Barriers and enablers to optimal diabetes care for adults with learning disabilities: A systematic review

Background: Individuals with diabetes and a learning disability have poorer health outcomes than those without a learning disability. In the UK, the health inequalities faced by people with learning disabilities are often the result of barriers they face in accessing timely, appropriate and effective health care. The aim of the study was to review relevant literature to identify the barriers and enablers to optimal diabetes care for adults with learning disabilities. Methods: Systematic review methodology was used to answer the research question: What barriers and enablers to optimal diabetes care exist for adults with learning disabilities? This review is reported according to PRISMA guidelines. Key databases were searched using relevant terms. Included studies were


| INTRODUC TI ON
Individuals with learning disabilities have poorer health outcomes than those without learning disabilities (NHS Digital, 2020). In the UK, the health inequalities faced by people with learning disabilities are often the result of barriers they face in accessing timely, appropriate and effective health care (Emerson & Baines, 2011). This inequality applies to specific health conditions. For example, evidence shows that diabetes management for individuals with a learning disability and a diagnosis of diabetes are poor (Shireman et al., 2010). The ongoing COVID-19 pandemic shines a spotlight on these inequalities. Individuals with learning disabilities were estimated to have had a death rate from COVID-19 between 4.1 and 6.3 times higher than the general population in England in Spring 2020 (UK Government, 2020). In addition, those with diabetes who acquire COVID-19 have a greater risk of severe illness (Muniyappa & Gubbi, 2020).
Diabetes mellitus is a metabolic disorder that results in abnormally high blood sugar levels (hyperglycaemia). There are two main types of diabetes mellitus: type 1 or insulin-dependent diabetes mellitus and type 2 or non-insulin-dependent diabetes mellitus. Insulin is required to allow the human body's cells to absorb and use sugar (glucose). Type 1 diabetes is characterised by individuals being unable to produce insulin at all, whereas type 2 diabetes produce insulin, but this is less effective and eventually leads to reduced insulin production (American Diabetes Association, 2010). Type 2 diabetes is the more prevalent of the two. Between 85 and 95 per cent of diabetics have type 2 diabetes, and 8 per cent of diabetics have type 1 diabetes (Diabetes UK, 2020).
Optimal diabetes management is required to prevent chronic hyperglycaemia and long-term complications. Potential complications include kidney damage leading to renal failure; peripheral neuropathy with a high risk of foot ulcers and amputations; and autonomic neuropathy causing cardiovascular, genitourinary and gastrointestinal symptoms. Patients with poorly controlled diabetes have an increased incidence of cardiovascular and cerebrovascular disease (American Diabetes Association, 2010).
Diabetes is more prevalent in people with learning disabilities than in the general population (McVilly et al., 2014). For those with learning disabilities, the estimated prevalence is between 9 and 11% (compared with 4%-5% in the general population) (Emerson, 2011;MacRae et al., 2015). Individuals with learning disabilities have a higher rate of hospital admission resulting from diabetes-related conditions that are usually managed in an outpatient or community setting (Dunn et al., 2018). The aim of this study was to review relevant literature to identify the barriers and enablers to optimal diabetes care for adults with learning disabilities.

| ME THODS
Systematic review methodology was used to answer the research question: What barriers and enablers to optimal diabetes care exist for adults with learning disabilities? The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework guided the reporting of this review (Moher et al., 2010). Prior to commencing this research, the International Prospective Register of Systematic Reviews (PROSPERO) and the Cochrane Library were searched to ensure no similar reviews had previously been completed. The review protocol was not registered.

| Search strategy
Searches were conducted using electronic databases (for peerreviewed and grey literature), hand searching of reference list citations and search engines for policy documents. The population, interest (phenomenon of) and evaluation (PIE) study design framework was used to formulate the review question and identify search terms (Table 1) (Methley et al., 2014). The search strategy used free-text synonyms for "learning disabilities" such as "intellectual impairment" and "cognitive deficit" (Appendix A).
These synonyms were obtained from a 2017 systematic review examining a similar population (Robertson et al., 2018

| Selection criteria
All publications that met the following criteria were included: (i) literature related to adults with a learning disability (those aged 18 years old and over); (ii) literature related to formal or informal caregivers of individuals with a learning disability; (iii) any type of publication (including peer-reviewed articles and grey literature); and (iv) all types of study design that include some form of qualitative data collection.
Publications that met the following criteria were excluded from the review: (i) article not available in the English Language (due to the time constraints of this project); (ii) population described in the article do not have learning disabilities; (iii) no reference to management of diabetes; and (iv) no enabler or barrier to optimal diabetes care identified.
All the identified citations were imported into Mendeley, and duplicates were removed. Articles were screened at the title and abstract level, and the full text of potential articles was retrieved and assessed for inclusion based on the eligibility criteria.

| Quality assessment
Studies were evaluated using an adapted version of the Critical  Verboom et al. (2016) to include items that consider descriptive, interpretive and theoretical validity. This tool, labelled the "CASP instrument," is designed to aid interpretation of review findings by assigning an overall score of methodological quality (either "high quality," "moderate quality" or "low quality"). A copy of the CASP instrument template is included as an additional file (Appendix B).

| Data analysis & synthesis
Qualitative data were extracted from the identified studies, and thematic analysis was carried out using the methods described by Thomas and Harden (2008). A judgement was made on the reliability of each identified theme. A reliability index based on that used by Bach-Mortensen et al. (2018) was used to combine the quality of each study that contributed to the identified themes, and the number of times each theme was represented in the included studies.
Reviewer 1 (BH) conducted the literature searches. If any concerns were identified, these were discussed with Reviewer 2 (AL). Likewise, the themes were identified by Reviewer 1 (BH) and double-checked by Reviewer 2 (AL).
These themes were evaluated through narrative synthesis, and the identified themes were applied to the social-ecological model to assist evaluation. The social-ecological model is a theory-based framework that uses an ecological approach to identify factors (such as barriers and enablers) that occur at multiple social and ecological levels (demonstrated in Figure 1) (Ohri-Vachaspati et al., 2015).
It was anticipated that optimal diabetes care for those with learning disabilities is impaired and facilitated due to the inter-

| Characteristics of included publications
The ten publications in this review include eight reports of research studies and two guidance documents identified from the grey literature. All references relate to work undertaken in highincome countries and have been published within the last nine years.
Of the eight research studies, six studies use a qualitative study design and semi-structured interviews for data collection methods.
These qualitative studies included between 4 and 29 participants.
The settings of each study vary, but include a number of learning disability services and residential locations Cardol et al., 2012a;Cardol et al., 2012b;Dysch et al., 2012;Hale et al., 2011;Maine et al., 2017).
One research study (Maine et al., 2019) is a process evaluation that uses the updated Medical Research Council guidelines -Vachaspati et al., 2015).
for evaluating complex interventions in health. This included focus groups with participants. The final research study (Taggart et al., 2018) is a two-arm, individually randomised, pilot superiority trial undertaken to assess the feasibility of an intervention.
A report produced by a diabetes charity, Diabetes UK, summarises how individual services can make reasonable adjustments to improve the care of diabetes for adults with learning disabilities and draws advice from published research (Diabetes UK, 2018). NHS RightCare (2017) has produced a health policy document that is included in this review. They outline how commissioners and providers of diabetes services can enable reasonable adjustments for those with a learning disability. Further detail relating to each of the included studies can be found in Tables 2 and 3.

| Quality appraisal
The CASP instrument (Appendix B) was used to critically appraise the eight research studies included in this narrative review. The two guidance documents from Diabetes UK (2018) and NHS RightCare (2017) were not quality-assessed as they did not include any qualitative research methodology or findings. A summary of the quality assessment results is shown below in Table 4.

| Thematic analysis
The synthesis identified 12 unique factors operating as barriers across the 10 publications. The barrier reported by the most publications was "low level of diabetes knowledge and understanding." 14 unique factors operating as enablers to optimal diabetes care were identified from the 10 publications. The enabler reported by the most publications was "person-centred planning and reasonable adjustments." These results are presented in Tables 5 and 6.
The identified barriers and enablers were grouped together into themes and allocated to the appropriate level of the socialecological model. For both barriers and enablers, the identified factors were spread across all four levels of the model.
A judgement was then made on the reliability of each theme to appreciate the consistency between the identified barriers and enablers and the quality of the studies reporting them (see Key for reliability rating in Tables 5 and Table 6) (Bach-Mortensen et al., 2018).
Two barriers are considered to have high reliability as they have been supported by high-quality studies. The majority were recognised as having medium reliability with only "independent living" considered to have low reliability.

TA B L E 2 Publications included in this review
The identified enablers to optimal diabetes care were grouped into 14 separate themes. One enabler had a high reliability score, and two enablers were considered to have low reliability. Two facilitating factors, training for staff and caregivers and social barriers addressed, did not receive reliability scores. This was because they were extracted from health policy documents, which were not appropriate for quality assessment due to the lack of research study design. No reliability judgement could be made due to the lack of quality assessment. Much like the identified barriers, the majority of enablers were considered to have medium reliability.

| DISCUSS ION
This review has identified 12 barriers to optimal diabetes care for adults with learning disabilities and 14 enablers to optimal care from 10 diverse publications. Where possible, quality assessment pro- However, these education programmes need to be adapted to the needs of this population.
In the UK, the general population are encouraged to attend one of two programmes specific to their diabetes type. DAFNE (Dose One barrier that an integrated healthcare system may be able to address is the finding that current systems do not provide the capacity or opportunity for reasonable adjustments. This may be due to a large amount of pressure on services, reduction in service budgets or restructure of specialist and community services.
Alternatively, there may be a lack of awareness of the need to make these adjustments. This finding highlights an urgent need for further system change to recognise and allow reasonable adjustments for adults with learning disabilities living with diabetes.
It is important to consider the barriers and enablers described in the context of the current global situation regarding COVID-19. Public Health England analysis shows that people with learning disabilities are estimated to have had a death rate from COVID-19 between 4.1 and 6.3 times higher than those without (UK Government, 2020). The risk is greater for specific learning disabilities, with studies estimating a fourfold increased risk of COVID-19-related hospitalisation is four times greater, and COVID-19-related death is 10 times greater in people with Down syndrome . Another study estimates the risk of COVID-19-related death is over 32 times higher in females with Down syndrome (Clift, Coupland, Keogh, Diaz-Ordaz, et al., 2020). As previously discussed, individuals with learning disabilities are known to have poor physical health, including higher rates of obesity and diabetes, putting them at increased risk of severe illness and death from COVID-19 (Emerson et al., 2016). This clearly demonstrates the importance of optimal diabetes management and the wider impacts of barriers to accessing timely, appropriate and effective health care for chronic health conditions.

High reliability
The identified barrier is supported by several studies of medium quality and one high-quality study, or the study is supported by at least two high-quality studies.

Medium reliability
The identified factor is supported from several medium-quality studies, or the factor is identified from at least one high-quality study.

Low reliability
The identified factor is supported by several low-quality studies and/or one medium-quality study.
a It should be noted that the confidence in these findings (reliability) does not relate to the generalisability of the findings.

| Limitations of the study
Publication bias may exist in this review because the majority of searches were undertaken in electronic databases of peer-reviewed journals. There is increasing awareness that studies that include positive or interesting findings are more likely to be published, and indexed higher in these electronic databases. To mitigate this limitation, grey literature searches were also conducted using search engines. The review included publications that were health policy reports. The benefit of this is that any relevant barriers or enablers could be included in the findings from this work. However, as two of these studies did not include a study design it meant that quality appraisal could not be applied to these publications. This meant that no judgement could be made regarding the reliability of the extracted themes from those reports.

High reliability
The identified barrier is supported by several studies of medium quality and one high-quality study, or the study is supported by at least two high-quality studies.

Medium reliability
The identified factor is supported from several medium-quality studies, or the factor is identified from at least one high-quality study.

Low reliability
The identified factor is supported by several low-quality studies and/or one medium-quality study.
the adapted educational programmes for adults with learning disabilities and diabetes.

| Policy and practice implications
Although there is limited high-quality literature on this topic, this narrative review provides information on a number of key barriers and enablers to optimal diabetes care for adults with learning disabilities. From these themes, initial recommendations for commissioners and service providers can be made. They should consider introducing structured educational programmes that have been adapted for adults with learning disabilities and diabetes. They should create capacity in healthcare systems to allow personcentred reasonable adjustments for adults with learning disabilities and diabetes. Service providers should be encouraged to make reasonable adjustments for adults with learning disabilities. Accessible information about diabetes that is appropriate should be provided for adults with learning disabilities. Educational programmes need to be introduced for healthcare professionals and caregivers supporting those with learning disabilities and diabetes.

| CON CLUS ION
This narrative review sought to identify the barriers and enablers to optimal diabetes care that exist for adults with learning disabilities. Using a robust search strategy, ten publications were identified.
From this literature, 12 barriers and 14 enablers to optimal diabetes care for adults with learning disabilities were identified. The lack of high-quality literature means that no judgement on the generalisability of the identified barriers and enablers can be made. However, this work highlights key actions that can be undertaken to address inequalities in diabetes care for people with learning disabilities.
With additional research in this field, further progress can be made to improve the lives of those adults living with learning disabilities and diabetes.

CO N FLI C T O F I NTE R E S T
The authors report no potential conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.