Screening for delirium and cognitive impairment in older, acute care in-patients

Hendry, Kirsty (2017) Screening for delirium and cognitive impairment in older, acute care in-patients. PhD thesis, University of Glasgow.

Abstract

Background:
Delirium, an acute neurobehavioral syndrome, occurs across all healthcare settings and is suggested to be the most common psychiatric condition experienced by older hospitalised patients. It affects around a fifth of those in general medical wards with higher prevalence in surgical and intensive care unit patients.
Delirium and chronic cognitive impairment share a complicated two-way relationship. Those with dementia are at greater risk of developing delirium while length of delirium episode is also associated with increased risk of long-term cognitive decline. Delirium is associated with a number of other serious negative outcomes including increased risk of falls, institutionalisation and mortality. Identification of delirium in hospitalised older patients is necessary to facilitate good patient care as well as to allow for the appropriate support for concerned relatives and carers.
Guidelines are in general agreement that screening for delirium and cognitive impairment is important in hospitalised, older patients. Identification of delirium is the first necessary step to then allow for the management of this syndrome. However, there are a wide range of screening tools available for cognitive impairment and delirium with limited research evidence or validation of these tools in large, representative cohorts. Furthermore, clinical awareness of delirium is low compared to many other conditions; this may be improved by implementing clear delirium screening guidelines along-side the necessary training.
Methodology:
Before delirium screening tools can be implemented in routine practice, an evidence-based approach should be followed to assess feasibility and diagnostic accuracy within older in-patient cohorts.
In this thesis, I investigate screening for cognitive impairment systematically in a series of linked studies. I review the existing published evidence as well as investigating screening for delirium in older, acute medical units locally and nationwide. I collate existing evidence for the use of brief screening tools for delirium, dementia and mild cognitive impairment (MCI) across healthcare settings. I also carry out analysis of an existing data set looking at the feasibility and accuracy of two single questions for delirium and dementia, separately. Furthermore I gather data relating to cognitive screening from lead clinicians across hospital sites within elderly acute care units in Scotland. I also carry out a local service evaluation to determine documented delirium prevalence as well as what tools were being used to screen for cognitive impairment and delirium. These results inform a diagnostic test accuracy evaluation of delirium and cognitive impairment screening tools recommended for routine clinical use with acute care in-patients. This evaluation is in a relatively large-scale, representative sample and assesses the feasibility as well as accuracy of these tools against a gold standard clinician diagnosis of delirium.
My diagnostic test accuracy evaluation was based on a clear local problem of lack of routine delirium screening in older in-patients and aimed to inform future recommendation policy by examining which tools are feasible and accurate within this setting. I also aimed to add to the existing delirium screening evidence base by examining a range of recommended tools within a large, consecutive patient cohort. This was contrary to much of the published literature which generally examine one screening tool and often within small or case-controlled patient samples.
This evaluation of screening tools for the assessment of possible delirium within the acute care setting examined the feasibility and test accuracy of cognitive tests which were recommended by clinical guidelines for both delirium and cognitive impairment. The tests evaluated were the Abbreviated Mental Test (AMT 10/4), the 4 A’s Test (4AT), the brief Confusion Assessment Method (bCAM) (a rapid, operationalised version of the Confusion Assessment Method (CAM)) and the Single Screening Question in Delirium (SQiD). I also explored the performance of reciting months of the year backwards (MOTYB), present as part of both the 4AT and bCAM. All screening tests were compared to gold standard diagnosis using delirium criteria from the Diagnostic and Statistical Manual of Mental Disorders – fifth revision (DSM 5) which was completed by senior geriatricians.
Findings:
My systematic literature review revealed heterogeneity of methods in the published evidence for very brief, single item cognitive screening tools. However my secondary data analysis revealed high sensitivity for a single informant question for dementia and reasonable sensitivity for a single question for delirium.
The clinician survey showed a lack of consensus regarding the choice of screening tools used for delirium and dementia at a national (Scottish) level. Within geriatric units in Scotland there appears to be notable variability in the way delirium screening is carried out. The clinician survey revealed a particular issue for delirium screening in the West of Scotland where there appears to be a lack of standardised tools used to screen for delirium. Furthermore, local ward service evaluation revealed a lack of documentation of delirium diagnosis with little awareness of delirium across acute elderly wards within a large teaching hospital in Glasgow.
Evaluation of cognitive impairment screening tools found that the AMT 10, AMT 4, 4AT and MOTYB were feasible and accurate tools for the assessment of delirium within a cohort of 500 acute in-patients age > 65 years. The AMT 10 was found to have reasonable sensitivity at a cut point of <4/10 and the AMT 4 was found to have good sensitivity at a cut point of <3/4; use of the full 10-point AMT seemed to carry no substantial advantage over the shorter AMT 4. The bCAM was found to have poor sensitivity, missing 3 in 10 cases of delirium. I did not find the informant-based SQiD to be feasible in this population, with a return rate of 28%, but displaying a sensitivity of over 90%. These results suggest that a range of tools exist which display good diagnostic test accuracy and feasibility in an older, acute care in-patient cohort. These can all be completed quickly and are simple to administer. Informant information using a standardised single screening question (SSQ) such as the SQiD may still hold value in aiding the diagnosis of delirium when this can be obtained.
Conclusions:
In conclusion, the studies in this thesis aim to add to the pool of literature available for the screening of delirium and cognitive impairment. I used a logical and informed ordering of the research conducted. The results from my systematic review, secondary data analysis, clinician survey and service evaluation all fed in to the planning of my clinical patient evaluation of delirium screening. Results from my literature review and data analysis did not discount the use of a single question to screen for delirium but did suggest a need for further research with a gold standard clinician diagnosis for comparison. My clinical evaluation results revealed that relatively accurate screening of delirium is possible using existing, simple and brief screening tools which are already suggested in guidelines for routine clinical use.
Screening for delirium should be regarded as a first step in the care pathway for those who are identified as having possible delirium. The value of delirium screening depends on the implementation of specific care pathways for those who then go on to receive a clinical diagnosis of delirium. Patients with delirium have an increased risk of falls, dehydration and infection alongside the associated long-term complications. Good patient care should aim to cater to these patients’ specific needs in the same way it does with other medical conditions.
Healthcare Improvement Scotland (HIS) recommends all older patients should be routinely screened for delirium but acknowledges that this is not the case, with delirium being ‘frequently overlooked or misdiagnosed’. It may not be enough to make these recommendations without implementing a system of education to promote and raise awareness for the importance of screening for delirium.
I suggest that further research is needed to assess the accuracy and feasibility of delirium screening tools for older, acute care in-patients while implementing a care pathway for patients who are then diagnosed with delirium. This would inform the best possible future care for patients with delirium. The potential for improved outcomes for these patients is also of interest. Evaluation of interventions in large scale, representative patient samples are needed to further progress our knowledge of treatment of delirium as a serious and often overlooked disorder of the brain caused by physical illness.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: Delirium, dementia, cognitive impairment, geriatric medicine, older adults, diagnostic test accuracy.
• Subjects: B Philosophy. Psychology. Religion > BF Psychology
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Cardiovascular & Metabolic Health
• Supervisor's Name: Stott, Prof. David J. and Evans, Prof. Jonathan
• Date of Award: 2017
• Depositing User: Miss Kirsty Hendry
• Unique ID: glathesis:2017-7925
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 28 Feb 2017 15:23
• Last Modified: 12 Apr 2017 08:09
• URI: https://theses.gla.ac.uk/id/eprint/7925