Background
Dementia is characterized by progressive impairment in domains such as memory, orientation, comprehension, language, and judgement and can also affect personality, mood and behaviour. In Ireland, approximately 41,700 people were estimated to have dementia as of 2006, and this figure is projected to rise to between 140,580 and 147,000 by 2041 [
1], making it a significant public health issue. According to the World Health Organisation [
2], the rapidly increasing prevalence of dementia will place unprecedented demands on health care services, challenging governments to develop and improve the services providing care for people with dementia, in order to meet these demands.
Older people with dementia on average have three or more physical comorbidities, which at some point necessitate hospital admission [
3] and the prevalence of dementia in hospital ranges from 29-42 % in adults over 70 [
4,
5]. Hospital admission is distressing for this population and is associated with cognitive and functional decline [
6], greater institutionalisation post-discharge [
7,
8], and higher mortality rates [
5,
7].
Acute hospitals are not currently equipped to provide best dementia care [
9]. Patients are generally not hospitalized for the dementia itself [
10] and therefore the dementia is rarely the care or treatment priority. Further, staff training and knowledge around dementia care can be poor [
9,
11] and it often goes undetected in the acute setting [
12]. This can result in unmet needs, and an increase in behavioural and non-cognitive symptoms of dementia (BPSD) [
13], which staff report to be burdensome [
14]. Other factors contributing to poor quality dementia care in hospitals include poor multidisciplinary assessment, [
15,
16], under-diagnosis and treatment of pain [
17], failure to collect collateral history [
18], and the inappropriate prescription of antipsychotic drugs [
19].
A diagnosis of dementia is also related to greater length of stay (LOS) in hospitals [
20], and an increased cost of care. Dementia care is estimated to cost €21 million per annum in Ireland, based on the number of Hospital In-Patient Enquiry (HIPE)-coded discharges of people with a recorded formal diagnosis of dementia [
1]. A more recent study has estimated that dementia-related admissions lead to approximately 246,908 additional days per annum in hospitals in Ireland, at an associated cost of €199 million [
21]. Reasons cited for increased LOS are largely systemic, include poor discharge planning, delayed discharge relating to post-hospital care availability for complex conditions, and lack of communication and coordination between health service providers [
22,
23].
It is clear from the above literature that improving dementia care in acute hospitals should be an urgent priority in Ireland. Countries including England and Wales [
15,
16] have carried out audits of dementia care to gain a baseline picture of the quality of care provision in their hospitals. The objective of this audit is to assess, for the first time, the quality of dementia care in acute public hospitals in the Republic of Ireland.
Method
As there are no standards in place for dementia care in acute hospitals in Ireland currently, this audit measured care against international best practice guidelines.
All 35 acute public hospitals admitting adults within the Health Service Executive (HSE) were included in this audit, and the data was collected between April and September 2013. The audit tools were developed for use in the National Audit of Dementia Care in General Hospitals in England and Wales [
15], and with permission from the Healthcare Quality Improvement Partnership, minor alterations were made to the wording of questions, where necessary, so that they made sense in the context of the Irish healthcare system.
For this audit, (i) a retrospective chart review tool, as well as (ii) hospital level organisation (iii) ward level organisation (iv) and ward environment tools were employed. The results of the ward environment module of the audit are not reported here.
(i)
Chart Review: This tool tracked the quality of care received by patients (n = 660) retrospectively, from admission through to death/discharge during a single admission, collecting information on demographics, assessments carried out on or during admission, referral to specialist services and discharge planning. Charts were identified using the HIPE system which collects national demographic, clinical and administrative data on discharges from, and deaths in all acute public hospitals. Inclusion criteria included a recorded HIPE diagnosis (primary or other) of dementia (International Classification of Disease (ICD)-10 codes: F00, F01, F02, F03, F05.1), a LOS greater than 5 days, and a hospital discharge date between 1st September 2012 and 31st January 2013.
The data was collected by the audit team as well as Specialist Registrar trainees in geriatric medicine, all of whom received comprehensive training in the use of the audit tool. Where possible, hospitals had their charts reviewed by independent auditors (76 %). Furthermore, 25 % of charts were independently re-reviewed by the audit team. The tool was found to have good inter-rater reliability and percentage agreement, similar to that found in the England & Wales audit [
15].
(ii)
Hospital Organisation: The audit team used this tool to collect data on hospitals’ (n = 35) policies, procedures, structures and guidelines relating to dementia care, through structured interview with senior hospital management and clinicians.
(iii)
Ward Organisation: This tool collected data on staffing levels and supports, availability of specialist services, and systems in place to support people with dementia on wards through structured interview with the ward manager. In each hospital, 2-3 wards that admit adults were selected, with at least one medical and one surgical ward chosen per hospital. In total, 77 wards were audited by the audit team.
Data analysis
Most of the data collected was categorical in nature, and valid percentages are reported in this study to describe this data. Median and interquartile ranges were used to describe the distribution of continuous variables, including age and LOS as the data did not have a normal distribution. For this same reason, the Mann-Whitney U-Test (non-parametric equivalent of the T-test) was used to investigate group differences in continuous variables. Data were analysed using SPSS version 20.
Ethics
As this was an approved national audit, performed in partnership with the Health Service Executive’s National Quality and Patient Safety Directorate, we were exempt from seeking multiple hospital-level ethical committee approvals. Data collectors external to a hospital signed confidentiality agreements and access to charts at each site was coordinated through the hospital’s audit/quality department. Each hospital was assigned a code, kept securely by the audit coordinator, and individual hospital-level data was only released to that hospital’s Executive Officers. Similarly, each patient’s chart was assigned a code at source (to allow the re-audit of some charts for quality assurance) and no identifying patient details were recorded. The linking code was kept securely in the hospital’s audit/quality department and destroyed at the end of the audit.
Discussion
The objective of this audit was to investigate, for the first time, the quality of dementia care across all 35 acute public hospitals in the Republic of Ireland.
Generally, high levels of physical assessments were carried out, echoing findings relating to hospital dementia care in England and Wales [
15,
16,
24] particularly for mobility, continence needs, and pressure sore risk; however a standardised assessment of functioning was carried out on only 36 % of patients with dementia. Assessing for changes in functional ability is important for prognostication, minimising hospital associated disability, and predicting mortality and other health outcomes during and after hospitalization [
25], as well as for determining dependency to secure an adequate care package after discharge. All older patients, particularly those with dementia, should receive a standardized assessment of their functional ability during hospital admission.
Assessment for the presence of pain was poor, with almost one-quarter of patients not being asked about the presence of pain, or receiving any formal assessment. This reflects previous findings that pain is under-assessed in people with dementia [
15,
24,
26]. Pain is common in older people with dementia, with up to 45 % experiencing pain [
27]. While patients with dementia may not always be capable of communicating pain through self-report, pain can be assessed with tools through caregiver report and direct observation [
28]. There are a number of validated tools that assess for indicators of pain, including the Pain Assessment in Advanced Dementia Scale [
29]; such tools should be used in cases where verbal communication is not possible.
Mental status assessment was poor, which other research has found in this population [
24] in a sample of 7934 people. Despite the fact that this sample had a known diagnosis of dementia, just over 40 % had any assessment of cognition, reflecting other findings in Ireland (55 %) [
30], and in England and Wales (43 %) [
15]. Cognitive assessment is vital in people with dementia to track course and prognosis, and determine the supports needed to manage the condition [
31]. This population are also at high risk of developing delirium [
32], however less than 30 % of patients were screened during their admission. Poor hospital delirium screening in dementia has been reported in other countries [
16,
33‐
35]. Delirium superimposed on dementia is related to adverse outcomes including increased functional dependence, institutionalisation and mortality [
36]. According to NICE guidelines [
37], screening for fluctuations in behaviour should be carried out at least daily in the acute setting. The absence of this in over 70 % of charts reflects particularly poor practice in the care of a population at high risk of delirium onset.
Mood and BPSD were only assessed in 14 % of patients. Under-recognition and assessment of BPSD is common [
38]. BPSD, including depression, are experienced by 66-97 % of people with dementia during the course of their condition [
39]. Undetected and untreated symptoms can lead to significant caregiver burden [
40] and consequently, earlier institutionalisation. Finally, less than half of patients had any collateral history recorded in their charts. This is imperative for the comprehensive assessment of change from baseline across a number of domains [
41] and is central to treatment and prognostication.
Regarding hospital size, operationalised by bed capacity, the largest hospitals were more likely to carry out delirium screening and BPSD and mood assessments, and record collateral history, than the smallest hospitals. However, the largest hospitals were poorer at carrying out assessments of functioning and nutrition; the smallest hospitals these assessments more frequently. Cognitive assessments were completed most frequently in smaller hospitals. While these findings indicate important trends in dementia care by hospital size, bed capacity provides limited information for service planners about how to improve the quality of dementia care, as this approach does not account for organisational factors including the individual services provided within hospitals or the flexibility of staffing based on fluctuating demand [
42,
43]. This is highlighted here by the large variation in assessments completed between the hospitals within each hospital size group, where resources would be expected to be reasonably similar.
There was a disparity between hospital policy regarding assessment for cognition, functional ability, nutritional status and BMI, and actual assessment in practice. In order to improve multidisciplinary assessment in hospital dementia care, better governance must happen at the hospital level; policies and guidelines must be put in place and enforced in practice at ward level. It is important that this issue of under-assessment, particularly mental assessments, for people with dementia is addressed in Irish hospitals; poor quality acute care has been suggested as contributing to an increased risk of adverse outcomes for older adults including those with dementia [
44] Furthermore, those with dementia have an increased cost of care when admitted to hospital, due to greater lengths of stay compared to those without dementia [
45]; improvements in multidisciplinary assessments to detect physical and mental problems would likely reduce the length of stay and consequently, cost of care.
According to ward managers, there was very poor access to psychology, continence services and social work, with 91, 66, and 47 % of wards reporting no access to these services respectively. People with dementia can have complex physical and psychosocial needs, particularly during a distressing hospital admission, and these services are vital to providing best care. Of note, there was some disagreement between wards in the same hospital about access to the above services, 9 % of wards disagreed about the availability of psychology and social work within the same hospital, and 17 % disagreed about continence services. This lack of clarity about the availability of services is reflective of fragmentation and a lack of coordination between different levels and settings of care [
46]. Better communication is needed between hospital level and ward level management about which services are available within their hospitals. Much better access to pharmacy, physiotherapy, dietetics, speech and language therapy, occupational therapy, liaison psychiatry, and palliative care was reported. Future research should investigate the impact of specialist service availability in acute hospitals on the care and health outcomes of people with dementia, as well as the issue of fragmentation in relation to specialist services.
The provision of dementia awareness training is poor across hospitals; only two hospitals included it in their induction programmes. None of the 35 hospitals had mandatory training for any staff in dementia awareness, indicating that it is not currently considered a priority. Just over half of hospitals had dementia awareness training available to their nurses and doctors over the previous 12 months. However, almost two-thirds of ward managers reported that staff cover could not be provided for training relating to dementia care, indicating that staffing levels and lack of resources may be a barrier to staff attending any available training. This picture of inadequate dementia training provision for healthcare staff in acute settings has been found elsewhere also [
15,
47]. Dementia awareness training should be mandatory for all staff that care for people with dementia in Ireland as it is integral to increasing overall quality of care and reducing staff burden.
Almost 70 % of wards had at least one vacancy between their permanent healthcare and nursing staff, and the mean number of nursing vacancies was 1.83, with a maximum of nine nursing vacancies. Lower staffing levels in hospital have been associated with reports of poorer quality care and nurse burnout, as well as adverse patient outcomes including higher mortality [
48]. Low staffing levels arise from systemic problems, however in an effort to improve services for dementia, more resources will need to be allocated to ensure sufficient nurse-patient staffing ratios in hospitals.
There are very few supports in place for staff working with people with dementia, with mentorship and clinical supervision not available on 78 and 56 % of wards respectively. Further, peer support groups or reflective practice groups were not available to staff on 88 and 95 % of wards, and 84 % had no dementia champion to provide guidance on issues relating to dementia care. Dementia champion training is now available in Ireland, and so the number of hospitals with dementia champions may increase significantly in the near future. Perceived support from co-workers is important, and has been shown to enhance job performance and decrease the job stress among nursing staff [
49]. It is important that management take responsibility for putting supports in place for staff caring for people with dementia to prevent burnout.
Discharge planning was quite poor; the majority of planning (72 %) was not initiated within 24 h of admission, and only 39 % of patients had a single detailed plan for discharge in their charts. The literature indicates that good discharge planning is imperative; inefficient planning is associated with adverse patient outcomes, increased length of stay, re-admission, and cost [
22,
50].
Discharge destination was linked to LOS in this sample; the average LOS was 24.73 days, however further investigation revealed that those newly discharged to residential care had a significantly greater LOS than those not requiring such placement. The discharge of 35 % of people with dementia originally admitted from home to long term care in this sample is high compared to 11.5 % of older people without dementia in a recent study also carried out in 5 acute public hospitals [
7]. This finding indicates that acute hospitals may be serving as an access point for entry to long term care for people with dementia struggling in the community. Overall, these findings suggest that the system for planning discharge teamed with poor availability of suitable residential placement for people with dementia is problematic in acute hospitals in Ireland, and is contributing to an increased LOS and consequently, greater cost of dementia care.
The results also indicated poor summarising of information, for example support needs, level/cause of cognitive impairment, and delirium and BPSD at discharge. This is poor practice, as the gap between hospital and community care is then not bridged and health outcomes cannot be effectively maintained or improved [
22]. A more integrated approach to care across settings is required in order to reduce the unnecessary use of hospital beds once the patient is medically fit for discharge; this includes improved discharge planning, e.g. building relationships with post-acute care providers, and developing shared information systems to facilitate speedy transition [
51].
Carer’s needs were found to be poorly assessed prior to discharge also, and they were found to have short notice of discharge. This is in line with reports that post-discharge, families feel ill-equipped to take on the carer role, and feel that staff neglected the component of discharge planning relating to accessing assistance and resources in the community [
52]. Assessing carer support needs prior to discharge is important as caregiver burden is a risk for early institutionalisation [
53].
This study has some limitations; the chart review data has been collected retrospectively, and these findings reflect what has been recorded in the charts. Theoretically, care relating to multidisciplinary assessment and discharge planning may have been better in practice; however any failure to record such information is a failure in communicating essential healthcare information to colleagues, which is, in and of itself, poor practice.