Addressing Age-Related Complexity in Intellectual Disability (AARC-ID): an economic analysis of different support models. Study protocol.

2.1 Population

WP1–3: IDS-TILDA provides a sample of older adults (aged ≥40 years) with intellectual disability to establish influences on ageing and health in this cohort. To date, there have been five data collection waves since 2009. IDS-TILDA has shown successful retention of participants, with 87.1% of surviving Wave 1 participants retained (McCarron et al., 2022), attributed to a personal, flexible, and innovative approach to the conduct of this longitudinal study. At Wave 4, the sample was refreshed to replace participants in the 40–49 years category who were, at that stage, ≥50 years. The sample was replenished again at Wave 5, with 141 participants recruited, totalling 762 at wave 5. Table 1 provides the number of participants at each wave.

WP4: We will recruit older adults with intellectual disability, family members, caregivers and service providers through services as well as using a social media campaign. Recruitment details for this are provided in section 2.3 below ‘Work package 4 – To elicit the perspectives of older people with intellectual disability, their families and caregivers, and service providers on the different models of support and changes needed to the mix of available services.’

2.2 Sources of data

This study uses data from multiple sources – IDS-TILDA, NASS, previous calculated costs from the HSE Disability Strategy and Planning division as well as qualitive data collected by the research team. A summary table of data sources is provided below, indicating the relevant work package the data relates to, see Table 2.

IDS-TILDA collects data on a rich array of individual characteristics and activities including demographics (for example, age, sex, early life experience), residence, health (for example, diagnoses, cognition, functional limitations) and health care use. This is completed via a pre-interview questionnaire (PIQ) and a face-to-face computer -assisted personal interview (CAPI). The PIQ is sent to each participant a minimum of one week prior to the CAPI; questions asked in the PIQ may require respondents to check records and this ensures they have time to do so. The PIQ is collected by the field researcher when they conduct the CAPI. All field researchers receive full training prior to beginning data collection. IDS-TILDA facilitates for different levels of intellectual disability with different interview styles employed: respondent only interview conducted directly with the individual, interview with the person supported by a family member or caregiver, or, a proxy interview with a family member or caregiver very familiar with the person [must know the participant for a minimum of 6 months (McCarron et al., 2022)].

NASS is an amalgamation of the National Intellectual Disability Database (NIDD) and the National Phsyical and Sensory Disability Database (NPSDD). NASS is a national database that collects information on HSE funded disability services, received by people who have a intellectual disability, developmental delay, physical, sensory, neurological, learning, autism spectrum or speech/language disability (Casey et al., 2019). It collects a core set of information for each person, including demographic variables (such as age, gender, level of disability, present living arrangements) and the support services they currently receive (such as residential campus accommodation, day services, respite breaks). Participants are those with an intellectual disability and utilising, or in need of, a specialist service now or in the next five years (Dodd et al., 2010). The data is used for management of services.

Existing costing data from the National Disability Authority (2021) derived from observational field research data and surveys to calculate staffing and associated pay costs, published in ‘Moving In, Moving On’ will be utilised. Previous healthcare costs calculated by co-author SS (Smith et al., 2021) will be used in conjunction with National Disability Authority estimated costing. Co-author, TO’B, has previously completed multiple costing exercises within the HSE Disability and Strategy and Planning and will support with estimated costing data, where necessary. This may include previous costing research on total cost per residential campus placement between service provider and HSE commissioning area (Cullinan et al., 2024).

2.3 Work packages

This study includes five work packages, four individual and one overlapping work package, ensuring that the aim stated above is achieved. The five work packages are presented below.

Work package 1 – Estimate the total volume of support need at the population level, 2024–2034

Objectives

1.1 – Estimate the number of people in Ireland in 2024 aged ≥40, with intellectual disability, and living (semi-)independently, in a family home, or community group home.

1.2 - Estimate the associated support needs of this population in 2024 taking into account age-related complexity

1.3 - Estimate the total volume of support need at the population level to 2034.

Study design

To achieve objective 1.1, a descriptive secondary data analysis will be conducted using the NASS. This will provide information on the numbers of adults aged ≥40 by intellectual disability level, age, sex and living arrangements.

This information will then be combined with analysis of IDS-TILDA data, to meet objective 1.2. Data from all five waves of IDS-TILDA will be explored and the participants reporting residence as independent or community group home will be extracted (see Table 3 for all categories of residence within the AARC-ID study). Demographics including age, sex and complexity (such as neurological conditions, physical health, mental health, behaviours of concern, functional limitations) of this cohort will be investigated. Levels of support will be defined, and each participant assigned a level of support. The levels of support were defined using a modified version of a pre-established framework on support (Cullinan et al., 2024), to be more specific to the population of people with intellectual disability and to correspond with available data from IDS-TILDA. See Appendix 1 for modified level of support framework. These defined levels of support may be reviewed in light of accessing IDS-TILDA data.

Residence categories will include different residence types, as presented in Table 3 below.

Please note that residence types comes from residential options provided to IDS-TILDA participants. Participants, or their proxy, self-select the residence type they believe most strongly represents their current living situation, at the time of interview. The research team is aware that there may be cross over between some categories. For example, a person with intellectual disability can live semi-independently but also in the family home.

For participants who have participated in multiple waves, they will be assigned a level of support based on their information provided at each timepoint. Therefore, the level of support assigned to a participant may have increased from the previous wave. In this study, we are defining level of intellectual disability as non-modifiable and level of support as non-reversible. An individual can transition from any support level to any higher support level, providing that this is possible via a change in health circumstances; transitions requiring a change in level of intellectual disability are considered impossible.

Population weights will be used to adjust for differences in age, sex and level of intellectual disability between IDS-TILDA participants and the population sample (generated from the NASS data). The prevalences of complexity and support need will be combined with the NASS data to estimate the total volume of support need at the population level in 2024, thus completing objective 1.2.

To achieve objective 1.3, population simulation method will be utilised to project support needs among future population with intellectual disability in Ireland to 2034.

Three files will be created – stock file, transitions file and a replenish file. The first will be a stock file aggregating the baseline cohort of all people in the population of interest in 2024 by sex and age band (40–49years, 50–59years, 60–69years, 70+). This will comprise of the data gathered as part of objective 1.2. Secondly, a transitions file representing the probability of individuals within each sex-and-age band having growing support needs through changing health in a one-year cycle will be documented. This will be derived from IDS-TILDA data, with adjustments for time interviews between waves to inform one-year incidence rates (IDS-TILDA waves are every three years). Markov models will be utilised to estimate the probabilities of moving between levels of support need and the probability of mortality, conditional on observed predictors: age, sex, health, and disability. Thirdly, the replenishing file will be generated by combining NASS data on people aged under 40 and IDS TILDA data on the distribution of support needs among 40-year-olds.

Sensitivity analyses using high and low confidence interval bands, will be conducted on transition probabilities and to cover for the estimated number of the population at the level of need due to sampling variation in IDS-TILDA. Where possible, we will externally validate IDS-TILDA transition probabilities against NASS data. WP 1 has begun and findings from this will feed into work package 2.

Ethical approval from Faculty Research Ethics Committee in the authors institution is already in place regarding IDS-TILDA data. Aggregated data will be received from NASS and therefore ethical approval is not required. Necessary data is being requested from IDS-TILDA and NASS.

Output

This work package will be submitted for peer review in an open access journal. Where possible, and in line with General Data Protection Regulations 2018 [GDPR (GDPR, 2018)], the three data files (stock, transition and replenish) and simulation data will be made available for reviewers and readers to ensure transparency and reproducibility. An accessible easy read of the finalised peer review article will be produced. Results may be presented at academic conference.

Work package 2 – Estimate the costs and outcomes associated with different models of support for those with different support needs, at micro level

Objectives

2.1 – Define the four models of transition to be costed

2.2 – Estimate the costs associated with meeting different levels of support need in each of the models

2.3 – Estimate the quality of life outcomes associated with each of the models

Study design

We will collaborate with the NFVSP, HSE Office of Disability Strategy and Planning and HSE Home Support Reform Unit to define four models of transition. Initially, these are i) status quo, ii) ageing in place with additional supports (for example, home care), iii) residence in a specialised setting with intellectual disability appropriate expertise, iv) residence in a nursing home with the general population. Specific personnel inputs required, particularly in ii and iii, will be derived from expert knowledge. Additionally, best practice residential guidance for people with intellectual disability who develop dementia (National Intellectual Disability Memory Service Support Materials) will be consulted during this process. This will complete objective 2.1.

To achieve objective 2.2, we will utilise existing calculated unit costs. In existence are estimated unit costs of different types of community group homes for people with intellectual disability (National Disability Authority, 2021) and standardised estimated health care costs for general population [including nursing home residence (Smith et al., 2021)]. The HSE Disability Strategy and Planning division will provide access to their existing calculated costings. Consideration will be given to the most appropriate methodology for identifying and costing inflationary costs. The levels of support need in each of the models will be amended based on findings from objectives 1 and 2.1.

Completion of objective 2.3 will be achieved through secondary analysis of IDS-TILDA data. For this specific objective, residence will be categorised as family home, (semi-)independently, community group home, residential campus, nursing home, or other. Multivariate logistic regression will be conducted to investigate the association between quality of life and place of residence. Where possible, we will derive causal estimates of relationships using data from participants who have moved between different residential settings and change of quality of life, if any. Essentially, we aim to answer if a participant moved place of residence, did this impact quality of life? This will be in comparison to participants who did not report a change in place of residence between waves. Additionally, we will update the systematic review of quality of life in decongregation for people with intellectual disability, led by team member MMcC (McCarron et al., 2018).

Output

The models, costing and quality of life outcomes feed into work package 3. A publication from work package 2 is anticipated; an updated systematic evidence review on quality of life outcomes and costs associated with moving from congregated settings to community living arrangements for people with intellectual disability will be published in a open access peer reviewed journal.

Work package 3 – Estimate the costs and outcomes associated with meeting population- level needs to 2034 under different funding scenarios for the models of transition

Objectives

3.1 – Define the policy/funding scenarios that will be compared – status quo including limited additional funding for ageing in place, and alternatives to the status quo

3.2 – Estimate the costs and outcomes, and where possible, the cost-effectiveness of the different policy/funding scenarios, taking into account declining availability of unpaid family care, at the population level

Study design

Within objective 3.1, we will define the scenarios. At this initial stage we expect the status quo to include limited new resources for ageing in place and will assume no new places created in context-appropriate settings (either group homes or specialist settings). Costs and outcomes under the status quo will be compared to different approaches. These include i) providing sufficient resources for people to remain in their current place of residence and ii) developing new services that meet low-support needs in the current home and high-support needs in a specialist setting.

This objective will be achieved through collaboration with NFVSP and HSE to clarify various budget scenarios. Population-level costs and outcomes associated with each scenario will be calculated using data previously gathered in work packages 1 and 2. Different budget levels between two extremes will be considered including the quantifying of trade-offs regarding spending in insufficient budget. For example, if only half of the funds are available, we will investigate the trade-off of supporting a relatively large number of low-support need people with intellectual disability or a relatively low number of high-support need people with intellectual disability. This objective will produce budget scenarios with clear rules on available funding as well as options for prioritising insufficient funding.

To complete objective 3.2, IDS-TILDA participants living in family homes or community group homes will be modelled in a two-step process. Firstly, a simulated model will calculate the change in support need between year t and t+1. Here, we will use the data from WP 1 to model distribution of support needs in 2024, transition probabilities and the replenishing cohorts for living in family home or community group homes separately. For future years, this will be calculated for all living situations. Secondly, living situation at each year will be determined, using the decision trees identified in Appendix 2 and Appendix 3.

The costs of the different scenarios will be estimated by combining unit costs from WP 2 with the estimated numbers of population within each group. Quality of life outcomes previously calculated as objective 2.3 will be included in the analysis here. To limit the number of possible transitions, and to improve study feasibility, policy scenarios will be considered fixed and specialist setting considered absorbed (i.e., once a person moves in, they do not move out). However, support need and mortality within these specialist settings will be monitored as this will impact costs and the availability of spaces. Sensitivity analyses will be conducted to account for possible variation in unit cost estimates. Within this work package, availability of home care supports and community group home places for different levels of need are determined by policy scenario. Capacity in specialist settings may become available if a person dies under the status quo. Capacity in nursing homes will be considered as infinite as people with intellectual disability are a relatively small population and if a suitable setting is not available then typically, they will be placed in a nursing home.

Output

This work package will submitted for peer review in an open access journal. Where possible, and in line with GDPR, costs, quality of life scores, transitions and mortality figures will be made available for editors, reviews and readers to ensure transparency and reproducibility. An accessible easy read of the finalised peer review article will be produced. Results may be presented at academic conference.

Work package 4 – To elicit the perspectives of older people with intellectual disability, their families and caregivers, and service providers on the different models of support and changes needed to the mix of available services.

Objectives

4.1 – to conduct individual interviews or focus groups with older people with intellectual disability

4.2 - to conduct individual interviews or focus groups with family members and/or caregivers to people with intellectual disability

4.3 – to conduct individual interviews or focus group interviews with service providers

Study design

This work package will employ a descriptive qualitative approach, given its fidelity to participant experience, as well as its acknowledged simplicity and flexibility (Doyle et al., 2020; Kim et al., 2017). The Consolidated Criteria for Reporting Qualitative Studies (COREQ) reporting guidelines (Tong et al., 2007) will be utilised here.

A purposive sampling strategy will be employed to recruit older adults with intellectual disability, family members/caregivers and service providers. Consideration will be given within the purposive sampling strategy to ensure that the experience of people who have intellectual disability and autism is considered. Recruitment of participants in this work package will be conducted via at least two service providers and using social media. We anticipate having at least one service provider classified under Section 38 and one service provider under Section 39 of the Health Act 2004 (Government of Ireland, 2004). IDS-TILDA has established relationships with service providers and this will be utilised to support recruitment to the study. Application for ethical approval from level 3 ethics committee in the author’s institution has been submitted. Ethical approval from each of the service providers will be sought. Each service provider will appoint a gatekeeper who will distribute study information to potential participants. This will include an invitation letter, participant information leaflet and consent form; accessible easy read format of these will be distributed to adults with intellectual disability.

The Trinity Centre for Ageing and Intellectual Disability (TCAID) will advertise this study on their social media channels (LinkedIn, X formerly Twitter) as will the NFVSP. This recruitment method will be used to recruit family members or caregivers to adults with intellectual disability and service providers. Those interested will be asked to contact the researchers who will provide study information (invitation letter, participant information leaflet and consent form). The HSE National Clinical Programme for People with Disability (NCPPD) will also advertise the study via their networks.

As is common in qualitative research, data adequacy will determine sample size. It is estimated that 10–15 individual interviews will be sufficient for this (Constantinou et al., 2017; Hennink & Kaiser, 2022). To achieve thematic saturation in focus groups, it is anticipated that three to eight focus groups is sufficient (Hennink & Kaiser, 2022; Namey et al., 2016). However, the characteristics of the focus group will be important in determining the number of participants within the focus group. For example, a focus group with people with intellectual disability may need to be smaller than a focus group with people who do not have intellectual disability. It is acknowledged that sample size in qualitative research is a contested issue (Sebele-Mpofu & Serpa, 2020) and therefore the number of interviews (individual or focus group) will be reviewed as data collection progresses within this work package in line with the study and work package aim (Young & Casey, 2019).

Individual interviews and/or focus groups will be conducted at a time and place convenient to both the interviewer and the interviewee(s). The research team will be flexible and make adjustments to facilitate participation of people with intellectual disability. Photo/object elicitation may be used where appropriate in addition to the interview topic guide, when interviewing people with an intellectual disability. A support person may be present if requested by a participant with intellectual disability. This will be explained in the participant information leaflet. Focus groups and interviews will be conducted face-to-face (likely at the location of the service provider) or online. Face-to-face interviews will be recorded using a digital audio recorder. As soon as possible after the interview has been completed, the recording will be uploaded to the secure study folder and deleted from the digital audio recorder. Interviews conducted online will be held via Zoom and recorded using the ‘record’ function; only the audio file will be saved. As soon as possible after the interview has been completed, the audio recording will be uploaded to the secure study folder.

Interviews will be transcribed verbatim, however, any identifying factors (such as participants’ names or names of services) will be pseudonymised. Transcripts will be transcribed by a transcription company who has existing contractual agreements with Trinity College Dublin. Recordings will be securely transferred to them to transcribe, as will be outlined in the agreement. As soon as the researcher has confirmed the accuracy of the transcription, the original recording will be deleted. Transcripts will be pseudonymised to allow for potential secondary analysis of transcripts in future related projects. This will be outlined in the participant information leaflet and participants will agree to this in the consent form. A key code, identifying the transcript to the participant will be stored in a password protected file in the study folder on the secure TCAID server. This folder will only be accessed by key study members (MM and AG) and key members of TCAID including centre directors, manager and data management team. The key code, along with pseudonymised transcripts and consent forms, will be securely stored for 5 years in line with current GDPR and Health Research Regulations 2018. All participants will receive a certificate of participation.

The reflexive thematic analysis approach of Braun and Clarke (2006, 2022) will be used to analyse transcripts. This includes six phases as shown in Table 4. This approach views researcher subjectivity as a resource, analysis as an interpretive practice and meaning as unfixed (Braun & Clarke, 2023). Applying reflexive thematic analysis enables the recognition of patterned meaning across a study dataset with reference to the stated study aim (Braun & Clarke, 2021).

Output

A publication from this work package will be submitted for peer review and publication in an open access journal. An accessible easy read of the finalised peer review article will be produced. Results may be presented at academic conference.

Work package 5 – To ensure key stakeholders are involved in a meaningful manner throughout the lifecycle of the project, including the dissemination of key findings

Objectives

5.1 – to ensure patient and public involvement (PPI) throughout the research cycle of this project

5.2 – to conduct multi-channel dissemination to the intellectual disability community, family members/caregivers, policymakers, service providers, general public and other key stakeholders

5.3 – to publish a statement from the research team synthesising outstanding questions, limitations and next steps

IDS-TILDA have established relationships with advocacy groups and service providers throughout Ireland. Through these existing relationships, key stakeholders, including people with intellectual disability, will be invited to participate in the development of the project via involvement in a Steering Group. Recruitment to this will be circulated through the research centre’s social media, and the projects existing collaborations with NFVSP and HSE. It is anticipated the steering group would meet twice yearly online via Zoom. The Steering Group will assist with the study directions and monitor the study’s progress and outcomes and make recommendations.

TCAID operates a PPI panel; the panel will be consulted as required throughout the research project. This would involve reviewing study results, reviewing the recruitment documents for people with intellectual disability in work package 4 as well as the study materials such as topic guide and accompanying images. It is anticipated that dissemination will involve the establishment of accessible easy read findings report which the PPI panel would be involved with.

This project will engage in national and international dissemination via peer review publications in open access journals, conferences and invited talks.

3.1 Strengths

IDS-TILDA is the first longitudinal study on ageing among people with an intellectual disability in Europe. The measures are repeated with the same population at five timepoints allowing for trends to be tracked.

To ensure this project is person-centred and valuable insights are gathered, PPI has been included in the project to date and will continue to the dissemination stage. The AARC-ID project will establish a independent Steering Group to provide advice on the project and review any issues which might arise.

3.2 Limitations

Due to the range of information required to assign a level of support, this may result in a high level of missing data.

IDS-TILDA does not gather information from healthcare professionals’ practice notes and therefore the reporting of conditions such as mental health condition is a self-report of having received a doctor’s diagnosis. As is common in epidemiological retrospective studies, it is acknowledged that recall bias may have impacted responses provided.