Delirium severity: how do (or should) we conceptualise and measure it?

Contributed by Zoë Tieges, PhD, Psychology Research Fellow, Geriatric Medicine, Usher Institute, The University of Edinburgh

Severity grading of delirium in research and clinical practice may have important value in monitoring clinical course and recovery, in providing prognostic information for risk stratification, in informing treatment, and as endpoints in clinical trials. Fine-grained measures of delirium severity can provide valuable information beyond that provided by a yes/no diagnosis, through more fully capturing the complex and heterogeneous phenomenology of delirium [1]. Scoring systems of delirium severity have generally operationalized severity through summing ordinal measures of various domains of the delirium syndrome such as inattention, hallucinations and delusions, and motor features. These scales generate a single severity score reflecting both the absolute number and intensity of symptom domains present.

Over the years a wide range of instruments for measuring delirium severity have been developed. In a recent comprehensive review, Jones and colleagues [2] identified 42 different instruments, of which 11 covered multiple delirium symptom domains and provided quantitative ratings of delirium severity; these included the Delirium Severity-Rating Scale 96 (DRS-R98) [3], the Memorial Delirium Assessment Scale [4] and the Confusion Assessment Method-Severity [5].

What do we mean by delirium severity?

We have tools to measure delirium severity, but do we agree on what we are trying to measure? What do we mean to say that someone has a severe delirium? This may seem like an obvious question, but the answer is far from simple. Does severity of delirium refer to the number of symptoms a patient is experiencing, or their intensity, frequency or duration, or perhaps a combination of all of these? Does severity refer to the distress and suffering of a person as a result of their delirium? To the probability of the delirium resulting in death or other undesirable outcomes? Severity clearly is a complex construct which means different things to different stakeholders [6]. Note that DSM-5 defines severity for some disorders, but not for delirium [7].

To my mind, this uncertainty about how to conceptualise and operationalise delirium severity remains a fundamental knowledge gap which hampers progress in the field of delirium.

Survey of expert opinion on the nature of delirium severity

In search for answers, we surveyed expert opinions (>300 delirium specialists) on delirium severity [8]. With consensus defined as ≥75% agreement, experts agreed that poor outcomes, symptom intensity, patient safety and distress, and delirium duration were all important markers of severity. Interestingly, there was no consensus that the number of symptoms indicates severity (51% agreement). Respondents also had widely differing views on which delirium symptom domains could predict poor prognosis, with consensus only on “altered level of arousal” (83%). These findings highlight the contrast between expert opinion and current severity instrumentation, with the latter emphasising symptom counts and not always including altered arousal.

This is problematic, because many (older) hospitalised patients with acute illness and delirium have reduced level of arousal – or ‘reduced consciousness’ – that is so severe that they are unable to engage with cognitive testing or interview. As a result, drowsy, hypoactive patients will typically score low on severity rating scales, because many scale items can be ascertained only through bedside testing: attention, language, memory, orientation, etc. For example, the DRS-R98 is a very detailed tool with excellent instructions, but in our studies we often found low scores in hypoactive delirious patients. Some tools, such as the Delirium Index [9], account for this by assigning maximum item scores when patients are ‘untestable’ – which is often another way of saying that the patient has severe hypo- or hyper- arousal.

The survey also showed that poor prognosis was the most agreed marker of delirium severity, i.e. delirium that is the most likely to result in poor outcomes should be classed as most severe. This idea is supported by studies reporting associations between higher delirium severity scores with worse outcomes [5], though the evidence is mixed [10]. Studies that have linked motor subtypes to outcomes broadly suggest that the hypoactive delirium subtype (drowsy, withdrawn patients) has a worse prognosis [11, 12].

Delirium severity and outcomes

To date there has been surprisingly little investigation into the associations between individual symptom domains of delirium with outcomes. The available studies have largely focused on altered arousal, broadly showing that this feature is associated with higher mortality [13]. Of interest, a recent systematic review in hospitalised patients found an adjusted 6-fold greater risk of 30-day mortality in patients with altered arousal, though none of the studies considered delirium [14]. We need more research to increase understanding of which (if any) delirium symptom domains drive the association between a delirium diagnosis with poor outcomes, which will impact on how we think of delirium severity.

Understanding the purpose of measurement of delirium severity

The question of how severity should be defined and operationalised is closely linked to another question: what is the purpose of using a severity measure? Is it to quantify the phenomenological intensity, to reflect underlying pathology, to guide treatment, to recognise patient distress, or for risk stratification? Schulman-Green et al. [15] helpfully distinguish between the phenomenological perspective which focuses on what delirium is (features intrinsic to the delirium episode) versus the impact perspective which focuses on what delirium does. Current severity tools focus mainly on symptom counts and intensity (phenomenology), but this may well be considered the least important in clinical practice. On the other hand, defining severity in terms of prognosis (impact) is tricky because prognosis is intertwined with treatment and other factors unrelated to the disorder.

Where do we go from here?

We don’t have the knowledge yet to fill these gaps, but there is reason for optimism. Firstly, there has been clear recent progress in understanding delirium severity. A prime example is the study by Vasunilashorn et al. [16] who found that, out of nine different severity measures, a combination of symptom intensity and duration showed the strongest association with outcomes. Vasunilashorn et al. [6] also did important work on selecting delirium items based on psychometric and clinical input and field-testing them, laying the foundation for a new delirium severity instrument.

Secondly, a number of novel assessment tools have been introduced in recent years for measuring key symptom domains of delirium in a graded manner. These include the DelApp smartphone test for the objective assessment of altered arousal and attention deficits in delirium [17], and the Observational Scale for Level of Arousal [18]. Scores on these tools appear responsive to within-person fluctuations in delirium severity scale scores over time, supporting the utility of these tools for assessing an aspect of delirium severity.

Thirdly, the research base on delirium is rapidly expanding and the overall quality of studies is steadily improving. The volume of potentially available high-quality data on delirium and patient outcomes and the possibilities for using these resources to progress the delirium research agenda are exciting. Further, data from tools like the 4AT which both (a) has built-in arousal items [19] and also (b) large-scale clinical implementation data [20] could be used to analyse relationships between delirium features and outcomes at scale. The NIDUS Delirium Research Hub provides a great resource of indexed delirium studies to help facilitate collaborative efforts.


To conclude, there remains considerable uncertainty regarding the definition, measurement and applicability of delirium severity. Delirium severity is a multi-dimensional construct that can be measured in different ways depending on the stakeholders’ perspectives. Nevertheless, I think that at the heart of any delirium severity tool should be reliable, graded assessment of measurable core symptom domains (inattention, altered arousal). Severely reduced arousal in the context of delirium means that other cognitive and psychiatric symptom domains can often not be ascertained, and this should be accounted for in severity scoring systems. Distress in relation to severity and prognosis is poorly addressed and requires further study. Psychosis can be distressing and may be phenomenologically intense, but the association between florid psychotic features and patient outcomes is not very clear. We need more research to solve these important pieces of the delirium puzzle.


  1. Tieges, Z., et al., The neuropsychology of delirium: advancing the science of delirium assessment. Int J Geriatr Psychiatry, 2017. 33: p. 1501-1511.
  2. Jones, R.N., et al., Assessment of Instruments for Measurement of Delirium Severity: A Systematic Review. JAMA Intern Med, 2019. 179: p. 231-239.
  3. Trzepacz, P.T., et al., Validation of the Delirium Rating Scale-revised-98: comparison with the delirium rating scale and the cognitive test for delirium. J Neuropsychiatry Clin Neurosci, 2001. 13: p. 229-42.
  4. Breitbart, W., et al., The Memorial Delirium Assessment Scale. J Pain Symptom Manage, 1997. 13: p. 128-37.
  5. Inouye, S.K., et al., The CAM-S: development and validation of a new scoring system for delirium severity in 2 cohorts. Ann Intern Med, 2014. 160: p. 526-33.
  6. Vasunilashorn, S.M., et al., New Delirium Severity Indicators: Generation and Internal Validation in the Better Assessment of Illness (BASIL) Study. Dement Geriatr Cogn Disord, 2020. 49: p. 77-90.
  7. Zimmerman, M., T.A. Morgan, and K. Stanton, The severity of psychiatric disorders. World Psychiatry, 2018. 17 p. 258-275.
  8. Tieges, Z., et al. Diversity of opinions on delirium severity: an international survey of 344 delirium specialists. in 12th Annual Meeting of the European Delirium Association, 2017. Oslo, Norway.
  9. McCusker, J., et al., The delirium index, a measure of the severity of delirium: new findings on reliability, validity, and responsiveness. J Am Geriatr Soc, 2004. 52: p. 1744-9.
  10. Rosgen, B.K., et al., The association of delirium severity with patient and health system outcomes in hospitalised patients: a systematic review. Age Ageing, 2020. 49: p. 549-557.
  11. Bellelli, G., et al., The impact of psychomotor subtypes and duration of delirium on 6-month mortality in hip-fractured elderly patients. Int J Geriatr Psychiatry, 2018.
  12. Avelino-Silva, T.J., et al., Prognostic effects of delirium motor subtypes in hospitalized older adults: A prospective cohort study. PLoS ONE, 2018. 13: p. e0191092.
  13. Han, J.H., et al., Impaired arousal at initial presentation predicts 6-month mortality: an analysis of 1084 acutely ill older patients. J Hosp Med, 2014. 9: p. 772-8.
  14. Todd, A., et al., Reduced level of arousal and increased mortality in adult acute medical admissions: a systematic review and meta-analysis. BMC Geriatr, 2017. 17: p. 283.
  15. Schulman-Green, D., et al., Use of an expert panel to identify domains and indicators of delirium severity. Qual Life Res, 2019. 28: p. 2565-2578.
  16. Vasunilashorn, S.M., et al., Quantifying the Severity of a Delirium Episode Throughout Hospitalization: the Combined Importance of Intensity and Duration. J Gen Intern Med, 2016. 31: p. 1164-71.
  17. Tieges, Z., et al., A smartphone-based test for the assessment of attention deficits in delirium: A case-control diagnostic test accuracy study in older hospitalised patients. PLoS ONE, 2020. 15: p. e0227471.
  18. Hall, R., et al., The Observational Scale of Level of Arousal: A brief tool for assessing and monitoring level of arousal in patients with delirium outside the ICU. Int J Geriatr Psychiatry, 2020. 35: p. 1021-1027.
  19. Tieges, Z., et al., Diagnostic accuracy of the 4AT for delirium detection in older adults: systematic review and meta-analysis. Age Ageing, 2020.
  1. Royal College of Physicians, National hip fracture database annual report September 2018. 2018, NHFD.


Posted in NIDUS Blog and tagged , , .

Leave a Reply