Older adults with an intellectual disability

People with an intellectual disability are living longer, although they experience accelerated and earlier ageing, often beginning from approximately 45 years of age, and earlier still (at 40 years) among people with Down syndrome [16–18]. Older adults with an intellectual disability frequently experience social exclusion, reduced access to mainstream social opportunities, and fewer friendships outside disability services [19,20]. They also report disproportionately high levels of loneliness, depression, and low quality of life [3,4,9,21,22], despite consistently expressing a desire to build friendships and participate in their communities [19,20].

A systematic review reported loneliness among people with intellectual disability at levels of nearly five times the prevalence seen in the general population [22]. Barriers included communication challenges, limited opportunities to develop social networks, and societal attitudes that reinforce exclusion [23].

Although befriending schemes exist for people with intellectual disability, evidence of their effectiveness remains limited. A small feasibility study suggested potential benefits but lacked statistical power [24], while a feasibility randomised trial in adults with intellectual disability in London UK, struggled to recruit sufficient participants and did not progress to a full trial [25]. Furthermore, befriending schemes may not promote sustainable social participation: once a volunteer withdraws, social connections may not continue. Thus, there is no robust evidence that befriending schemes improves clinical or cost outcomes for people with an intellectual disability, and existing befriending schemes may not build enduring community integration. Many people with intellectual disability attend community-based groups that are offered only to other people with an intellectual disability. However, there is now growing research that is examining how people with intellectual disability engage in non-disability community-based groups.

Policy guidance emphasises inclusion of people with intellectual disability in their local community groups [26–29]. UK guidelines from the National Institute for Health and Care Excellence (NICE) (NG96) recommended that older people with an intellectual disability should have equal access to local community-based groups [18]. However, there is limited evidence on how best to operationalise these recommendations.

The MATILDA intervention

MATILDA (Managing Activities Together to Involve older adults with a Learning Disability in their Local Areas) is a novel community-based, mentor-supported intervention that promotes social inclusion, quality of life and wellbeing. MATILDA is based on the Australian “Transition to Retirement” intervention, which enabled older adults with an intellectual disability to join community-based groups for six-months with volunteer support. MATILDA adapts and extends this approach within UK health and social care contexts [30,31].

This study was guided by the RE-AIM framework (Reach, Effectiveness, Adoption, Implementation, Maintenance) to structure the evaluation of the feasibility, acceptability, and potential sustainability of the MATILDA intervention. RE-AIM provides a pragmatic implementation lens suited to community-based public health interventions and helps identify whether MATILDA could be scaled in routine practice. This study also tested the practicality and acceptability of conducting a definitive full-scale trial to evaluate MATILDA’s effect on social connectedness, well-being, and quality of life.

Aim

The primary aim of this study was to assess the feasibility, acceptability, and procedures required in order to conduct a randomised controlled trial, and to determine whether the prespecified progression criteria for advancing to a full trial were met. Secondary outcomes explored the clinical outcomes, process evaluation and health economics. The process evaluation was guided by the RE-AIM Framework (Acceptability, Reach, Effectiveness, Adoption, Implementation and Maintenance) [32].

Objectives

1. Determine feasibility outcomes and examine whether progression criterion have been achieved (eligibility, recruitment, consent, randomisation, matching to community groups and mentors, attendance, retention, dropout)
2. Explore stakeholder views (older adults with intellectual disability, mentors, community groups) on the acceptability of the MATILDA intervention
3. Examine whether mentoring relationships and community participation can be sustained over the study period?
4. Assess appropriateness and acceptability of the clinical outcome measures
5. Explore potential cost-effectiveness.

Study design and setting

This was a two-arm, randomised feasibility study with 1:1 allocation, which was conducted in two regions of the United Kingdom: one Health and Social Care Trust in Northern Ireland and one Trust in London, England. We approached 95 potential participants and 57 older adults with an intellectual disability agreed to join the study. We randomly allocated them to either the MATILDA intervention plus usual care or usual care with three group recreational activities with other older adults with an intellectual disability (active control arm). The intervention lasted six-months. The primary outcome measures were feasibility outcomes (eligibility, recruitment, consent, randomisation, matching to community groups and mentors, attendance, retention, dropout). Secondary outcome measures focused on the clinical outcomes (i.e., quality of life, anxiety, loneliness). Both primary and secondary outcome measures were collated at baseline, 6- and 12-months post intervention. A process evaluation was undertaken using the RE-AIM framework [32] and a health economic evaluation was also explored. This study was aligned with the CONSORT 2010 extension for pilot and feasibility trials [33] (SEE Fig 1).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. MALIDA Consort Diagram.

https://doi.org/10.1371/journal.pone.0341940.g001

Participants and recruitment

The recruitment process ran from July 2023 to March 2024. There were some delays in the London site, so identification and recruitment of participants commenced there in October 2023. Statutory and voluntary disability community groups helped to identify potential adults with an intellectual disability who met the inclusion criteria using a user-friendly information sheet. A member of the research team at each site then met with the potential participant, and if needed a family/ paid carer for additional support, to ensure that each participant had the capacity to consent and understood what was involved in the trial, including completing of questionnaires at three time points, randomisation, and if selected participating in a community group for six-months.

Inclusion criteria

The inclusion criteria were that each participant had to be at least 45 years of age, living in the community, have a mild/ moderate intellectual disability, sufficient communication to engage in a community group, and sufficient mobility to attend the community group. All participants had to have the capacity to give their own written consent.

Sample size calculation

As this was a feasibility trial and the purpose was to provide estimates of key parameters for a larger definitive trial rather than to power this study to detect statistically significant differences, a formal power calculation was not conducted [34]. A target sample of 64 participants was selected pragmatically, based on the anticipated capacity of the Volunteer Co-ordinator to recruit, manage and support participants across the community groups involved in the study. This sample was considered adequate to assess the primary feasibility outcomes of recruitment, retention, intervention delivery and outcome data collection, and to inform the planning of a future definitive trial.

Randomisation

Following informed consent and completion of baseline assessments, participants were randomised on a 1:1 ratio to either the MATILDA intervention arm plus treatment as usual or an active control arm plus treatment as usual. Randomisation was stratified by study site (Northern Ireland or London). The allocation sequence was generated by a member of staff at Ulster University who was independent of the study team and not involved in participant recruitment, enrolment, assessment, or intervention delivery. Concealment was achieved using sealed opaque envelopes. No permuted blocks were used. At randomisation, each participant was assigned a unique Participant Study Number, which was used throughout the study for identification purposes. Following randomisation, research staff informed the participants of their group allocation.

Due to the nature of the intervention, participants and mentors could not be blinded to treatment allocation. In addition, the research staff who conducted baseline and follow-up assessments were aware of participants’ group allocation. To minimise the potential for assessment bias, standardised assessment procedures and validated outcome measures were used consistently across both sites using participants unique Participant Study Number.

MATILDA intervention

Older adults with intellectual disability were matched to one or two mentors from a local community group. This matching process was managed by a Volunteer Co-Ordinator (VC) in Northern Ireland and London. The VC at each site had to assess each person’s capabilities, motivations, interests/ hobbies, financial situation, health co-morbidities and living context (i.e., living on their own, with family or in supported living/ residential care) transport arrangements and match the person to a local community-based group within the same locality, with similar interests, and pair them to one or two mentors. The VC supported each adult with an intellectual disability to meet their mentor(s) prior to commencing the group. Then the VC physical accompanied each adult with an intellectual disability to attend their local community group and meet with their mentor(s), then gradually withdrawing their support over the following weeks. The total length of the MATILDA intervention was six-months.

One of the core responsibilities of the VC was to develop a rapport and trusting connection between the adult with the intellectual disability, their family or/and paid carers, and the community group and mentor(s) to allow the participant to move outside of their safe and protected world of disability services. Participants randomised to MATILDA also received treatment as usual (TAU), ensuring that they did not lose any treatments or care that was standard. All mentors completed a short 2–3 hour training session on the role of a mentor to ensure they were aware of the study purpose and how they as mentors would be able to provide support.

Active control arm

Those in the active control group received TAU and were offered three group recreational activities with other older adults with an intellectual disability within this arm: this included going for a meal, going to the museum, zoo, or cinema. TAU was established at the start of the study and again at the end of the study. Participants in the control group completed the same data gathering instruments as the intervention group at baseline, 6- and 12-months.

Feasibility outcomes/ Process evaluation

A process evaluation was conducted in line with the Medical Research Council’s Guidelines [35] to assess whether MATILDA was delivered and received as intended, and to understand how trial processes interacted with the implementation context. This included an examination of both feasibility outcomes and clinical outcomes, including the acceptability of outcome measures and the perceived benefits and challenges of implementing MATILDA.

We used the RE-AIM framework (Reach, Effectiveness, Adoption, Implementation, and Maintenance) [32] as part of the process evaluation. Reach refers to the number of older adults with an intellectual disability who engaged with their local community groups and mentor(s) at 6-months with support from the VC. We further explored whether those adults still engaged with their local community groups at 12-months. Effectiveness assessed the impact of MATILDA on the participants’ quality of life, anxiety, loneliness and community connectedness from baseline to the two follow-up time points using the standardised measures, focus groups and VC reports. The effectiveness data were analysed using descriptive statistics. Adoption explored how the mentors, other community group members and the staff in the supported living facilities practically and emotionally supported the adults with intellectual disability to engage in their groups. Implementation concerned whether MATILDA was delivered as intended. We collated the attendance records of the adults with intellectual disability at the groups, additional activities undertaken from the mentors and the VC’s attendance logs and fidelity checks, respectively. Maintenance assessed the extent to which the MATILDA’s effects were sustained over time at both the individual level and the organisational level.

By considering these five domains, RE-AIM provided a structured approach to undertaking the process evaluation evaluating both the internal validity and external applicability and supporting generalisability and implementation of the MATILDA intervention into routine practice.

Clinical outcome measures for the adults with intellectual disability

The secondary outcome measures focused on the clinical outcomes and were collated at baseline, 6- and 12-months post intervention for both arms. We used the WHO QOL-Dis as a measure of health-related quality of life developed by the WHO [36]. This 26-item short version consists of two benchmark items on general health (not used in the scoring) (one general QOL item, one general health item) and 24 specific items which generate a total score and four domains: physical health (7 items), psychological health (6 items), social relationships (3 items), and environment QOL (8 items). Each item is answered on a 5-point Likert scale, which assesses the intensity, capacity, frequency, and evaluation of QOL facets with respect to the last two-weeks. The WHOQOL-Dis has been validated in other intellectual disability population studies and found to have good psychometric properties [36]. Higher scores mean better QOL.

We used the Glasgow Anxiety Scale for (GAS-ID) which was developed for adults with intellectual disability [37]. This is a 27-item self-report scale. It has been used in several studies with adults with disability and found to have good psychometric properties [37]. Scores ranged from 0 to 54: score of 13 or over indicates anxiety.

We also used the Modified Worker Loneliness Scale, a 12-item questionnaire that measured Aloneness and Social Dissatisfaction in adults with an intellectual disability [38]. It has been used in several other disability studies and found to have good psychometric properties [38]. We also used the WHO Wellbeing 5, a short 5-item scale to measure subjective wellbeing [39].

We explored a health economic evaluation to examine issues likely to be encountered in the conduct of a full economic appraisal including sources of data, how best to collect these data and the inclusion of spill-over effects into the analysis. We used the Client Receipt Services Inventory [40] to collate service use and costs for people with an intellectual disability over the 12-months of the study. It was completed by the adults with an intellectual disability and their family/ paid carer. We also used the EQ-5D-Y [41] a measure of health-related QOL which provides a description of health using five dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression). Each domain has three response levels of severity: no problems, some problems, extreme problems. Scores range from 5 (no problems) to 25 (severe problems). This scale has been used previously in studies with adults with disability [42].

We collected data from the mentors at baseline and 6- and 12-months post intervention measuring psychological well-being (Warwick Edinburgh Mental Wellbeing Scale (WEMWBS)) [43], and QOL (WHO QOL DIS) [36] and Attitudes Towards People with a Learning Disability (ATPLD) [44].

Intervention schedule

We undertook a series of 1−1 interviews and focus groups with twelve adults with intellectual disability, nine mentors, three community group managers and five paid disability staff from residential facilities across the two sites: all were involved in the MATILDA intervention. The interview schedules and data collection were informed by RE-AIM framework [32].

Data analysis

The study was designed as a feasibility trial and hence not powered to evaluate intervention effectiveness. Statistical analyses were conducted in accordance with the published study protocol. As this was a feasibility trial, analyses were primarily descriptive and focused on feasibility outcomes including recruitment, retention, intervention adherence, acceptability and data completeness.

Clinical and health-economic outcome measures were summarised descriptively by treatment group and assessment time point. No formal hypothesis testing was undertaken because the study was not powered to detect between-group differences. Missing data were reported but were not imputed. Secondary and exploratory outcomes were included to assess the feasibility of data collection procedures, estimate variability, and inform outcome selection. A quasi sample size calculation for a future definitive randomised controlled trial was also untaken. Site stratification was incorporated within the randomisation procedure, but no formal stratified outcome analyses were prespecified.

With regards health economics, we assessed the feasibility of calculating quality adjusted life years (QALYs) using both the EQ5DY [41] and WHOQOL-DIS [36] and including spill-over effects experienced by mentors and carers in the evaluation [40] using the Client Receipt Services Inventory. The focus groups were audio recorded and transcribed and analysed using RE-AIM [32].

Progression criteria

Pre-specified progression criteria were established a priori and related to the primary feasibility outcomes (eligibility, recruitment, consent, randomisation, matching to community groups and mentors, attendance, retention, dropout). Achievement of these criteria was used to inform decisions regarding progression to a definitive trial.

Ethics

Ethical approval was obtained from the Office of Research Ethics N Ireland (ORECNI) (HSC REC B (Ref 22/NI/067)).

The trial was conducted in accordance with the published study protocol [31]. Any deviations from the protocol were minor, operational in nature, and are described where relevant within the paper. No changes were made to the study aims, intervention, randomisation procedures, or primary feasibility outcomes.

Baseline characteristics

Baseline demographic, disability and outcome characteristics are presented in Table 1. Overall, participants in the intervention and control groups were broadly comparable, although some minor differences in baseline characteristics were observed. Given the feasibility nature of the trial and limited sample size, these findings should be interpreted descriptively rather than as evidence of between-group differences.

Recruitment and retention

Staff within community statutory and voluntary intellectual disability services advertised the study on our behalf and spoke to potential participants who met the inclusion criteria. For those participants who verbally expressed an interest, their contact details were passed to the research team who then met the person and/or their family/ paid carer. Capacity to consent was assessed and checks were made by the research team to ensure that each participant met the inclusion criteria.

Recruitment took place from July to October 2023 in Northern Ireland, and October 2023 to February 2024 in London (see Table 2). This was an average of approximately 8 participants per month. We recruited 57 participants of our target population of 64 (89%). Because of delays in starting in the London site, recruitment for mentors started in November 2023 and was completed in April 2024.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Progression criteria.

https://doi.org/10.1371/journal.pone.0341940.t002

As shown in Table 2, of the 57 participants, 29 were randomised to the control arm and 28 were randomised to the MATILDA intervention arm. Of the 28 adults with disabilities randomised to MATILDA, two participants unexpectedly died and one fully withdrew. Our intervention arm therefore reduced from 28 to 25 adults with an intellectual disability across both sites.

Matching participants to community groups/ mentor(s)

We were able to match all 25 participants to a community group (100%) (such as Art & Crafts groups, Walking groups, Church Bowl groups, Evening Social Club, Lego club, Men Shed’s, Cuppa & Chat groups, Coffee Morning groups, or a Singing group). These groups were offered from biweekly, weekly, fortnightly to monthly depending on the time of the year and weather.

We were able to match 23 participants with at least one mentor (92%). Most participants were matched to a mentor, although some were matched to 2 mentors in their local community group. One man in Northern Ireland was matched to a local Men’s Shed, and one man in London was matched to a community group, although the community groups did not want to identify a specific mentor. However, all the members of the two community groups offered their support to the person with the intellectual disability on a weekly basis to embed them into the group. Two of the adults were attending their community groups weekly and continued to attend at the 12-month follow-up period.

The matching process

“I think certainly the volunteer coordinator’s input was very good as in it wasn’t a case of bringing X along one week and going, there you go, leaving them alone… they fully supported them.” (Disability Staff Member)

The matching process took longer than anticipated in many cases: we had originally planned 2–4 weeks; but in fact, this took up to 12-weeks. Part of this increased timeframe can be attributed to the time taken to identify community organisations, ensure they were suitable environments, and to explain the MATILDA intervention to group members (potential mentors) in sufficient detail. The VC’s also physically supported the person with the intellectual disability to attend the community group alongside the mentor(s) for several weeks, until the participant and mentor(s) became more comfortable with each other: the length of time depended upon each dyad’s circumstances (range 3–12 weeks). When confidence and trust were developed by both parties, the VC’s gradually withdrew their support over a 6-month period: again, this depended upon each person (range 5–26 weeks). For most participants with an intellectual disability, they asked about the VC when the researcher was collecting the follow-up data at 6- and 12-months, a genuine testimony reflecting the trusting relationship that developed.

Recruiting community groups

“Volunteer coordinator approached us to see if X could possibly attend a community group that I belong to with a view to sort of getting him integrated into the club and becoming part of it. The MATILDA Intervention was a bridge to make that introduction and hopefully then to be a long-term thing. So, not just part of the intervention for six-months. And that’s how it’s worked out for us. X is coming along now under his own steam and is an active part of the club.” (Mentor)

Recruitment and reception of the community groups to the MATILDA intervention demonstrated notable variation across the two sites (see Table 3). In Northern Ireland, 30 older persons’ community groups were approached and most groups agreed to participate. One group declined due to unwillingness to complete the mentor documentation required, although the participant matched to this group continued to attend independently of the study. A second organisation declined due to instability of funding and inability to take on additional perceived work.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Numbers of older persons community groups contacted.

https://doi.org/10.1371/journal.pone.0341940.t003

The London site approached 82 community groups, 51 groups were ineligible for inclusion, due to post COVID-pandemic merging and dissolution of groups or a lack of long-term activities running (see Table 3). Among those 31 groups deemed eligible, thirteen groups declined to take part. Common barriers included the belief that involvement in the MATILDA intervention would be an additional burden on their group members, some community groups were understaffed or already at capacity for new members. Furthermore, there was noted apprehension regarding possible disruption to the existing group dynamics, with some community groups overestimating the risk associated with people with intellectual disability. Additional assurances were required to challenge biases held by group members, as well as clarification regarding practical concerns such as organisational liability and insurance coverage.

Acceptability of the MATILDA Intervention

“I love mixing with people and they like me; I like to join the Bingo. The other members say hello to me, nice to see you. You know.” (Older adult with an intellectual disability)

Findings from the focus groups indicated that the adults with intellectual disability clearly accepted the novelty of engaging with their local community group via the support of their mentor(s) where they both shared common interests. MATILDA was different to other community groups as no other people with an intellectual disability attended. Further qualitative analyses of the focus groups with the adults with intellectual disability and their mentors revealed that this was due to: a) the structural characteristics of the community groups (i.e., sharing similar interests, close to home, appropriate in duration, small group size); b) the careful matching of the person with disabilities to a mentor(s) (i.e., age, gender, interests/ hobbies); and c) the support of the VCs at both sites (i.e., a strong ‘implementation champion’ within the local community, providing ‘check-in points’ and offering continuity).

Fidelity checks

“I thought the VC built a really good rapport with X. I really did. I thought it was lovely…It’s like you felt confident that she had thought of everything, you know what I mean?” (Family member)

“I suppose when the VC was there, it was a bit of a safety blanket. Having someone there constantly beside her, I think that allowed her to be a bit more comfortable and settle in well.” (Disability Staff Member)

Both VC’s in Northen Ireland and London continually contacted the adults with intellectual disability and their mentor by phone, text and ad-hoc visits to the person’s community group regularly over the six-months of the intervention. This also included contacting the family and/or paid carers and community group managers, to ensure the person with intellectual disability was regularly attending, mentor(s) were still facilitating the person to embed within the local group, and to identify and resolve any concerns. These implementation strategies were deemed to be very successful by all as evidenced within the fidelity checks and focus groups with adults with intellectual disability, mentors, family and/or paid carers and managers.

Adherence

Of the 25 older adults with an intellectual disability matched to a local community-based group, attendance records showed that 18 participants (72%) were regularly attending their group at 6-months and 17 participants (68%) were still attending their community group at 12-months. A clear statement that the MATILDA intervention was positively working.

We had matched four older adults with an intellectual disability to a local community group, trained the mentors, and these participants were already attending these groups for several weeks. However, these four participants requested a rematch to a new community group as they described the current group/ mentor(s) as being ‘too old/ boring/ not very posh’, or experienced scheduling conflicts with existing activities such as day center/ college. Three of these older adults with intellectual disability were re-matched successfully with an alternative local community group and mentor(s); however, one participant chose not to be re-matched due to ongoing physical health difficulties that they were experiencing. Another participant transitioned to an alternative local community singing group specific for people with dementia after receiving a dementia diagnosis.

Reach

“I made a planter…growing lettuce and carrots and spuds and stuff like that”. (Older adult with an intellectual disability)

“…we go on the trips and all with them (mentors), like garden centre or for lunch and we get away in the summertime, different places we go with them.” (Older adult with an intellectual disability)

Data collected from the VC’s attendance records, alongside focus groups with the adults with intellectual disability and their mentors, indicated that 12 of the 17 participants (71.5%) engaged in other social activities in addition to engaging with their local community group. Additional activities included involvement in local events (e.g., visits to garden centres, luncheon clubs, bingo, theatre performances, and Christmas pantomimes), and the development of new social connections with individuals outside of their residential setting without disabilities. The adults with intellectual disability also reported acquiring new skills, such as knitting, playing indoor bowls, engaging in creative and craft-based activities, and basic woodwork. This clearly demonstrated the wider community reach and engagement that MATILDA has offered these adults with an intellectual disability. A potential implementation strategy to ensure the older adults with intellectual disability remained in their community group after the VC’s withdrew at 6-months was the development of ‘succession plans’ with all stakeholders. These ‘succession plans’ were arranged before the end of the 6-month period and was key to the longevity of the intervention.

Barriers to engaging with local community groups

“It cost me £20 for the taxi to go to the group, from my house down to X and then from X back up to my house”. (Older adult with an intellectual disability)

“Unfortunately, person in the same group home] he cannot stay home as well without the staff in the building…” (Disability Staff Member)

Using data collated from the focus groups, attendance records, and engagement with the VCs, there were several implementation barriers that prevented some of the adults with intellectual disability from regularly attending their local community group or stopped them attending altogether. These ranged from deteriorating in personal health, low mood/ motivation, holidays, transport, and the group clashing with other appointments/ activities. Another core reason for three older adults with intellectual disability stopping attending their local community group, was that the support workers within their supported living scheme could no longer provide support and/or transport to attend their community group.

Lost to follow-up

All 57 adults with intellectual disability completed the baseline questionnaires. By the time of the 6-months follow-up, sadly three participants had died (two had been randomised to the intervention arm but had not commenced their community group, and one person who died was from the control arm). Another person withdrew completely due to multiple family bereavements, and another person was uncontactable at this data collection point. That meant 52 people had their 6-month data collected (91.2%).

At the 12-month follow-up, data could not be collected from one participant who had developed sudden-onset dementia. In addition, two adults with intellectual disability could not be contacted, and one participant declined to take part in the 12-month assessment, resulting in missing follow-up data for a total of four participants. That meant, of 52 people who had their 6-month data collected (91.2%), we lost another three participants at this 12-month data collection point (94.2%). All the adults with intellectual disability indicated that the data collection process (i.e., time to needed to complete questionnaires, supports (researcher and/or carers) available to complete questionnaires), and the easy read versions of the questionnaires themselves, were all acceptable and helpful.

Mentor’s completion of questionnaires

With regards the mentors, we only were able to get two mentors to complete their baseline questionnaires across both sites. We discussed the challenges with this low response rate with the two VCs and with the mentors themselves during community group visits. Potential reasons for non-completion included age of mentors (over 60 years of age), poor levels of literacy, not understanding why the questionnaires were needed and viewing questionnaires as an additional burden on mentors. Therefore, to minimise the burden on mentors and as a means of encouraging completion at the 6- and 12-month follow-up data collection points, we decided to reduce the number of items on each questionnaire based upon the factor loadings and the questions most relevant to the mentor group. We also attempted to give the questionnaires out at the focus groups at the 6-month follow-up.

The WHO QoL questionnaire examined four domains. We kept all four domains and selected items based upon the highest factor loadings and/or item most relevant to the aims of the intervention. We reduced the items to four items per sub-scale, in total 16 items. Factor loadings were examined using the work of [44] with the most relevant items at face validity extracted.

The Attitudes to People with Learning Disability questionnaire examined four domains, we removed one sub-scale related to the causation of intellectual disability as this was not relevant to the MATILDA intervention. We observed the factor loadings of the other three domains and reduced the number of items to 8, keeping those items with highest factor loadings and/or most relevant to aims of the intervention. This was based on the work of Bollen who wrote that if questions are removed from a one factor model, the factor is still a factor [45]. Despite our efforts, only 8 mentors completed their 6-month follow-up questionnaires, and only 4 mentors completed their 12-month follow-up questionnaires.

Clinical Outcome Data

Table 4 shows that the Glasgow Anxiety Scale scores for the participants in the MATILDA arm decreased from a mean of 20.7 at baseline to a mean of 18.1 at 6-months and a mean of 16 at 12-months. This is compared to the participants in the control arm who displayed a slight decrease across the three time-points (baseline mean (19.6), 6-month mean (17.1) and 12-month mean (18.2)).

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Mean and standard deviations of the secondary outcome measures for the participants in the MATILDA and Control Arms.

https://doi.org/10.1371/journal.pone.0341940.t004

It can be observed in Table 4, that there were minimal differences in the participants scores across the four domains of the WHO-QOL DIS (physical health, psychological health, social relationships, and environment QOL) across the three-points (baseline, 6- and 12-months) for those in the MATILDA and control arms.

Table 4 shows that the Modified Worker Loneliness Aloneness sub-scale scores for the participants in the MATILDA arm decreased from a mean of 3.2 at baseline to a mean of 2.9 at 6-months and a mean of 2.3 at 12-months. This is compared to the participants in the control arm who also displayed a decrease across the three time-points (baseline mean (3.4), 6-month mean (2.8) and 12-month mean (2.3)). With regards the Modified Worker Loneliness Social Dissatisfaction sub-scale there was minimal differences in scores across the three time points for participants in both arms.

Likewise, there were minimal differences in the participants scores across both the Matila and control arms on the WHO (Five) Well-Being Index and EQ-5DY across the three time-points.

Health economics

The data indicates that it is possible to collect health-related QOL data and data on aspects of service use. It was clearly feasible to collect this data across both sites, and up to and including the final 12-month time point. That a range of values are recorded using HR-QOL instruments suggests: first, that there is real engagement with instruments rather than token engagement as might be indicated by clustering in responses or a high level of non-response; second, the instruments can differentiate between individuals in terms of HR-QOL, that is they appear not to be subject to ceiling or floor effects and; third, the low number of missing values at this third time point remains low, suggesting attrition may not be a major factor. These findings are mirrored across both sites.

With respect to resource use, the participants reported using a wide variety of hospital, primary care, social and community care services. In addition to GP services, chiropody, optometry, dentist and social services were among those widely used by participants in both London and Northern Ireland. Responses to the CSRI also demonstrated the adults with intellectual disability readiness and ability to provide detailed information on prescribed medicines including dosage. Variations in use were evident, which similar to the outcomes data, supports the contention that the adapted CSRI has discriminative validity.

Adverse events

There were three sudden deaths, but these events were not related to the intervention.

Strengths/ Limitations

Key strengths include the use of a multi-site randomised design, the focus on an underserved and ageing population with intellectual disabilities, and the comprehensive assessment of feasibility, acceptability, implementation, and preliminary outcome measures. The study generated valuable information on recruitment, retention, intervention adherence, outcome assessment procedures, and economic data collection, all of which are essential for informing the design of a future definitive trial. Furthermore, the RE-AIM framework provided a structured approach to evaluating implementation processes and identifying factors that may influence the scalability and sustainability of the MATILDA intervention in real-world services.

Several limitations should be considered when interpreting the findings. First, the study was conducted across two sites in the UK and included a relatively small sample, which may limit the generalisability of the findings to other settings and populations. Second, as this was a feasibility trial, the study was not powered to detect intervention effectiveness, and any observed clinical changes should be interpreted cautiously. Third, the duration and nature of the control condition may have limited direct comparability between groups over time. In addition, mentor-completed outcome measures were not available for all participants, resulting in some missing data. Fourth, the older adults with intellectual disability, mentors, and research staff conducting outcome assessments were not blinded to treatment allocation, which may have introduced the potential for assessment bias. However, standardised outcome measures and assessment procedures were employed to reduce this risk

Future research should focus on a fully powered randomised controlled trial involving a larger and more diverse sample. Such a study should further evaluate the effectiveness, implementation and cost-effectiveness of the MATILDA intervention, while examining longer-term outcomes relating to community participation, social inclusion and wellbeing.