Development and validation of the 4AT: a new rapid screening tool for delirium
Award Number
11/143/01Award Type
HTA CommissionedProgramme
Health Technology AssessmentStatus / Stage
CompletedDates
1 February 2014 -1 June 2017
Duration (calculated)
03 years 04 monthsFunder(s)
NIHRFunding Amount
£686,253.17Funder/Grant study page
NIHRContracted Centre
The University of EdinburghContracted Centre Webpage
Principal Investigator
Professor Alasdair MacLullichPI ORCID
0000-0003-3159-9370WHO Catergories
Development of clinical assessment of cognition and functionDisease Type
Dementia (Unspecified)CPEC Review Info
| Reference ID | 178 |
|---|---|
| Researcher | Reside Team |
| Published | 12/06/2023 |
Data
| Award Number | 11/143/01 |
|---|---|
| Status / Stage | Completed |
| Start Date | 20140201 |
| End Date | 20170601 |
| Duration (calculated) | 03 years 04 months |
| Funder/Grant study page | NIHR |
| Contracted Centre | The University of Edinburgh |
| Contracted Centre Webpage | |
| Funding Amount | £686,253.17 |
Abstract
Delirium affects at least 15% of acute hospital patients. It has multiple severe consequences, but is grossly under-detected. We have developed and implemented a new screening tool called the ‘4 A s Test’ or 4AT. It has four items: (1) alertness, (2) orientation (Abbreviated Mental Test – 4), (3) attention, and (4) an assessment of whether the onset of the confusion was acute. The initial development of the 4AT involved road-testing, feedback and modification before clinical implementation. The 4AT is now in clinical use in several UK hospitals. Early evaluations have shown promising performance. This preparatory work has provided a strong platform for the development of a more refined instrument. This is the focus of the present application. Phase 1 will further refine the 4AT, maintaining the option to make major changes if necessary. We will do (i) two surveys, one of doctors and nurses working in settings where delirium is common, and one targetted at staff already using the 4AT (eg. London, Edinburgh, Watford, Aberdeen); (ii) a qualitative study involving interviews with staff, distributed across sites, and observation of assessments at key points in the user journey (Emergency Department (ED), medical and geriatric admissions units and wards) to evaluate the existing instrument and any revisions; (iii) field testing of 4AT items and 4AT instructions, examining ease of use by staff, duration of completion, acceptability to patients and carers, and pathway from assessment to recording of results; (iv) Consensus Conference, which the data from the will be presented and discussed, and changes to the 4AT agreed. Phase 2 is a study of the performance of the 4AT as a screening instrument for delirium in the Emergency Department (ED) and general wards across three sites, compared against the reference standard of DSM-IV criteria. Secondary objectives are: comparison with the performance of short Confusion Assessment Method (CAM) and ability to detect general cognitive impairment. Following eligibility screening by clinical staff, consent from patient (or proxy/consultee agreement) will be sought by a study Research Nurse. Patients will be randomised centrally via a standard process at the Edinburgh Clinical Trials Unit to undergo either the 4AT (N=450) or the CAM (N=450) by the Research Nurse. A different Research Nurse will perform the reference standard and general cognitive impairment assessments in all patients. Economic evaluation: To estimate the delivery costs associated with the diagnostic tools, a decision-analytic model will be developed. An economic model will incorporate the results from Phase 2 on the diagnostic accuracy in addition to synthesising the evidence from the existing literature using a meta-analytic approach. The model conceptualisation and development will present the cost considerations regarding the implementation of the diagnostic tools from the perspective of the NHS. Setting: Emergency Departments in Sheffield and Edinburgh. Medical and geriatrics admissions units in Bradford and Edinburgh hospitals. Strategy for reviewing literature: We have conducted several literature reviews and consulted clinical guidance on delirium on existing tools for delirium screening, e.g. a recently completed review on attentional tests in delirium. We have also reviewed conference proceedings and attended annual meetings of the European Delirium Association and American Delirium Society. Ongoing searches will include quarterly searches of Pubmed, Embase and Web of Science. Target population: 900 patients aged >70 attending Emergency Departments via ambulance and/or patients newly admitted to medical admissions wards. Health technologies being assessed: the 4AT and the CAM. Measurement of cost and outcomes: We will evaluate the delivery costs of the 4AT and CAM from the perspective of the NHS. Healthcare resource use will be assessed using a patient or proxy-respondent self-report resource use questionnaire. Monetary values will be attached to resource use, training and labour costs as well as the indirect costs of delivering the screening tools. Generalised linear models will be used to analyse the resource consequences of a positive score by synthesising routine and within-trial data to guide the design of a decision analytic model. A microsimulation decision model will be developed to estimate delivery costs associated with the patient pathway as a function of the outcome of sensitivity and specificity of the triage tools and subsequent resource consequences at various timescales. Clinical and economic outcomes, such as mortality, length of stay and readmissions, will be collected up to 12 weeks from recruitment representing key model input parameters in addition to diagnostic accuracy. Sample size: Phase 1 N=180 (60 patients per site). Phase 1 will also include a total of approximately 500 professionals in the consultation process (surveys) and around 60 participants (professionals, patients and caregivers) in the Consensus Conference), 30 professionals in interviews, 10 patient/professional dyads in observation sessions and up to 20 in field testing of specific aspects of the 4AT. Phase 2 N=900: 300 patients will be recruited in each of the three sites, with half of patients from EDs and half from geriatrics and medical admissions units. Among the 450 patients randomised to assessment using 4AT, 67 (15%) would be expected to have delirium. The specificity of the 4AT triage tool could therefore be estimated precisely, its two-sided 95% confidence interval (CI) having a width of at most +/-0.05. For the sensitivity of 4AT, the two-sided 95% CI width would range from +/-0.072 (Sensitivity=0.9) to +/-0.12 at most (Sensitivity=0.5). CAM performance would be evaluable with the same level of precision. Project timetables including recruitment rate: Phase 1: Months 1-9 (3 months to set up, and recruit research nurses, 1 month to train nurses, 5 months to recruit 180 patients). Phase 2: Months 10-27 (18 months, recruiting an average of 1 patient per day per site). Phase 3: Months 28-30 (analysis). Expertise in team: This project requires a team with skills and experience in qualitative studies, statistics, health economics, test design, clinical practice, clinical research, ethics, and project management. The team expertise addresses each of these domains with extensive research experience in this field.
Aims
The main aim of the research is to develop a test that quickly and easily identifies patients with delirium, and which can be used by non-specialist staff. We will find out how long the test takes to complete, and how acceptable it is to staff, the patients and their carers.