Statistics from Altmetric.com
Which bedside instruments are accurate for detecting delirium in hospitalised adults?
Sensitivity, specificity, positive likelihood ratio (LR+) and negative likelihood ratio (LR−).
MEDLINE (from 1950 to May 2010) and EMBASE (from 1980 to 2010). Additional articles were identified by searching the bibliographies of retrieved articles.
Study selection and analysis
Inclusion criteria were published English language studies conducted in hospitalised patients (not intensive care unit) and that included participants with and without delirium. Studies were prospectively designed and used an appropriate reference standard (DSM-III, DSM-III-R or DSM-IV) that was performed by a specialist physician (geriatrician neurologist or psychiatrist) and applied the same index test to >80% of patients. The review did not include studies of patients with mostly alcohol-related delirium or a paediatric population or studies where the index and reference tests were performed by the same individual.
Nine studies were included assessing 11 bedside instruments. The prevalence of delirium ranged from 9% to 63%, the highest in a study of patients hospitalised in an oncology or palliative care ward presenting with mental status change. The tools which had likelihood ratios greater than 5.0 for positive tests were the Global Attentiveness Rating (GAR), Memorial Delirium Assessment Scale (MDAS), Confusion Assessment Method (CAM), Delirium Rating Scale Revised-98 (DRS-R-98), Clinical Assessment of Confusion (CAC) and the Delerium Observation Screening Scale (DOSS). Negative likelihood ratios of less than 0.2 were found with the GAR, MDAS, CAM, DRS-R-98, Delirium Rating Scale (DRS), DOSS, Nursing Delirium Screening Scale (NuDESC) and the Mini Mental State Examination (MMSE). The CAM had the best diagnostic OR out of time efficient instruments performed by nurses (summary LR+: 7.3, 95% CI 1.9 to 27; summary LR−: 0.08, 95% CI 0.03 to 0.21). For physicians the summary LR+ with CAM was 19, 95% CI 0.13 to 0.27). The MMSE was the least useful for identifying a patient with delirium (LR+: 1.6, 95% CI 1.2 to 2.0; LR−: 0.12, 95% CI 0.04 to 0.38).
The GAR, MDAS, CAM, DRS-R-98, CAC and DOSS are appropriate bedside tests for delirium. They give an adequate likelihood ratio for positive results. The DRD, NuDESCand MMSE give an adequate LR for negative results. The MMSE is the least useful tool. The choice of tool may be dependent on the time available and the discipline of the examiner. The evidence is greatest for the 5-min CAM which can be used out by both nurses and clinicians.
Delirium is a very serious complication of the plethora of medical disorders that influence brain function. Like all syndromes, a variety of presentations are possible, and although delirium usually presents acutely and resolves rapidly, it can also have an episodic, chronic or progressive course. As a result, the diagnosis of delirium can be challenging and inter-rater reliability is typically poor. Wong and colleagues hope to improve this situation by comprehensively reviewing 11 candidate instruments that may help with delirium assessment. The main strength of this review is that the authors provide a diagnostic validity meta-analysis, although this was only possible for five instruments (a further six were examined individually) and only two (the Confusion Assessment Method and the DRS) had more than three attempts at replication. The second strength of this review compared with previous attempts is that the authors attempt to restrict their analysis to the most robust studies employing DSM-III-R and DSM-IV as a reference standard. That said, it is far from certain that DSM offers any superiority over ICD or indeed expert clinician consensus.1 This issue can only be fully resolved by developing a better gold standard based on a common pathophysiology of the condition.
The review itself is well conducted and informative although the question of acceptability and implementation of the tools in clinical practice is underdeveloped. For example, one scale that seems to be underrepresented is the ubiquitous MMSE. There are several studies comparing MMSE with DSM criteria that have not been included and this may be important as the MMSE is still a popular choice among clinicians.2 The MMSE has many shortcomings, but improving upon it requires either greater accuracy or greater acceptability. Brief, efficient single domain tests like the Digit Span and Clock Drawing Test are an increasingly popular initial screening method for dementia,3 and their value in delirium is unclear and clarification urgently needed. The overall paucity of short tools with a good evidence base may explain why most clinicians do not routinely test for delirium and why delirium continues to be seriously overlooked in clinical settings.
Source of funding Tier 2 Canada and the Alberta Heritage Foundation for Medical Research.
Competing interests None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.