2020) showed that active use of social media (e.g., commenting on messages or photos) during the COVID-19 pandemic was positively associated with well-being. Future studies should look into ways in which staff from care facilities can offer support for safe use of social media, for example through training or psychoeducation (Caton & Chapman, 2016).

| Strengths and limitations
This study used a rigorous, two-step approach to systematically review the evidence on the use and feasibility of ICT for social contact by people with disabilities living in care facilities. A potential limitation is that the researchers' perspective has guided the selection of literature and extraction of data and no stakeholders were involved in the interpretation of the findings. To improve objectivity in the interpretation of the findings, the data were extracted by two independent reviewers. In addition, the nine studies included in the review were Three studies included in this review suggested that even people with severe intellectual and multiple disabilities could enjoy using the telephone to call their loved ones and to watch videos of family members talking, even though their responses during these interactions were minimal. In addition, these technologies were perceived posi- Lastly, all of the included studies focused on visual and auditory means of ICT-mediated communication. However, the use of these technologies may not be feasible for some populations, such as people with visual and hearing impairments. In addition, regular communication technologies such as the telephone and video conferencing may not provide a 'feeling of presence', due to the absence of touch and the inability to make real eye contact (Cohen et al., 2017). Recent technological advances by Cohen et al. showed that 3D virtual reality and haptic technology can be used to recreate a feeling of presence, which might be expected to increase the positive emotional impact of such contact (Petrova & Schulz, 2021). Making such technologies available for the wider population would be a valuable direction for innovation and research.

| CONCLUSION
This systematic review showed that the use of virtual social contact may be feasible and enjoyable for people with various types and levels of disabilities living in care facilities. However, there is very little research addressing the question whether this type of social contact meets the needs of people with disabilities and their families and friends, and how it might affect well-being when in-person contact is (temporarily) not possible, such as during quarantine situations or due to other circumstances (e.g., when elderly parents are unable to visit their children with disabilities). Given the impact that pandemic restrictions appear to have (e.g., Schuengel et al., 2020), more empirical evidence is needed about the feasibility and effectiveness of ICT for social contact of people with disabilities living in care facilities, before this might be recommended as an alternative for in-person contact with loved ones, both during and after the COVID-19 pandemic.