Health inequities in unscheduled healthcare for children with intellectual disabilities in Ireland : a study protocol

Health inequities for children with intellectual disabilities (ID) Background: are prevalent within different health systems, and children with ID have shorter life expectancies than the general population, higher mortality rates before the age of 17 and have a greater risk of potentially preventable hospitalisations. A health systems approach to research in this area provides a useful means through which research can inform policy and practice to ensure people with ID receive equitable healthcare; however, there is a paucity of evidence regarding how to address differences that have been described in the literature to date. The overall aim of this research is to establish the extent of health inequities for children with ID in Ireland compared to children without ID with respect to their utilisation of primary care and rates of hospitalisation, and to gain a better understanding of what influences utilisation of primary care and ED services in this population. The design of this research adopts a Methods and analysis: mixed-methods approach: statistical analysis of health data to determine the extent of health inequities in relation to healthcare utilisation; discrete choice experiments to explore General Practitioners’ decision making and parental preferences for optimal care; and concept mapping to develop consensus between stakeholders on how to address current healthcare inequities. By applying a systems lens to the issue of health inequities for Discussion: children with ID, the research hopes to gain a thorough understanding of the varying components that can contribute to the maintenance of such healthcare inequities. A key output from the research will be a set of 1 2 3 4


Introduction
The health needs of people with intellectual disabilities (ID) are often complex and this population are known to utilise health services more often than people without ID; an effect that remains generally stable cross-culturally 1,2 . Children with ID have poorer reported health status than children without and such health disparities are more marked in children and young people compared to other age groups 3 . This population has a shorter life expectancy than the general population, with higher mortality rates before the age of 17 in children with disabilities compared to those without 4 . In Ireland, the mortality rate of people with ID under the age of 19 is seven times higher than the general population 5 . Rates of hospitalisation also tend to be higher for this population 3 with findings from Australia showing that children with Down syndrome were hospitalised at a rate five times that of the general population 6 and, in Canada, they have higher ambulatory physician visits and a greater risk of hospitalisation due to injury, respiratory illness and diabetes 4 . A greater risk of further ill-health is exacerbated by an increased risk of socioeconomic disadvantage experienced by children with ID 7 .
Access to healthcare constitutes the fit between the individual and the health system 8 and is influenced by a myriad of complex factors such as availability, utilisation, effectiveness and equity 9 . Access to quality primary care is associated with improved patient outcomes and reduced hospitalisation rates, however, universal access to primary care is not synonymous with equity of access 10 . General practitioners (GPs) experience challenges while treating people with ID related to communication difficulties, deriving incomplete medical histories, lack of knowledge regarding existing supports, and a lack of training 1,11 . GPs have also expressed that time restrictions may impact upon the quality of care that they provide for patients with disabilities 11 . Ambulatory care sensitive conditions are conditions which can be managed with access to timely and appropriate outpatient care 12 . Better access to primary care may decrease utilisation of emergency departments and rates of hospitalisation for such conditions 12,13 .
There is a significant lack of information related to accessing healthcare services for children with ID in comparison to the general population in an Irish context. Ireland has a national database that records details pertaining to service provision for people with ID, which aims to elucidate the service needs of people with ID in Ireland. While it is a valuable resource, not all people with disabilities are registered with the database, and needs pertaining to access and utilisation of primary care and hospitalisation rates are not recorded. Healthcare provision, planning and coordination tend to be poorer for people within the ID population compared to the wider population 3 , and such inequities are amenable to change by improvement in quality of healthcare 14 . The UN Convention on the Rights of People with Disabilities (UNCRPD) sets out that the guiding principles of the convention will need to be considered in relation to existing policy and practice. Article 25 of the convention states that health professionals are required to provide the same standard of care for people with disabilities and outlines the importance of specialist training and appropriate ethical standards in order to meet the needs of people with disabilities. Moreover, Article 31 recommends that state parties use statistical and research data to support policy planning that will give effect to the Convention and are obliged to identify and address barriers that affect the rights of people with disabilities. While it is important to establish any health disparities that exist for this population, it is also vital that models and strategies for reducing any existing inequalities are also addressed 1 . Previous research has suggested that the quality of care GPs can provide people with ID is limited due to factors such as time and lack of knowledge 11,15 . However, to the best of our knowledge, no previous research has sought to systematically examine and model the trade-offs that influence GP referral practices when treating children with ID.
While international evidence suggests that people with ID, including children, experience inequality in accessing healthcare 16 , there is a paucity of evidence relating to the decision-making by frontline staff and parents in relation to this population. Decision-making is critical within healthcare where limited resources are an ongoing concern and ultimately result in a complex interplay between stakeholder choices and behaviours, which often dictate where competing resources are allocated 17,18 . Recognising the complexity of the factors that lead to health disparities for this population, beyond establishing the differences in utilisation, will be critical to identifying avoidable determinants of health disparities and how these can be modified to improve health care provision for this population 1 . For instance, continuity of care has been highlighted as a preference for parents of children with developmental disabilities 4 , whereas socioeconomic disadvantages may influence access to care for this population 8 . Moreover, risk factors for emergency department utilisation and hospitalisation may be unique for this population 19 . Within health services research, public and patient preferences can inform policy and practice, as ensuring the consideration of all viewpoints will increase the likelihood of demands being adequately met when planning for service provision 17 .
There are significant gaps in the evidence base around health inequalities for people with ID whereby, the evidence was generally of low quality and heavily skewed towards psychiatric interventions 20 . While there is a small evidence base on rates of hospitalisations for physical health conditions for this population 3 , there is a paucity of evidence related to access to paediatric health care for children with ID and, critically, the decision-making that drives or inhibits this access. One review examined the factors that influenced access to secondary healthcare for people with ID and found that a myriad of issues affected a persons' experience of care, including poor communication by staff, lack of skills and knowledge about working with people with ID, and poor signage and layout in hospitals 21 . The authors argued for more targeted funding for research that adequately explores the healthcare of people with disabilities. A second systematic review sought to develop a research agenda for primary care research for people with ID and argued that a systems level approach to research would strengthen primary care and improve equitable service 15 .
The proposed research herein will provide information about access and utilisation of primary and emergency care of children with ID, particularly in relation to inequalities compared to children without an ID. Modelling the health status and service utilisation of children with ID at a population level is critical to determine their needs and priorities and to build a robust evidence base for policy, planning and service provision 4 . Rigorous mixed-methods research, which adopts a systems approach, will support the identification of targeted strategies and interventions that will strengthen service provision through policy and practice 15 . The current research project will aim to examine the utilisation of first-contact healthcare for children with ID in Ireland compared to children without ID, gain greater understanding of parental preferences and GP decision-making that drives this utilisation and finally, devise strategies for improved healthcare for this population during a time of large scale changes within the Irish health system

Methods and analysis
The proposed research will employ a mixed-methods approach over three work packages to address the research objectives. Work package 1 will establish the extent of health inequities between children with and without ID with respect to access to health status, attendance at primary and emergency care, as well as rates of emergency hospitalisations. Work package 2 will examine the decision making and referral practices of GPs and elicit parental preferences for unscheduled healthcare for their children. Finally, work package 3 will seek to utilise the evidence that emerges from the first two work packages to develop strategies for improving access to unscheduled healthcare. The culmination of work packages 1 and 2 will seek to provide a comprehensive picture of the multifaceted issues that impact access and utilisation of healthcare for this population, while work package 3 aims to provide a model of strategies generated by stakeholders in order to foster meaningful impact of the research findings.
Work Package 1. Describing differences in health care utilisation and hospitalisation between children with and without intellectual disabilities The aim of this work package will be to compare the use of unscheduled health services and emergency hospitalisation of children with ID to children without ID. A similar approach was employed by researchers in Canada who examined 12 years of administrative data to compare health status and use of health services between children with and without developmental disabilities matched by age, gender and place of residence 4 . This method enabled the team to identify significant differences between the health status of the children with developmental disabilities compared to the matched comparison group, with greater risk of mortality before the age of 17 as well as greater risk of respiratory illness, diabetes, injuries and ambulatory care visits, also a risk factor for children with developmental disabilities. The approach and results were deemed crucial to identifying appropriate interventions to meet the service needs for this population 4 .
A cross-sectional population-based study with a matchedgroup design will be conducted by re-analysing data from an existing database. The databases will contain attendance data for paediatric populations from primary care across Ireland and attendance data from five emergency populations and inpatient data from children admitted to the hospital through the emergency department. Children with ID will be matched 1:3 on specific variables (e.g., age, gender, medical care status, co-morbidities) with children without ID to allow for any differences between the two populations to be established.
The research questions for this work package are as follows: 1. Do children with ID have more attendances at primary care and the emergency department compared to those without in Ireland?
2. Do children with ID have a greater risk of emergency hospitalisations for ambulatory care sensitive conditions compared to those without ID?
3. What is the profile of children with ID who experience unscheduled hospitalisations in Ireland (i.e., age, gender, reasons for emergency hospitalisation [e.g., primary diagnosis], co-morbidities and length of stay)?
4. What is the relative risk of emergency hospitalisations for children with ID compared to those without?

Data analysis plan:
Propensity score matching will be used to match the children with ID to those without ID on specific covariates that may influence health outcomes. Descriptive statistics will be used to profile the patients, while chi-square tests and t-tests will allow comparisons to be made between the children with and without ID on the aforementioned factors and test the statistical significance of any variability. Relative risk (RR) will be used to determine if having an intellectual disability is a risk factor for emergency hospitalisations and 95% confidence intervals will be used to establish statistical significance.
Work Package 2. Exploring and modelling the decisionmaking factors that influence referral practices when treating children with ID and eliciting parental preferences for care A discrete choice experiment (DCE) is a survey-based methodology that elucidates the relative importance of certain factors or attributes that influence decision-making and preferences. They are increasingly being used in health care research as they provide real-world clinically-relevant scenarios to model decision-making at a more granular level by exploring the trade-offs that typically occur when multiple factors are considered during the decision making processes 17,22 . The underlying assumption of this methodology is that services, such as healthcare provision, can be broken into numerous characteristics, with individuals assigning differing values to each 22 . For example, they have been used to examine preferences for access to primary care which highlighted that waiting time for an appointment was only important to patients when attending a new health concern and, from a parental perspective, if the appointment was for a child 8 . DCEs have adequate external validity and have been shown to accurately mimic real-world decisions for choices within healthcare 22,23 . The proposed research will utilise DCEs in order to generate an understanding of the factors that influence GP decision making and referral practices when working with children with ID and to elicit parental preferences for unscheduled healthcare for their children: 1. Exploring decision-making and referral practices of GPs when treating children with ID. The attributes and factors that influence GP decision making and practices regarding referral to the emergency department and wider paediatric services will be examined using the DCE methodology.
2. Eliciting parental preference for primary care for their children. Using the above DCE methodology, the research will seek to model parents' preferences regarding primary care for children with ID In keeping with best practice in DCE design, the following four-step approach will be taken to conduct the DCEs: Step 1. Attribute development: An exploratory step is crucial within a DCE design to establish the attributes that potentially influence decision making and preferences and to define the levels of each attribute 24-27 . Attributes may be patient, service or clinician-focused and can have numerous levels within them. For instance, in a DCE examining preferences for access to primary care, one key attribute was waiting time for an appointment with two levels being within 48 hours and in 4 days 8 . An iterative approach is necessary to select attributes and levels that contribute to most variation in decision making 26 . This will include a systematic review to identify the relevant literature in this area and qualitative inquiry 27 to explore factors that influence GP referral practices with this population and parent preferences related to their child's healthcare.
The qualitative inquiry to develop the attributes will be an iterative two-step process consisting of conceptual development to establish the attributes and then refining the language used to ensure it is meaningful for the intended population 26 . Interviews and focus groups 26 will be utilised with participants including both GPs and parents of children with ID. Purposive sampling will be used to ensure maximum variation of viewpoints are obtained. Topic guides for the qualitative data collection will be developed from the results of the systematic review. The qualitative data will be analysed using the constant comparison approach 28 , which will allow for questions to be adapted in response to emerging data, which is particularly valuable within DCE 26 .
Step 2. Structured prioritisation exercise to finalise attributes: Given the large number of possible attributes that may be relevant in the research, it is important to narrow the focus of the DCE to ensure that the included attributes and levels are feasible and meaningful to the wider research question and to safeguard the face validity of the DCE. A structured prioritisation exercise (SPE) will determine the relative importance of the attributes and factors that emerge from the qualitative work. These will be ranked in order of priority for inclusion in the DCE and to ascertain the levels required for each attribute. Considerations for attribute inclusion will comprise of issues relating to sample size calculation, ecological validity and ensuring adequate information is provided in the DCE. A panel made up of researchers, GPs, a health economist, parents and disability advocates will use the evidence from the SPE to decide on final design of the DCE.
Step 3. Pilot study: An experimental design will be used to generate the DCE choice cards that will allow for combinations of attributes and levels to be presented to participants in manageable subsets, presented in either table format or as vignettes. The result will be a series of hypothetical scenarios that each present combinations of attributes and levels to each participant. These will be piloted to test ease of use and to determine the length of time the DCE takes to complete, as well as establishing the plausibility to ensure clinical validity.
Step 4. DCE and analysis: Purposive sampling will be utilised when recruiting participants and sample size will be determined during the design of the DCE as it is contingent on the number of factors that emerge during the exploratory phase 24 . Recruitment of GPs and parents will target participants to ensure representation based on factors that emerge during the research, e.g., to ensure a geographic spread amongst respondents. The DCE will record relevant participant information such as age, years of experience, training, socio-economic status etc.
Random utility models (including conditional logit models and mixed logit models) will be estimated to establish which factors affect decision making and preferences and the characteristics (e.g., age, level of experience, training, socioeconomic variables) of those making the decisions 18 .
Work package 3. A consensus conference to generate evidence-based strategies for improving access to healthcare for children with ID The final work package will employ a multi-stakeholder concept mapping design within a consensus conference to identify strategies for improving access to healthcare for children with ID based on the results from the two previous work packages and as a means to lessen the gap between policy and practice. Evidence from work package 1 will establish the magnitude of health access disparities for children with ID, particularly relating to a risk of preventable hospitalisations, while the DCEs in work package 2 will explicate GP referral practices and parental preferences for unscheduled healthcare for their children. Work package 3 will build on these findings in order to address areas for improvement by discussing and generating strategies in partnership with patients, parents, healthcare professionals, disability advocates and policy makers. In order to identify workable strategies for improving access to healthcare, it is crucial to engage with stakeholders from all levels of the health system to ensure that the results are aligned with needs and priorities. Concept mapping 29 can provide a structured format for key stakeholders, with varying degrees of expertise, and support equal contribution to the development of strategies for improving access to care and discussing the feasibility of these strategies. This methodology has been successfully implemented to address complex issues in primary care where the delivery of primary care was comprehensively explored by stakeholders to provide a practice index for GP integration 30 . Moreover, a modified version of this approach was used to generate priorities and strategies for improving access to maternity services among women with disabilities who have experienced domestic abuse 31 .
Concept mapping is a methodology that seeks to map ideas or concepts generated by a diverse group of stakeholders through the integration of activities such as brainstorming and unstructured sorting with statistical analyses to map the ideas generated 29 . The result is a visual map that represents the composite thinking of the group and can provide a framework for planning and programme development that incorporates complex elements perceived to be both important and feasible to stakeholders. A modified version of this approach will be adopted in this work package in the following steps: 1. Sampling and recruitment. Purposive sampling will be employed to ensure adequate representation across gender, disability-type and level of ID (i.e., mild/moderate to severe/profound). Up to 30 participants will be invited to participate and will include people with disabilities, parents of children with disabilities, disability advocates, policy makers, health and social care professionals (e.g., GPs, paediatricians, nursing staff, social care workers) and researchers and academics. Invitations will be sent out to GP surgeries, disability services, and children's hospitals.
2. Procedure. The findings from work packages 1 and 2 will be presented to the group. Participants will then engage in a brainstorming session to derive strategies to address the issues, challenges and areas for improvement that emerge from the research in the previous work packages. Participants will then be asked to individually rank the strategies in terms of priority on a 5 point Likert scale ranging from 1 (Very High) to 5 (Not a priority) and in terms of feasibility on a 5 point Likert scale ranging from 1 (Very Feasible) to 5 (Impossible). This step is key in the production of data for concept mapping. The participants will then discuss the means through which the results of the can be used as a framework to enhance policy or practice in order to improve access to healthcare for children with ID.
3. Analysis. Findings will be synthesised and mapped to support the development of a set of recommendations for policy and practice to improve access to healthcare for children with ID. Hierarchical cluster-analysis will be used to rank and identify priority areas from the perspective of the different stakeholder groups. A bivariate plot of the two sets of ratings (priority and feasibly) will produce a 'go zone' graph to map the stated importance and feasibility of the strategies. The plot is divided into quadrants based on the average priority (x axis) and feasibility (y axis) scores where the top right quadrant will represent the strategies rates as being of highest priority and the most feasible. These 'go-zones' will identify the potential courses of action which are highly useful for planning purposes as they provide a detailed outline of the strategies that key stakeholders collectively view as important and feasible.
This procedure will enable the generation of a set of recommendations for policy and practice. Embedding stakeholder involvement throughout this process will provide a valuable mechanism to support meaningful and feasible impact of the research findings.

Ethical considerations
The proposed research has been granted full ethical approval by the University College Dublin Research Ethics Committee (Reference: LS-19-64-Nicholson). All participants will provide written consent on their own behalf for their participation in the study.

Public and patient involvement (PPI)
The project will adopt a disability-centred approach whereby patients, parents and people with disabilities will have input in the design, analysis and dissemination of the research throughout the project. At the outset of the project, a panel of PPI members comprising of parent and patient representatives from the disability community as well as disability advocates will be recruited to the project. This panel will contribute to the governance of the research and will be integral members of each work package sub-committee where they can advise and contribute to the design and analysis of the research. Members of this panel will also facilitate the consensus conference in work package 3.

Study status
Work package 1: Early stages of analysis.
Work package 2A systematic review to inform the qualitative inquiry was recently completed.
Work package 3: Not yet started.

Discussion
The overarching aim of the research is to establish the extent of health inequities for children with ID in Ireland compared to children without ID, with a focus on their utilisation of primary care and rates of hospitalisation, to gain a better understanding of what influences utilisation and avoidable hospitalisations in this population. The focus and timing of this research is of importance given the ratification of the UNCRPD in Ireland.
The convention sets out guiding principles that will need to be considered in relation to existing policy and practice; however, there is a paucity of research evidence in Ireland to ascertain health inequities for this population. Applying a systemsapproach, whereby the various components of the health systems in which the health inequities and the intricacies of the relationships between these will be considered, can provide evidence towards understating health inequities for this population and seek to identify interventions that can address them 15,32 .
While describing the health inequities experienced by people with ID is an important element for study, research must endeavour to establish the causal factors behind inequities, such as increased hospitalisations 3 , and examine whether they can be considered avoidable and unjust 1 and critically, amenable to improvements in the quality of care 14 . The inclusion of the DCE methodology can provide explanatory models that have important applications within policy making in health and, critically, can point to modifiable factors that contribute to any disparities in referral patterns experienced by this population.
GPs are required to make highly complex decisions within health systems of scarce resourcing and thus, establishing the trade-offs they dictate, their behaviour is important for planning purposes 25 . Moreover, preferences for healthcare beyond outcomes alone are important for planning quality care 18 and eliciting parents' stated preferences for first contact care for children with ID will be key to identifying opportunities for interventions that can support meaningful change. Finally, visits to the emergency department and subsequent emergency hospitalisations are more expensive than primary care 19 and generally, overcrowding can result in poorer health outcomes for children 33 .
In keeping with the disability-centred approach of the project, ongoing and continuous knowledge exchange activities will be adopted to support an accessible and wide-reaching dissemination plan. Drawing on an evidence-based model for knowledge transfer of health research 34 , dissemination will strategically target stakeholders with materials designed specifically for their needs and through avenues chosen to maximise their reach.
Four key components will be considered when disseminating the research: messages, stakeholders, processes and contexts. For instance, workshops will be held with health care professionals to outline the results from the studies and receive feedback. A key output from the project will be a set of recommendations developed in collaboration with stakeholders that will be aimed at policymakers. Alongside traditional peer-reviewed publications, accessible materials, such as infographics, will be created to communicate the results from each work package to the disability community. These will outline the results in clear and accessible formats using lay terminology and will be designed with the PPI panel members and developed by a graphic designer to ensure their suitability for dissemination to the general population.

Limitations
Potential inconsistencies in the reporting of ID within the Irish health system may be a limitation for the proposed research, as it may hinder our ability to accurately identify patients with ID from the data administrative systems. For instance, people with mild ID may be significantly underrepresented in the data systems as their care needs will not be considered as great as people with severe and profound ID. Understanding the extent of reporting of ID will be beneficial in and of itself, in order to raise awareness of the differences in health presentations for this population, which may not be widely known among those involved with practice and planning 5 . Monitoring trends and identifying any health disparities for this population is crucial for the development of appropriate interventions that will facilitate good health for people with ID 35 . The recent ratification of the UNCRPD requires the government to use research and statistical data to develop policies that give effect to the Convention 19 and the availability of appropriate data is key for policy and planning purposes.

Conclusion
The proposed programme of research will apply a systems lens to the issue of health inequities for children with ID, specifically in relation to unscheduled healthcare. Across three work packages, the research will establish the extent of inequities for this population in relation to utilisation of unscheduled health services, elicit parents' preferences for their children's healthcare, as well as model the factors that influence GP decision-making. Finally, the study will seek to identify feasible solutions and interventions that can address health inequities for this population.

Data availability
No data are associated with this article.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.