MICA: Development of a software application for detection and monitoring of attentional deficits in delirium

Delirium is an acute, severe deterioration in mental functioning affecting at least 1 in 8 general acute hospital inpatients, and at least 1 in 3 intensive care unit inpatients. It is associated with considerable patient and carer distress and other poor outcomes such as longer stay in hospital. Despite its importance, delirium is grossly under-detected in acute hospitals, with less than 1 in 3 detected.

Several validated tools for the assessment of delirium are available. However, there is a lack of detailed research on means of objectively measuring inattention, the key cognitive deficit in delirium. Some commonly-used cognitive tests can reliably detect inattention, but mostly they do not clearly discriminate delirium from dementia. Also, there is a lack of knowledge on grading the severity of inattention, and monitoring attentional function over time in order to assess change in the patient. These important gaps contribute to the under-detection of delirium in clinical practice.

To address these gaps, we developed a new test for the measurement of inattention in delirium, implemented on a purpose-built computerised device called the Delbox. The test involves counting sequences of slowly-presented lights. We found that the Delbox detects delirium successfully, and also discriminates delirium from dementia. Our studies found that patients with dementia perform similarly on this test to cognitively healthy volunteers.

Although the Delbox is effective, tests on devices that are universally available have potential for far greater impact. Therefore, we have now developed a prototype software application (DelApp) for smartphones based on the same tests.

A feasibility study in 20 hospital patients showed highly comparable performance between the Delbox and the DelApp. Further studies of the DelApp in 156 general ward patients showed that patients with delirium had statistically significantly lower scores compared to patients with dementia or cognitively normal volunteers. A statistic called the Area Under the Receiver Operating Characteristic Curve (AUC) used to assess performance of diagnostic tests had a value of 0.99, which indicates excellent performance. We have also recently completed a feasibility study of the DelApp in 47 patients in the intensive care unit (20 with delirium). Delirium was associated with statistically significantly lower DelApp scores, and an AUC figure of 0.98, again indicating excellent performance. We also found that the test showed good sensitivity to change in attentional functioning upon repeat testing in a subset of patients.

We now propose a new series of studies which will first develop the DelApp further, to make it easier to use, and then to test its performance in a group study and then a larger group of unselected patients. In contrast to our pilot studies, we will use separate researchers to assess the presence of delirium and to administer the DelApp. This is important to meet agreed standards for the evaluation of tests that could be used in clinical practice. As part of this project we will work with an industrial partner, Cambridge Cognition. This company is an international leader in computer-based cognitive testing and already supplies tests for dementia for the NHS.

Objectives:

(1) To further develop the DelApp, including optimisation of the test, user interface and data acquisition and display.

(2) To conduct studies in patients in general wards (delirium, dementia, control) and the intensive care units (delirium, control) to assess feasibility and diagnostic accuracy.

(3) To conduct further studies in unselected patients in general wards and the intensive care unit (delirium, dementia, control). The aim is to determine the diagnostic test accuracy of the DelApp in representative patient groups and clinical settings.

(4) To produce a software application which is suitable for use in clinical practice and which is commercially available.

Delirium affects at least 1 in 8 acute hospital patients. It has multiple severe consequences, but is grossly under-detected.

A fundamental cause of the low rates of detection is the lack of well-validated tests for the core cognitive diagnostic feature of delirium: inattention. Currently available assessments of inattention have several drawbacks including inadequate discrimination between delirium and dementia, specific training required to administer, and lack of objectivity.

We previously developed a new neuropsychological test for the objective measurement of inattention in delirium. This was implemented on a purpose-built computerised device (Delbox). We have now developed a prototype software application (DelApp) for smartphones based on the same test.

The DelApp has many advantages over existing tests, including objectivity, automated scoring and ease-of-use. Studies in 30 hospital patients showed highly comparable performance between the Delbox and DelApp. Further proof-of-principle single-rater studies using the DelApp in more than 200 hospital patients have provided additional evidence of good performance of the DelApp.

The primary objectives of the programme of work are:

(1) To further develop the DelApp, including optimisation of test parameters, user interface and data acquisition and display.

(2) To conduct Phase 1/Phase 2 case-control studies in older patients (aged >64) in general wards (delirium, dementia, control groups) and in adult-age patients (aged >17) in intensive care units (delirium, control groups) to assess feasibility, reliability, diagnostic accuracy, and to inform optimal cut-points.

(3) To conduct a Phase 3 study in unselected patients in general wards and intensive care units. The main aim is to determine the diagnostic test accuracy and other test characteristics of the DelApp in detecting delirium in representative patient groups and clinical settings.

Who might benefit from this research?

Several groups might benefit from this research.

Patients with delirium: less than two-thirds of delirium is detected.

Patients at risk of delirium: for example high risk elective surgical patients (eg. those with dementia) could be screened pre-operatively, and DelApp scores used as a baseline value to compare with post-operative values, using reductions in DelApp scores as a delirium marker.

Patients with dementia: two-thirds of hospitalised older patients with delirium also have dementia, and only half of that dementia is formally diagnosed. So delirium is a strong marker for dementia. Related to this is the fact that good care of dementia in hospitals and other institutions requires delirium detection and management, because delirium is an extremely frequent and harmful complication of dementia.

Hospital managers: delirium increases length of stay by 2-3 fold, and new institutionalisation risk by 3-5 fold. Delirium is also a risk factor for falls, dehydration, and other adverse outcomes. Because delirium affects at least 15% of patients (ie. >75 patients in a typical 500-bedded hospital), suboptimal management resulting from non-detection has important consequences in any hospital.

Policy-makers: increasing delirium detection rates is a critical part of improving the quality of care for older people in hospital. Recent reports such as the National Dementia Audit highlight that in NHS hospitals delirium is poorly detected.

Researchers: The DelApp provides a new tool for objective attentional assessment in delirium.

Cambridge Cognition or other companies who license the DelApp.

The above benefits could readily extend internationally, because the tests are not language-specific, though they do depend on basic knowledge of arithmetic.


How might they benefit?

Patients with delirium will benefit because detection and good management is associated with better outcomes for patients, including less distress. In the ICU, delirium is associated with a higher risk of future long-term cognitive impairment (eg. Pandharipande PP, et al. Long-term cognitive impairment after critical illness. N Engl J Med. 2013;369:1306-16.). Detection of delirium linked with longer-term follow-up might allow patients with post-ICU cognitive impairment to have better support.

Delirium detection also provides a mechanism for improved dementia detection rates, in that older patients with delirium but with no prior diagnosis of dementia might be followed-up and assessed for dementia (after delirium recovery). Many hospitalised patients with delirium have established dementia, or are at high risk of future dementia (Davis DH, et al. Delirium is a strong risk factor for dementia in the oldest-old: a population-based cohort study. Brain. 2012;135:2809-16).

Hospital managers and policy-makers would benefit from better delirium detection in multiple ways, including providing better quality of care in older people in hospital (via higher rates of delirium detection and consequent better care), and potentially reducing length of hospital stay and new institutionaliation.

Researchers might benefit through having better tools for assessment of inattention, with several benefits including increased comparability among studies and improved inter-rater reliability.

Companies such as Cambridge Cognition might benefit because the potential markets in the UK and internationally are large. Even if the initial use is focused on older patients in geriatrics wards, and ICUs, the numbers of patients (and associated healthcare) staff are considerable such that the DelApp is highly likely to be commercially viable.

More generally, given the enormous human and economic costs of delirium, methods that improve detection and thus care have the potential for considerable leverage, well beyond the costs of development and purchasing by end users or organisations.