Podcast: Play in new window

Subscribe: RSS

Date: April 2, 2026  

Reference: Lee et al. GRADE-Based Clinical Practice Guidelines for Emergency Department Delirium Risk Stratification, Screening, and Brain Imaging in Older Patients With Suspected Delirium. AEM Feb 2026

Guest Skeptic: Dr. Christina Shenvi is a board-certified emergency physician, educator, keynote speaker, coach, and academic leader. She is widely recognized for her work in geriatric emergency medicine, faculty development, and professional identity formation in EM. She brings deep clinical expertise along with thoughtful perspectives on systems-level change and guideline development.

Case: An 82-year-old woman with hearing impairment and mild baseline dementia is brought to the emergency department (ED) by her daughter because she became “not herself” over 24 hours. She is more sleepy, intermittently agitated, keeps losing the thread of conversation, and cannot say the months backward. She arrived by ambulance from home after nearly falling twice. Vitals show fever and mild tachycardia. The daughter reports foul-smelling urine and poor oral intake for two days. On examination, there is no head trauma and no focal neurologic deficit. The question in the ED is not simply “Is she confused?” but “Does she have delirium, how do we confirm it efficiently, and does she need a head CT as part of the workup?”

Background: Delirium is an acute brain dysfunction: a disturbance in attention and awareness that develops over hours to days, fluctuates, and is accompanied by additional cognitive disturbances such as memory, language, orientation, or perceptual changes. In older adults, it is common, dangerous, and often goes unnoticed. The latest GED Delirium Guidelines indicate that delirium occurs in about 6% to 38% of older ED patients, increases mortality, contributes to functional decline, and imposes a significant burden on health systems. ED-based geriatric screening tools also highlight that delirium is frequently under-recognized by emergency clinicians and that hypoactive delirium is most common, making bedside detection even more challenging. 

For emergency physicians, delirium matters because it is rarely the final diagnosis. Delirium is usually a clue that something else serious is also wrong. The practical ED task is to identify the syndrome, search for precipitants, and avoid worsening the situation. But one reason the new guideline is so useful is that it is honest about the evidence gap. Prior reviews found no consistent ED-based strategy to prevent incident delirium or to treat prevalent delirium, so this guideline appropriately focuses on the parts of care for which there is sufficient evidence to guide bedside decisions now. It addresses risk stratification, diagnosis, and brain imaging. 

This delirium guideline is also notable because it was built using the newer GED 2.0 model for subspecialty guideline development [1].  The Geriatric Emergency Department initiative moved beyond the older consensus-based 2014 framework and adopted a transparent GRADE process: multidisciplinary working groups, explicit PICO questions, systematic reviews and meta-analyses, Evidence-to-Decision frameworks, attention to feasibility, equity, and stakeholder values, plus external stakeholder review. This SGEM episode highlights the first EM subspecialty guideline effort to fully adopt GRADE, and this delirium guideline shows that process in action. 

Clinical Questions:

  1. Which older ED adults are at the highest risk on walking in, and who should then be further assessed for delirium? (or CLS addition, should have special prevention measures or expedited treatment or bed placement).

  2. Which tools should be used to identify ED delirium?

  3. Should acutely confused older ED patients undergo head CT as part of the delirium evaluation?

Reference: Lee et al. GRADE-Based Clinical Practice Guidelines for Emergency Department Delirium Risk Stratification, Screening, and Brain Imaging in Older Patients With Suspected Delirium. AEM Feb 2026

Authors’ Conclusions: “Rigorous ED-based research is needed to strengthen evidence and guide delirium care for older adults in geriatric emergency medicine.”

Quality Checklist for a Guideline:

  1. The study population included or focused on those in the emergency department? Yes 
  2. An explicit and sensible process was used to identify, select and combine evidence? Yes
  3. The quality of the evidence was explicitly assessed using a validated instrument? Yes
  4. An explicit and sensible process was used to value the relative importance of different outcomes? Yes
  5. The guideline thoughtfully balances desirable and undesirable effects? Yes
  6. The guideline accounts for important recent developments? Yes
  7. Has the guideline been peer-reviewed and tested? Yes/No
  8. Practical, actionable and clinically important recommendations are made? Yes
  9. The guideline authors’ conflicts of interest are fully reported, transparent and unlikely to sway the recommendations? Unsure

Key Recommendations: They came up with six recommendations that are conditional and all rest on very low certainty of evidence.

  • Risk Stratification
    • A Delirium Risk Score may be used to identify low-risk older adults. Conditional FOR; very low certainty. The delirium risk score gives you a baseline risk of delirium before the patient comes into the ED. It’s based on work by Jin Han and friends from 2009. In a cross-sectional convenience sample of about 300 patients, they identified three factors were independently associated with presenting with delirium in the ED: dementia, Katz ADL<=4, and hearing impairment. They then developed a score based on those three factors and found that higher scores were associated with more delirium If you have none of those factors, then you have <1% prevalence of delirium on arrival (prevalent vs incident delirium). If you have all three factors, then the rate is close to 50%. So, you could conceivably only screen patients who have at least 1 of those factors, now known as the delirium risk score, to avoid screening every single patient.
    • For adults >65 admitted to ED observation, Zucchelli’s tool with threshold ≥4 may be used to identify low- or high-risk patients [2]. Conditional FOR; very low certainty. Zucchelli’s age >= 75, dementia, hearing impairment, and psychotropic drugs. High risk is >= 4. This could be used to identify patients at high risk.
    • For adults >75 presenting to the ED, REDEEM threshold ≥11 may identify low- or high-risk patients, and a cutoff <5 may help rule out high risk [3]. Conditional FOR; very low certainty. REDEEM score of >=11 would mean about 1/5th or 20% of the ED older adult population, so then if you only screened those patients for delirium, then it may reduce the work. But that means you’re screening everyone with REDEEM and then 20% of them with a delirium diagnostic tool. And the redeem questions are more complex (see paper). REDEEM stands for Recognizing Delirium in geriacric Emergency medicine and is based off of a paper from 2021. It was an observational study of patients age >=75 coming to the ED who were screened for delirium and then regression analysis used to create a model with 10 variables – 7 from triage info and 3 obtained during early history. Some caveats here. Also scores from -3 to +66. This to me seems more complex than just screening everyone for delirium with a DTS and bCAM or the 4AT
  • Diagnostic Tools
    • 4AT, bCAM, CAM-ICU, mCAM, AMT-4, or RASSmay be used to rule delirium in or out. Conditional FOR; very low certainty.
    • Delirium Triage Screen (DTS) may be used to rule out, but not rule in, delirium [4]. Conditional FOR; very low certainty.

The DTS and bCAM are frequently used in concert. Start with the DTS (2 components: RASS=/0 or ALOC and attention – spell LUNCH backwards). If normal, NOT delirious. If not normal, do the bCAM:

  1. Acute change or fluctuating course – from family or pt
  2. Inattention eg months of the year backwards
  3. Altered level of consciousness or
  4. Disorganized thinking eg will a stone float on water

If you have 1, 2, and 3 or 4, then you are delirious.

4AT is often used internationally for 4 items and a single step.

  1. Alertness, normal/mild sleepiness/clearly abnormal
  2. AMT4 (abbreviated mental test for orientation)
    1. Age
    2. Date of birth
    3. Location
    4. Year
  3. Attention – months of the year backwards
  4. Acute change or fluctuating course

Score 0 – normal, 1 to 3 possible cognitive impairment, >=4 possible delirium or severe cog impairment

  • Brain Imaging
    • In the undifferentiated older ED patient with delirium or altered mental status (AMS), there are insufficient data to recommend for or against a head CT. Conditional EITHER; very low certainty.

1. Certainty: The biggest limitation is the certainty of the underlying evidence. The guideline authors are admirably transparent about the fact that every recommendation is conditional and based on very low-certainty evidence. In GRADE terms, this means the true effect may be substantially different from the estimate, especially when evidence is downgraded for risk of bias, indirectness, inconsistency, and imprecision. That matters because even a methodologically rigorous guideline can only be as trustworthy as the body of evidence beneath it. 

2. External Validity: A second threat is that much of the evidence base consists of single-center diagnostic or prognostic studies with limited external validation. The delirium paper explicitly notes that the risk-stratification tools were supported by single-center evidence and, in most cases, lacked separate validation research. From an evidence-based diagnosis perspective, this raises real concerns about spectrum effects and transportability: a tool that appears to work in one ED case mix may perform differently in another. 

3. Biases: Diagnostic test studies are particularly vulnerable to recognizable biases such as spectrum bias, incorporation bias, verification bias, and imperfect gold standard bias [5]. That is not a theoretical problem here; delirium has no perfect ED criterion, patient severity varies substantially, and some studies may include more selected populations than the real-world undifferentiated older adult seen at triage. As the diagnostic-bias literature emphasizes, these biases can systematically exaggerate or underestimate sensitivity and specificity [6].

4. POO: Many of the recommendation-driving outcomes are diagnostic-accuracy or likelihood-ratio outcomes rather than patient-oriented outcomes (POO). That is understandable, because the ED delirium literature has not yet identified a consistent intervention that clearly improves downstream outcomes. But it does mean the chain from “better screening” to “better patient outcomes” remains partly assumed rather than proven.

5. Thresholds: The guideline authors deliberately relaxed their likelihood-ratio threshold for including tools, noting that conventional thresholds would likely have excluded every delirium prognostic and diagnostic instrument from the guideline. That is a pragmatic and defensible choice in a sparse evidence base, but it also introduces a degree of arbitrariness into which tools “make the cut.” In other words, the recommendations are built partly on a methodological compromise designed to keep the field moving rather than on a clearly overwhelming body of evidence.

SGEM Bottom Line: This guideline provides EM docs with a reasonable roadmap for delirium risk stratification, bedside diagnosis, and selective imaging. However, the signal is useful only if we keep our skeptical radar set to high because the underlying evidence is still very shaky.

Case Resolution: Your 82-year-old patient is exactly the kind of patient who should trigger a delirium-first frame. She is older, arrived by ambulance, has baseline dementia and hearing impairment, and now has an acute fluctuating change in attention and awareness. These are all features that should raise our concern for delirium. At the bedside, use a validated ED tool such as the 4AT or bCAM [7] rather than Gestalt alone. If she screens positive, treat the likely precipitants aggressively while using nonpharmacologic supports and reserving medications for unsafe agitation. Because she has no trauma and no focal neurologic deficit, the guideline does not force a routine head CT; instead, it says the evidence is insufficient to recommend for or against routine imaging, so CT becomes a selective judgment call rather than an automatic reflex. In this case, I would confirm delirium with a validated tool, treat the probable infectious trigger, optimize the environment, involve family, and forgo immediate head CT unless new neurologic red flags emerge.

Clinical Application: Use the guideline as a guide, not as a mandate. If your ED already has the staff and culture to screen broadly, the paper does not argue against that. But if universal delirium screening is not in place, these recommendations support a pragmatic approach: use selective risk stratification to focus scarce attention, then use a single, consistently validated bedside tool rather than relying on gestalt. The strongest practical move is probably not choosing the “perfect” tool, but choosing one your department will use reliably. 

Dr. Christina Shenvi

For imaging, the key clinical takeaway is to exercise restraint in judgment. The guideline does not endorse blanket head CT for every older patient with delirium, but it also does not say never scan. It leaves room for selective imaging when there are additional red flags, such as trauma or focal neurologic findings, and it highlights just how uncertain the evidence remains. That is a useful message for emergency physicians, because it legitimizes individualized decision-making rather than checkbox medicine. 

Finally, use these recommendations to tighten the ED workflow around delirium, not to end the conversation. Delirium screening matters only if it changes what happens next: medication review, infection screening, hydration, mobility, sensory optimization, family engagement, and safer disposition planning. The guideline is best used as one part of a delirium pathway rather than as a stand-alone test selection document.

There are still many questions I would like to see answered, like what measures in the ED can reduce incident delirium, and what measures for prevalent delirium can reduce inpatient LOS and improve outcomes such as mortality, discharged to a higher level of care (i.e., loss of function), etc.

Keener Kontest: Last week’s winner was Dr. Steven Stelts from NZ. He knew the primary physiologic mechanism by which HFNO reduces the work of breathing is to reduce the anatomical dead space (flushing out CO₂ from the upper airway) and improve ventilatory efficiency.

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.

References:

  1. Gunaga S, Carpenter CR, Kennedy M, Southerland LT, Lo AX, Lee S, Swan K, Mowbray F, Skains RM, Hogan TM, Casey MF, Ouchi K, George NR, de Wit K, Gettel CJ, Selman K, Ragsdale LC, Chary AN, van Oppen JD, Arendts G, Maddow CL, Hunold KM, Tyler KR, Khoujah D, Hwang U, Liu S. A Model for Developing Subspecialty Clinical Practice Guidelines: The Geriatric Emergency Department Guidelines 2.0. J Am Coll Emerg Physicians Open. 2025 Sep 16;6(6):100247. doi: 10.1016/j.acepjo.2025.100247. PMID: 41019914; PMCID: PMC12476112.
  2. Zucchelli A, Apuzzo R, Paolillo C, et al. Development and validation of a delirium risk assessment tool in older patients admitted to the emergency department observation unit.
    Aging Clinical and Experimental Research. 2021;33(10):2753–2758.
  3. Oliveira J E Silva L, Stanich JA, Jeffery MM, Mullan AF, Bower SM, Campbell RL, Rabinstein AA, Pignolo RJ, Bellolio F. REcognizing DElirium in geriatric Emergency Medicine: The REDEEM risk stratification score. Acad Emerg Med. 2022 Apr;29(4):476-485. doi: 10.1111/acem.14423. Epub 2021 Dec 17. PMID: 34870884; PMCID: PMC9050857.
  4. Han JH, Wilson A, Vasilevskis EE, Shintani A, Schnelle JF, Dittus RS, Graves AJ, Storrow AB, Shuster J, Ely EW. Diagnosing delirium in older emergency department patients: validity and reliability of the delirium triage screen and the brief confusion assessment method. Ann Emerg Med. 2013 Nov;62(5):457-465. doi: 10.1016/j.annemergmed.2013.05.003. Epub 2013 Jul 31. PMID: 23916018; PMCID: PMC3936572.
  5. Waikar SS, Betensky RA, Emerson SC, Bonventre JV. Imperfect gold standards for biomarker evaluation. Clin Trials. 2013 Oct;10(5):696-700. doi: 10.1177/1740774513497540. Epub 2013 Sep 3. PMID: 24006246; PMCID: PMC3800226.
  6. Kohn MA, Carpenter CR, Newman TB. Understanding the direction of bias in studies of diagnostic test accuracy. Acad Emerg Med. 2013 Nov;20(11):1194-206. doi: 10.1111/acem.12255. PMID: 24238322.
  7. Han JH, Wilson A, Graves AJ, Shintani A, Schnelle JF, Ely EW. A quick and easy delirium assessment for nonphysician research personnel. Am J Emerg Med. 2016 Jun;34(6):1031-6. doi: 10.1016/j.ajem.2016.02.069. Epub 2016 Mar 3. PMID: 27021131; PMCID: PMC4899095.