Dysphagia Knowledge Hub — 吞嚥困難知識庫

EAT-10: The Eating Assessment Tool — Clinical Guide

Overview

The EAT-10 (Eating Assessment Tool) is a validated, 10-item self-report questionnaire designed to identify swallowing difficulties and quantify their impact on daily life. Developed by Belafsky and colleagues (2008), it has become one of the most widely used dysphagia screening instruments in clinical practice and research, with translations available in over 30 languages.

Its strengths lie in its brevity (completion takes under two minutes), patient-centredness, and robust psychometric properties. It does not replace clinical or instrumental assessment, but serves as a reliable first-pass screen that helps clinicians determine who needs further evaluation.

Development and Validation

The EAT-10 was derived from a larger item pool using item reduction methodology. Belafsky et al. (2008) validated it in a cohort of 482 patients presenting to a swallowing centre, demonstrating:

Internal consistency: Cronbach’s α = 0.960
Test-retest reliability: r = 0.72 at 2 weeks
Concurrent validity: significant correlation with Penetration-Aspiration Scale scores and VFSS findings
Cut-off score: ≥3 points indicates swallowing risk (sensitivity ~65–80%, specificity ~60–75% depending on population)

PMID: 18349030

The 10 Items

Each item is scored 0–4 (0 = no problem, 4 = severe problem):

My swallowing problem has caused me to lose weight.
My swallowing problem interferes with my ability to go out for meals.
Swallowing liquids takes extra effort.
Swallowing solids takes extra effort.
Swallowing pills takes extra effort.
Swallowing is painful.
The pleasure of eating is affected by my swallowing.
When I swallow, food sticks in my throat.
I cough when I eat.
Swallowing is stressful.

Total score range: 0–40. Cut-off for risk: ≥3.

Administration Protocol

Who administers it: Can be self-administered by literate patients, administered verbally by a nurse or healthcare assistant, or completed by a proxy caregiver for those with cognitive impairment.

Setting: Suitable for outpatient clinic, ward, community, or telehealth.

Time required: 1–2 minutes for most patients.

Language considerations: Use a validated translation in the patient’s primary language. Proxy completion by a caregiver is appropriate when the patient lacks literacy or has significant cognitive impairment, but this should be documented.

Cognitive caveat: The EAT-10 relies on subjective insight. Patients with moderate-to-severe dementia may under-report symptoms. In these populations, supplement with observational tools or caregiver-reported measures.

Scoring and Interpretation

Score	Interpretation
0–2	Within normal limits — routine monitoring
3–9	Mild swallowing risk — refer for clinical swallowing examination
10–20	Moderate risk — prioritise SLT referral; consider VFSS/FEES
≥21	Severe risk — urgent SLT review; assess nutrition and airway safety

A score ≥3 should trigger referral to a speech-language therapist (SLT) in most clinical protocols. Some centres use a combined screen: EAT-10 ≥3 plus water swallow test failure = high probability of aspiration.

Clinical Integration

Paired with water swallow test (WST): Combining EAT-10 ≥3 with WST abnormality increases specificity for aspiration risk compared to either tool alone (Rofes et al., 2014; DOI: 10.1007/s00455-014-9527-1).

Paired with GUSS: In stroke units, EAT-10 provides patient-reported context while GUSS provides the clinician-observed bedside screen. Together they offer complementary data.

Serial monitoring: EAT-10 is sensitive to clinically meaningful change over time (minimally important difference estimated at ~3 points), making it useful for monitoring treatment response in rehabilitation settings.

In oncology: Validated in head and neck cancer patients post-radiotherapy, where it detects late-onset dysphagia that patients often fail to report spontaneously (MD Anderson Dysphagia Inventory is a longer alternative for this population).

Limitations

Silent aspiration: The EAT-10 captures subjective symptoms and will not detect silent aspiration (aspiration without cough or patient awareness). A normal EAT-10 does not rule out aspiration.
Cognitive impairment: Self-report validity decreases with dementia severity.
Cultural variation: Eating-related questions (e.g., “going out for meals”) may carry different weight across cultures; validated translations should be used.
Not diagnostic: A high score indicates risk, not a specific diagnosis. VFSS or FEES is required to characterise the nature and severity of the swallowing impairment.

References

Belafsky PC, et al. (2008). Validity and reliability of the Eating Assessment Tool (EAT-10). Ann Otol Rhinol Laryngol, 117(12):919–24. PMID: 18349030
Rofes L, et al. (2014). Sensitivity and specificity of the Eating Assessment Tool and the swallow quality-of-life questionnaire for clinical screening of oropharyngeal dysphagia. Neurogastroenterol Motil, 26(9):1256–65. DOI: 10.1111/nmo.12382
Burgos R, et al. (2012). Adaptation and validation of the Spanish version of the EAT-10 (Eating Assessment Tool-10) for the screening of dysphagia. Nutr Hosp, 27(6):2048–54. PMID: 23588456