Delirium is a severe neuropsychiatric syndrome characteristic by acute onset and fluctuating course of inattention, altered level of arousal and other mental status disturbances such as delusions and hallucinations. Many synonyms for delirium exist including ‘acute confusional state’, ‘metabolic encephalopathy’ and ‘organic brain syndrome’. Up to one third of elderly hospitalised patients will suffer from delirium and up to two thirds of such patients will also have pre-existing dementia.1 Irrespective of aetiology, delirium is independently associated with poorer outcomes including increased length of stay, hospital acquired complications, dementia risk and overall mortality.2 Delirium also causes considerable patient and carer distress.3 To mitigate these sequelae it is essential to identify delirium so as to address its many modifiable causes, to treat distress and to create an environment conducive to recovery.

Unfortunately studies show that delirium remains widely under-detected. In a recent UK study Collins and colleagues showed that 72% of cases of delirium were missed by medical teams within the first 72 hours of admission to hospital.4 They concluded that the Abbreviated Mental Test (AMT) alone was unhelpful in assisting with diagnosis. Elie and colleagues reviewed emergency department medical notes and found that the majority of delirium cases were missed, even after efforts to actively encourage diagnosis were made.5

With routine consultation alone, 74% cases of delirium were missed and with a seven-point mental status checklist, which asked directly for delirium, 65% of cases were still missed. The authors speculated on reasons for this, including lack of training, time and perceived interventions. 

Inouye and colleagues, in a study involving 800 patients, found that routine observations carried out by nurses not trained in delirium detection was a poor method of detecting delirium, with sensitivities of 15–31%.6 They also found that certain risk factors made detection less likely: hypoactive delirium; age greater than 80 years; visual impairment and dementia.

Limitations in this study included inconsistencies in areas such as nursing expertise and handover methods. Han and colleagues examined emergency delirium detection tools in routine geriatrics care department and general medical doctors’ case notes and showed that both groups routinely missed delirium: 76% and 72% respectively.7 Strikingly, not one of the more than 300 patients involved had a documented delirium assessment. The authors concluded that there are significant failings in frontline assessment. A drawback of this study was that the reference standard assessment used, the CAM-ICU (Confusion Assessment Method for the Intensive Care Unit), is not validated in the emergency department and has been shown to have a low sensitivity in general populations.8 This means that the under-detection rates could be even higher than found in this study.

These studies suggest widespread delirium underdetection, starting at the beginning of the patient journey in the emergency department but also evident further downstream. One potential method of improving this situation is for healthcare systems to use tools that are suitable for routine use. Here we provide a selective review of delirium detection tools, focusing on those potentially suitable for routine use in geriatrics wards.

Search strategy
Our aim was to focus on studies examining delirium assessment in general and geriatric inpatients. Pubmed, Medline, AARP Ageline, Embase, Psychinfo and the British Nursing Index were searched from between January 2000 to January 2011. The search terms were ‘screening tool or diagnostic tool’ and ‘delirium or confusion’ and ‘risk-factor’ and ‘older people over 65 years’. We also drew from several published reviews of delirium assessment instruments. References thought to be important to this review but not captured in the initial search were also retrieved and included. Final selection of the tools was guided by potential utility in routine geriatrics inpatient care. 

Tools for delirium detection in routine clinical practice
Our search revealed a wide spectrum of tools used for delirium and cognitive screening in older hospital inpatients. These include: the Clinical Assessment of Confusion (CAC); the Clock Drawing Test (CDT); the Confusion Assessment Method (CAM); the Confusion Rating Scale (CRS); the Delirium Index (DI); the Delirium Observation Screening (DOS) Scale; the Delirium Rating Scale (DRS); the Digit Span Test (DST); the Delirium Symptom Interview (DSI); the Global Attentiveness Rating (GAR); the Memorial Delirium Assessment Scale (MDAS); the NEECHAM Confusion Scale; the Nursing Delirium Screening Scale (Nu-DESC); the Mini-Mental State Examination (MMSE) and handwriting tests. 

Some of these tests are not validated for delirium detection and identify only cognitive impairment without differentiating among delirium, dementia or another diagnosis. The following summary of clinically useable scales focuses on selected tools, which could be used by geriatricians relatively rapidly and without specialist training, to facilitate delirium diagnosis.

The Diagnostic and Statistical Manual of Mental Disorders and International Statistical Classification of Diseases
The Diagnostic and Statistical Manual of Mental Disorders (DSM) published by the American Psychiatric Association provides standard criteria for the classification of mental disorders such as delirium. Several editions have been published, the most recent, DSM-5 in 2013.9 A second standard, particularly used in Europe, the International Classification of Disease tenth revision (ICD-10), is a system of coding created by the World Health Organization. Both are designed for use by psychiatrists and are regarded as the gold standard for diagnosis of delirium. Some expert clinicians report directly using DSM criteria as a clinical rating scale without using other tools.10 However DSM and ICD are intended for use by specialists so cannot be recommended for routine clinical use by general staff. 

The Confusion Assessment Method
The most commonly advocated tool, the CAM diagnostic algorithm, is reported to have a sensitivity and specificity of approximately 90% when used by staff specifically trained in its use.11 Here a diagnosis of delirium is reached if 1 and 2 and either 3a or 3b are positive:
1. Acute onset and fluctuating course
2. Inattention
3a. Disorganised thinking
3b. Altered level of consciousness.

The CAM requires a preceding separate cognitive assessment (eg. MMSE) and so it typically takes around 10 minutes to complete. Specific training is required in its use.
Laurila and colleagues found a sensitivity of 81–86% and specificity of 63–84% and concluded that the CAM was an acceptable screening tool but concluded that delirium diagnosis should be ensured according to the DSM-IV.12 Furthermore they proposed that the use of the CAM should be restricted to ‘experienced physicians’. Contrastingly, Monett and colleagues trained nurses and research assistants over a five day period on using the CAM and found 86% sensitivity and 100% specificity.13 They suggested that systematic screening by trained nurses using the CAM was feasible and efficient in early detection. In conclusion, the CAM performs well as a screening tool for delirium when used by staff trained in its use and with adequate time to carry it out.

The Delirium Observation Screening Scale
The DOS Scale is based on DSM-IV criteria of delirium.14 It originally consisted of a 25 item scale but this scale was subsequently reduced to 13 items. This tool is commonly used in Dutch hospitals and is designed to be used by nurses during routine patient care to pick up early symptoms of delirium: over a 24 hour period nurses on each shift calculate a score and then an average rating is taken. Observations which are incorporated in the scale include patient easily distracted by their environment, patient slow to answer questions and patient picking at intravenous lines, catheters etc. It has good concurrent validity with the CAM and can indicate severity of delirium as well as acting as a screening tool.15

The Nursing Delirium Screening Scale
The Nu-DESC, an extension of the Confusion Rating Scale, has five items including orientation and psychomotor retardation over a 24 hour period and like the DOS is designed for administration by a nurse based on routine observations. It is reported to have good inter-rater reliability and it takes less than two minutes to complete.16 Leung and colleagues compared general medical nurses trained on the use of the Nu-DESC and the CAM and found that the Nu-DESC was more sensitive and rapid than the CAM, although the latter was more specific. The Nu-DESC had moderate agreement with the CAM (kappa 0.52). A lack of reliable baseline information may have impacted on results, namely the high false negative rate with the Nu-DESC. The authors proposed a three-stage ‘filter approach’ towards optimal detection of delirium: firstly nurses apply the Nu-DESC to high-risk patients; secondly physicians utilise the CAM algorithm to confirm those screened positive and thirdly psychiatrists are involved when diagnostic uncertainty exists. Clearly such an approach requires sufficient staffing and time to implement appropriately. 

The NEECHAM Confusion Scale
The NEECHAM Confusion Scale was developed to assess acute confusion on the basis of criteria identified by nurses assessing three levels of function: mental, behavioural and physiological. It requires observation of cognitive processing as well as physiological measures such as oxygen saturations. It has been shown to have good inter-rater reliability (0.91–0.96) and is fairly easily applied.17 However, its original purpose was to assess cognitive dysfunction rather than specifically delirium and may not been interchangeable.

The 4 ‘A’s Test (4AT)
The 4 ‘A’s Test or 4AT has four items which address core diagnostic features of delirium: level of arousal, cognitive impairment (orientation and attention) and acute deterioration in mental status.18 It generally takes under two minutes to complete and it can be applied by staff given minimal training. It incorporates two short items to provide basic cognitive screening. Initial studies suggest this tool to be sensitive and specific and possibly applicable in routine geriatrics care and in stroke medicine.19,20 These results are encouraging though further validation studies are required. 

Conclusions
There is a wide range of tests available for the assessment of delirium. However, relatively few have been evaluated in relation to use in routine clinical practice. We highlighted six tools in this review: The Diagnostic and Statistical Manual of Mental Disorders; The Confusion Assessment Method; The Delirium Index; The Delirium Observation Screening Scale; The Nursing Delirium Screening Scale; The NEECHAM Confusion Scale and the 4 ‘A’s Test. Each tool differs in its features, including the need for training, the target staff group, the aim of the tool (screening versus diagnosis) and the time taken to complete the tool. Additionally the size of the evidence base for each tool varies widely, with the CAM having many studies examining its performance, but other tools with more limited evidence.

The choice of tool depends on the resources available and the target staff group. The CAM is well-supported by the literature assuming that staff undergo specific training in its use and that adequate time (10 minutes per assessment) is provided for its use. The Delirium Observation Screening Scale is widely implemented in the Netherlands and elsewhere as a delirium detection tool for use by nurses in routine care. Some training is required but once implemented the DOS Scale is very rapid to administer. The 4AT is a new tool and has a small evidence base, but shows promise as a method for rapid detection of delirium in routine care without the need for staff training. Cognitive screening tests alone are useful in determining the degree of cognitive impairment, but the evidence does not support their use as stand alone tests for delirium detection.

We recommend that to aid delirium detection in routine clinical practice the focus should be on rolling out user-friendly, easily transferable tools such as the 4AT, with audit and other methodologies to further evaluate their use.

Acknowledgements
We would like to thank Carol Hallesy and Maureen Graham, librarians for NHS Lanarkshire, for their assistance with this literature review.

Conflict of interest: none declared

References
1. Banerjee S, Chambers L: Old Age Psychiatry. Rai GS, Mulley GP editors: Elderly Medicine, a Training Guide. Oxford 2007, Elsevier Limited; Chapter 17: 161
2. NICE 103. 2010. Delirium: Diagnosis, Prevention and Management. https://www.nice.org.uk/guidance/cg103 (accessed 10/03/15)
3. Partridge JS, Martin FC, Harari D, et al. The Delirium Experience: What is the Effect on Patients, Relatives and Staff and What Can be done to Modify This? Int J Geriatr Psychiatry 2013; 28(8): 804–12
4. Collins N, Blanchard MR, Tookman A, et al. Detection of Delirium in the Acute Hospital. Age Ageing 2010; 39(1): 131–35
5. Elie M, Rousseau F, Cole M, et al; Prevalence and Detection of Delirium in Elderly Emergency Department Patients. CMAJ 2000; 163(8): 977–81
6. Inouye SK, Foreman MD, Mion LC, et al; Nurses’ Recognition of Delirium and its Symptoms: Comparison of Nurse and Researcher Ratings. Arch Intern Med 2001; 161(20): 2467–73
7. Han JH, Zimmerman BA, Cutler N, et al; Delirium in Older Emergency Department Patients: Recognition, Risk Factors, and Psychomotor Subtypes. Acad Emerg Med. 2009; 16: 193–200
8. Neufeld KJ, Hayat MJ, Coughlin JM, et al; Evaluation of Two Intensive Care Delirium Screening Tools for Non-Critically Ill Hospitalized Patients. Psychosomatics 2011; 52(2): 133–40
9. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders: DSM V. 5th ed. 2013
10. Morandi A, Davis D, Taylor JK, et al. Consensus and Variations in Opinions on Delirium Care: A Survey of European Delirium Specialists. Int Psychogeriatr 2013; 25(12): 2067–75
11. Wei LA, Fearing MA, Sternberg EJ, et al; The Confusion Assessment Method (CAM): A Systematic Review of Current Usage. J Am Geriatric Soc 2008; 56(5): 823–30
12. Laurila JV, Pitkala KH, Strandberg TE, et al; Confusion Assessment Method in the Diagnostics of Delirium among Aged Hospital Patients: Would it Serve Better in Screening than as a Diagnostic Instrument? Int J Geriatric Psychiatry 2002; 17: 1112–19
13. Monette J, Galbaud du Fort G, Fung SH, et al; Evaluation of the Confusion Assessment Method (CAM) as a Screening Tool for Delirium in the Emergency Room. Gen Hosp Psychiatry 2001; 23(1): 20-25
14. Schuurmans MJ, Donders RT, Shortridge-Baggett LM, et al. Delirium Case Finding: Pilot Testing of a New Screening Scale for Nurses. J Am Geriatr Soc 2002; 50: S3
15. Schuurmans MJ, Shortridge-Baggett LM, Duursma SA. The Delirium Observation Screening Scale: A Screening Instrument for Delirium. Research and Theory for Nursing Practice 2003; 17(1): 31–50
16. Leung JL, Leung VC, Leung CM, et al; Clinical Utility and Validation of Two Instruments (the Confusion Assessment Method Algorithm and the Chinese Version of Nursing Delirium Screening Scale) to Detect Delirium in Geriatric Inpatients. General Hospital Psychiatry 2008; 30: 171–76
17. Neelon VJ, Champagne MT, Carlson JR, et al. The NEECHAM Confusion Scale: Construction, Validation, and Clinical Testing. Nursing Research 1996; 45: 324–30
18. http://www.the4at.com/ (accessed 12/03/15)
19. Bellelli G, Morandi A, Davis D, et al. Validation of the 4AT, a New Instrument for Rapid Delirium Screening: A Study in 234 Hospitalised Older People. Age Ageing 2014; 43(4): 496–502
20. Lees R, Corbet S, Johnston C, et al. Test Accuracy of Short Screening Tests for Diagnosis of Delirium or Cognitive Impairment in an Acute Stroke Unit Setting. Stroke 2013; 44: 3078–83