Facilitated by near-ubiquitous digital connectivity in the global north, digital technologies are increasingly used across all areas of life and bestow significant benefits upon those who use them [1,2]. In response, many organizations, not least public service providers, proceed as if access to the internet and digital technology is already universal [3]. However, although connectivity has increased, access to digital technologies and the competencies to use them are not yet universal [4,5]. In the United Kingdom, the 2021 Consumer Digital Index found that approximately 10 million people (14% of the population) do not have the basic digital competencies needed for use the use of everyday digital technologies [6]. Digital exclusion can no longer be seen as a product of rural or remote living, place-based infrastructural and connectivity issues, or older age. Instead, digital exclusion is currently most likely a product of wider social and economic inequalities. Sociodigital inequalities [7] refer to the interplay between traditional (ie, social, economic, and health) and digital inequalities and have led to systematic differences in the ability and opportunity of different groups to beneficially use digital technology and participate in society. These differences, in turn, contribute to deeper social inequalities, which lead to greater digital inequalities over time [8].

A growing body of literature has begun to document the extent of digital inequalities and highlight the specific barriers to digital use in different social contexts [8]. In the United Kingdom, the Digital Inclusion Strategy sets out how the government and partners from the public, private, and voluntary sectors will increase digital inclusion [9] by targeting the following 4 interconnected barriers to digital inclusion: digital access, skills, confidence, and motivation [10]. However, although national policies are imperative to improve inclusion from an infrastructure and digital connectivity [1] perspective, localized interventions are much better placed to address digital skills, confidence, and motivation across previously excluded groups [11]. In the United Kingdom, localized digital training interventions are typically delivered by intermediaries, such as voluntary and community sector organizations and public libraries, and include formalized group programs, one-to-one digital buddy programs [12], and peer support [13]. Specifically, the United Kingdom’s Local Government Association’s Digital Inclusion Programme is funding councils to reach residents and provide personalized digital training programs for those who do not have access to or confidence in using digital communication platforms [14]. However, there is limited evidence on whether these digital training interventions are effective in increasing “digital readiness,” the competency and increased motivation to use digital technologies [15-17]. However, evidence does suggest that individuals who participate in these digital interventions have a higher socioeconomic status, higher education, higher social participation, and a greater experience with technology [18,19]. Conversely, those who are already experiencing sociodigital inequalities remain the hardest to reach [20-22] owing to reported barriers, including physical health issues, a negative attitude toward technology, caring for a sick spouse, a lack of energy, and a lack of time [23,24].

Within the context of the digital needs of hard-to-reach populations, this study focused on social housing residents in Southwest England. In the United Kingdom, social housing associations (HAs) are not-for-profit organizations that provide rental properties at 50% to 60% of market rates to those whose circumstances exclude them from the private market [25]. Overall, 3.9 million people live in social housing for various socioeconomic reasons [25]. Studies have shown that the demographic and socioeconomic profile of social housing residents means that they are significantly more likely to be digitally excluded and harder to reach than other groups in the United Kingdom [10,26-28]. This is because the factors that are known to increase digital exclusion are found at higher incidences among social housing residents than among those outside the social housing system [29,30]. These factors include lower incomes, fewer qualifications, older age, physical and mental health issues, disabilities, and living in more deprived areas [31,32].

The Smartline project [33] worked with 200 social housing households in one of the most deprived areas of England [34] to understand the potential of everyday digital technology to address health and well-being challenges. A qualitative scoping study on the feasibility and acceptability of digital technology among Smartline participants found that although the participants had positive perceptions of technology and were keen to try new technologies, digital readiness and the desired digital destination (goals) varied greatly among the community [28]. Several concerns surrounding technology use were identified, including data security and privacy concerns and the fear of “making a mistake” or “pressing the wrong button.” Many participants expressed a strong desire for further training and support.

Following this research, a training intervention was conducted to help Smartline participants get on the web and use digital communication technologies, such as web-based video calls and messaging, with confidence. The Getting Online: Staying Connected (GO:SC) intervention was originally planned as a face-to-face intervention with peer-to-peer support on how to use video calling technology. However, as with many research interventions during this time [35], the outbreak of the COVID-19 pandemic in March 2020 meant that the study had to be redesigned as a smaller-scale telephone-based training intervention for social housing residents delivered in conjunction with the Cornwall Council’s Digital Inclusion Team (DIT). The aim of this study was to evaluate the feasibility of the telephone-based training intervention. Specifically, informed by the feasibility framework of Bowen et al [36] and RE-AIM (Reach, Effectiveness, Adoption, Implementation, and Maintenance) model of Glasgow et al [37], this study aimed to evaluate the (1) reach and demand, (2) implementation, (3) acceptability, and (4) potential efficacy of the telephone-based digital training intervention for social housing residents [36,37]. These are established frameworks for evaluating digital health interventions [38,39] and are suitable and credible for testing an unexamined intervention in a real-life setting where constraints exist over conditions [36,38,39].

This was a mixed methods feasibility study. The protocol and ethics application were informed by best practices for the process evaluation of public health interventions [40]. An overview of the study procedure, including recruitment and data collection, is provided in Figure 1 and Table 1 and detailed throughout the subsequent subsections.

Recruitment took place between October 2020 and June 2021. As we did not expect potential participants to have email addresses, recruitment was initially planned to take place via telephone, targeting only the Smartline participants. Participants who identified as having low digital competency or confidence in the Smartline baseline survey [41] were contacted via telephone and completed the recruitment survey asking respondents if they were interested in undertaking the digital training (Figure 1). The Smartline survey data enabled the analysis of the characteristics of those who did not respond and were not interested in the training offer.

As recruitment was lower than expected, it was decided to extend the intervention to all Coastline Housing residents in the greater Camborne area. Extending to all the residents of the HA meant that individual phone calls were no longer possible. The recruitment survey was sent to all Coastline residents who had an email address. A specially designed advertisement for digital training was also disseminated to the residents using the communication channels of the HA, including the residents’ monthly newsletter (printed and electronic), the HA’s Facebook (Meta Platforms, Inc) page, and printed posters placed in community areas. The recruitment survey asked the participants to outline their reasons as to why they were interested or not interested in receiving the training. The survey thus enabled the analysis of the characteristics of both the participants who were interested and participants who were not interested in the training intervention.

Residents who expressed an interest in the digital training either via telephone or email were posted the study information sheet, the consent form, and a baseline survey assessing well-being and digital competency (Multimedia Appendices 2-4). Participants returned the completed forms via a Freepost envelope to the research team, who forwarded the participants’ phone numbers and digital training interests to the Cornwall Council’s DIT via email. The DIT then directly telephoned each participant to arrange the provision of the training.

The inclusion criteria were adult (aged >18 y) social housing residents in Cornwall who defined themselves as lacking either the competency or confidence to use digital communication apps. These included existing participants in the Smartline project and residents of the wider Coastline HA. Participants were excluded if they did not have access to an internet connection or at least 1 internet-enabled device. Training was provided free of charge to participants (in line with the DIT’s protocol).

The intervention is described in line with the TIDieR (Template for Intervention Description and Replication) standards for intervention reporting [42]. The training intervention was based on a 4-week, face-to-face, and basic digital skills course that was established by the DIT and delivered in libraries and community venues across Cornwall before the COVID-19 pandemic. The course structure adhered to the UK government’s Essential Digital Skills Framework [43], encompassing fundamental computer skills, problem-solving skills, communication skills, transactional skills, and skills for handling digital information and content, as detailed in Multimedia Appendix 5.

In response to the social distancing measures in place during the COVID-19 pandemic, the DIT adapted the training to be delivered over telephone. This adaptation was informed by best practices in digital inclusion initiatives [44] and participant preferences identified in our prior qualitative scoping work [28]. The adapted training approach emphasized informality; flexibility; person-centeredness; and task-specific, on-demand delivery, diverging from traditional, predetermined, and technically focused programs imposed by training organizations [44]. This informal model aligns with research indicating that training success factors encompass participant autonomy, personalized learning, practice opportunities, and individualized support [13,45]. Informal learning, in particular, enhances self-efficacy and proves highly effective for digitally marginalized groups [46].

The overall purpose of the training intervention was to support participants in achieving active and continued use of digital devices and communication apps of their choice (eg, WhatsApp [Meta Platforms, Inc], Zoom [Zoom Video Communications, Inc], and Facebook). The training was delivered through one-on-one phone calls facilitated by a digital inclusion officer from the DIT, all of whom possessed a background in education. To provide course content guidance and ensure consistent and best-practice training delivery, the DIT used a bank of task-specific instructions (ie, step-by-step guide to installing WhatsApp), which could be printed and posted to participants on demand (Multimedia Appendix 6). Training sessions varied in duration (between 5 min and 2.5 h) according to each participant’s individual learning needs, often including follow-up calls as necessary.

Overarching ethics approval to conduct research with Coastline Housing (a social HA in Cornwall) residents was granted by the University of Exeter Business School Research Ethics Committee as part of the Smartline project (eUEBS002996 v4.0). Specific approval for the GO:SC project was granted by the College of Life and Environmental Science Penryn Research Ethics Committee (eCORN002229). Written informed consent was obtained from all participants involved in the study (Multimedia Appendices 3 and 4). Data were anonymized using pseudonyms. Participants who completed the follow-up survey received a £10 (US $12.7) shopping voucher as compensation for their time.

The following 4 qualitative and quantitative data sources were used to assess the feasibility and acceptability of the intervention: a recruitment log, recruitment survey, well-being and digital competency survey (administered at baseline and 6-month follow-up), and digital training call log (Table 1). Previously piloted by the Smartline participants, the well-being and digital competency survey (Multimedia Appendix 2) used quantitative rating scales and was completed at baseline (immediately before training commencement) and 6 months following the training, with the aim of assessing potential outcomes. This survey (disseminated via post) was based on a validated survey, the “Happiness Pulse” [47], and included 4 domains of psychological well-being (general, emotional, behavioral, and social). A bespoke module on digital attitudes, behaviors, and competence was developed by the research team for the wider Smartline project. This digital module was informed by behavior change and technology acceptance theories [48-51] and included questions adapted from existing sources, including the UK government’s Digital Inclusion Evaluation Toolkit [52]. The module contained specific questions on video calls and messaging in addition to questions on technology in general. The theoretical basis of this module is provided in Multimedia Appendix 7 [47,49-63].

To determine the reach of and demand for the intervention, as well as participants’ levels of engagement with the training, descriptive statistics (frequencies and percentages) from the recruitment log, recruitment survey (n=168), and training call log (n=12) were calculated. A probit model was used with the recruitment survey data to identify the socioeconomic factors associated with initial interest in participating in the digital training program. The analysis was performed in Stata (version 17; StataCorp) [64].

For the well-being and digital competency survey, the scoring protocol for the Happiness Pulse was followed [53], with means and SDs calculated from summary scores to describe each of the 4 well-being domains. As digital competency outcomes were measured on interval scales, medians and IQRs were calculated for these outcomes.

The COREQ (Consolidated Criteria for Reporting Qualitative Research) criteria [65] for reporting qualitative research were adhered to throughout the analyses. Qualitative data analysis used data from the recruitment log, recruitment survey, and digital training call log. To manage the qualitative data analysis process with transparency and traceability [66], NVivo (QSR International) [67] was used. In line with best practices [68], 3 rounds of inductive coding were conducted using a constant comparative method [69,70]. The first round of coding was open and focused on identifying and labeling discrete incidents. For example, “I have a smartphone, but I struggle to use it” contains 2 incidents: an object (smartphone) and a construct (competency). The second round of coding was axial, where open codes were compared (via contradiction, expansion, and support) and integrated into themes, and the third round of coding was selective, where connections between themes were compared and refined to build the grounded theory. The lead researcher (TW) conducted the initial coding, and the themes identified were discussed with a second researcher (SAB). To further improve rigor and reliability [71,72], a third researcher (KM) was consulted, and minor discrepancies were resolved through discussion. Pseudonyms were applied to protect participants’ identities.

The quantitative and qualitative data were integrated within the selected feasibility criteria of reach, demand, acceptability, implementation, and potential outcomes [36] to provide a complete picture of the feasibility, acceptability, and potential impact of the intervention. This use of mixed methods enabled triangulation to strengthen the validity of the findings and complementarity to explore different facets of a phenomenon [73].

The main implementation finding relates to the reduction in the number of participants between those who expressed initial interest in the training and those who completed the training. The study found that key facilitators included the HA’s established relationship with a digitally susceptible population and the flexible, informal approach of the DIT. Barriers involved issues with written consent, internet access, and device functionality. Here, we discuss each of these in turn.

The tangible and intangible positive roles that the HA played as an intermediary [76,77] in the implementation of the intervention cannot be overstated. In the United Kingdom, the role of HAs has evolved over recent decades to include supporting the health and well-being of their residents [78]. As a result, many HAs have built meaningful, trust-based relationships with their residents, and this factor played an important role in this project. From a practical perspective, this meant that the collaboration of Smartline with the HA enabled a wider reach in advertising the training offer. For example, the HA was able to use its customer relationship management system with phone numbers and email addresses to contact over 4000 residents. Indeed, examining the socioeconomic factors associated with initial interest in participating in the digital training program revealed that respondents who were contacted by the HA were more likely to be interested in the training program (P<.001; Table 3).

Another factor important for reach was customer liaisons by the HA with known susceptible residents who they felt would be interested in and would benefit from the training. Although recruitment remained difficult and the uptake of the intervention was low, the implementation of the recruitment survey and associated data collection would not have been possible without the help of the HA.

Participant dropout at the recruitment stage negatively affected implementation; only a small proportion of those who expressed an initial interest went on to complete the training. Among those (n=67) who were initially interested in the training, many (n=37, 55%) required multiple callbacks (up to 4) before they could be reached again; for reasons discussed next, many (55/67, 82%) dropped out. Making multiple callbacks was an administratively complex task that required many hours and email exchanges within the recruitment team. The need for written consent was a key factor for recruitment dropout, as was the need for a working digital device and an internet connection. Completing and returning the necessary consent forms was found to be a particular challenge for some participants. Participants noted that “filling in forms is a worry” (Julia, a woman resident aged 65 to 74 years) and that making time for the task was difficult: “I have the forms but have a lot on at the moment” (Jess, a woman resident aged >75 years). The possession of an internet-enabled device and internet connection was a requirement for participation in the training. Although we found high rates of digital technology ownership and internet connection possession among the participants contacted, 17.3% (29/168) of the respondents to the recruitment survey did not have a tablet, 5.9% (10/168) of the residents did not have a smartphone, and 7.1% (12/168) of the residents did not have an internet connection. Financial constraints, in particular, were a key reason for the lack of internet:

For other interested participants, the working condition of the device was also a barrier to participation in the training:

In addition, it is important to reflect on the implementation of the training itself. In practice, 3 different digital inclusion officers delivered the telephone-based training to 12 participants. The calls lasted between 5 minutes and 2.5 hours. To fit around participants’ caring responsibilities and day-to-day lives, all participants required multiple calls to arrange and rearrange training times. Training sessions lasted as long as required for the participant to learn to use the digital application, and follow-up support calls (up to 5) with the digital inclusion officer were arranged to ensure that the training objectives were achieved (Table 4). Successful delivery of the training required flexibility and persistence from the DIT. This indicates that there is a wide range of digital needs that are best served by informal and one-to-one support directed by individual needs [13,46].

Similar to the key findings on implementation, the study’s key findings on acceptability revolve around the substantial dropout rate, highlighting the challenges of translating training interest to training participation. Participants dropped out because of competing priorities, including health issues, caregiving responsibilities, and time constraints. However, all the 12 participants who started the training completed it; this suggests high acceptability of the intervention itself.

As with all research conducted during this period, the COVID-19 pandemic impacted all aspects of the intervention. We found the lockdown to have both a positive and negative impact on acceptability. For some, it was a driver for learning how to use video calls to access health services, social groups and classes, and resident groups, which had transitioned to function on the web:

The lockdown also had a negative impact on the acceptability of the training intervention, with potential participants noting a need to attend to everyday tasks and self-care, rather than learning new skills: “I have been unwell with COVID for months, I just need to focus on the day-to-day things at the moment” (Judith, a woman resident aged 45 to 54 years). Importantly, we found that the pandemic compounded and increased several preexisting barriers to undertaking digital training for our participants. For example, we found that preexisting health conditions arose as a key barrier to participation:

Participant’s lives were generally “busy” (Diana, a woman resident aged 65 to 74 years), and they did “not have time for this at the moment” (Ruth, a woman resident aged 55 to 64 years) or needed to focus on caring responsibilities. Regarding the acceptability of the training vis-à-vis the everyday lives of our participants, 2 further themes emerged: the method used to deliver the intervention and the timetabling of the training sessions. Understanding that the Smartline participants had a strong preference for peer-based, face-to-face activities [27], participants’ concerns about the effectiveness of support being delivered over the phone were expected:

The timing of the training was an important factor for participants, particularly for those who were working or had chronic health issues: “I have ME [myalgic encephalomyelitis or chronic fatigue], so afternoon is better for me” (Susan, a woman resident aged 55 to 64 years). Therefore, despite the increased need for digital communication at this time [79], the already complex social and health needs of the Smartline participants [80] meant that the recruitment and retention of interested participants was time intensive and a major challenge for the acceptability of the intervention.

Of the 12 people who participated in the training, 9 (75%) completed the baseline and follow-up well-being and digital competency surveys (Multimedia Appendix 2). Table 5 provides the summary well-being scores for participants who undertook the digital training intervention; there were no changes in the mean general, emotional, or social well-being between baseline and follow-up.

Counter to the overall aim of the intervention, we found a small reduction in behavioral well-being for participants who had undertaken the intervention [53]. The behavioral well-being measures included a question asking whether participants were attending courses and a further question on whether respondents were learning a new skill. As such, respondents would have indicated “yes” to these questions at baseline; however, at follow-up, participants were unlikely to be undergoing any other training given the ongoing COVID-19 restrictions. As such, this decrease could be explained by a confounding reduction in training levels from before to after the intervention, rather than an actual decrease in behavioral well-being.

Table 6 provides the summary digital competency scores of the intervention participants (n=9) at baseline and follow-up. From Table 6, it can be inferred that there were no clear changes over time in any of these scores.

Although conclusions on potential efficacy based on the survey data are limited owing to the small sample size, the qualitative data indicated that participants had achieved their original training objectives. Table 4 provides a summary of the training objectives, training needs, and the level to which they were met.

From Table 4, it can be inferred that increased competency in using video calling apps, particularly help with the installation and setting up of these apps, were key training needs. Participants’ training needs were motivated by both social and personal goals, such as contacting family members and becoming more competent with digital technology in general. Although advertised as training on video calling and messaging, participants received a diverse range of training, from training on how to order prescriptions on the web to training on more complex tasks such as recording and mixing music using web-based platforms. The flexibility to deliver such diverse and tailored training was not initially planned as part of the intervention yet proved a successful strategy. Overall, from the qualitative data, we found that participants achieved their training objectives and social and personal goals and have the potential to use other digital technologies in the future.

The findings indicate a demand for basic digital training among social housing residents, and the intervention was acceptable for those who received it. However, recruitment and implementation were challenging, with potential participants experiencing barriers that reflected the sociodigital inequalities commonly found in this group [10,26-28]. Barriers were strongly influenced by socioeconomic and health circumstances, which were closely related to preexisting digital readiness (eg, preexisting skills, confidence, motivation, and access) and further compounded by the COVID-19 pandemic. Our results confirm that the factors known to increase digital exclusion are particularly high among social housing populations [29,30] and highlight the interplay between traditional inequalities (ie, social, economic, and health) and digital inequalities [7]. However, social and personal goals were achieved by the participants who received the intervention. This demonstrates that tailored, flexible, and persistent training efforts can overcome barriers.

Regarding policy, the UK’s Digital Inclusion Strategy aims to “equip the whole country with the skills, motivation and trust to go on the internet, be digitally capable and to make the most of the internet” [81].

Although national policies are imperative to improve infrastructure, access, and digital connectivity [1,9,10], the implications of this study are that an effective policy also needs to focus on strategies for reaching digitally excluded groups [20-22]. The essential strategies and recommendations for practice derived from our findings are listed in Textbox 1.

By implementing these strategies, policy makers, organizations, and communities can address sociodigital inequalities. However, in making these recommendations, we recognize that this places a considerable burden on individuals delivering such interventions. Future feasibility research of this nature could investigate the burden on intervention deliverers and the associated economic costs, which were not examined here.

A strength of this study is its focus on social housing residents, an understudied population with associated socioeconomic inequalities that can present particular barriers to digital technology use. Another strength of this study is the collection of quantitative and qualitative data from various sources to enable rich insights into feasibility, acceptability, and potential impact, including the capture of data on those who initially expressed an interest in participating but did not go on to receive the intervention.

A limitation of the study is the small sample size for quantitative evaluation; however, as this is a feasibility study, quantitative outcomes (well-being and digital competency) were only intended to be indicative of the potential impact and were supplemented by qualitative findings on the achievement of training objectives. Owing to the unexpected difficulties in recruiting participants to the intervention, follow-up interviews were not possible within the time frame of the project.

We measured psychological constructs with individual items to balance theory alignment with reducing participant burden. However, we suggest that future feasibility and acceptability studies use established multiitem measures to assess such constructs. The final limitations to note are those with regard to intervention delivery fidelity and economic costs. The study did not assess how the DIT delivered the intervention, other than following the “standard operating procedure.” It is possible that variations occurred in training delivery with regard to relationships with participants. Future studies should consider structured approaches to assessing intervention fidelity [82].

Finally, the intervention’s cost could not be specified, as it was provided through a county council’s DIT. Despite their personalized nature, similar personalized digital training programs are common in UK councils [14]. Therefore, the results of this study are valuable for providers facing challenges in engaging specific resident groups.

This study contributes to the contemporary literature, theory of “sociodigital inequalities” [7], and need to redefine digital inequalities in terms of their relation to other forms of socioeconomic and health inequalities. To address sociodigital inequalities, this study highlights that future policies need to be more proactive in reaching excluded groups, and such initiatives need to be considerate of people’s everyday lives, which will be conditioned by social and health circumstances. To achieve this, initiatives need to be appropriately resourced and include a flexible combination of digital provision with an agile person-centered approach to training based on personal needs and goals.