Dysphagia Knowledge Hub — 吞嚥困難知識庫

IDDSI Compliance Audits for Care Homes: A Practical Self-Audit Checklist

TL;DR: IDDSI compliance is not about paperwork — it is about whether the Level 4 puree in today’s lunch tray actually holds its shape on a fork. A good self-audit looks at seven domains: the food itself, the fluids, the diet order chain, staff training, mealtime practice, equipment, and corrective action. This article gives a care-home manager a repeatable checklist that can be done in under two hours per kitchen, using only a standard dinner fork, a 10 ml slip-tip syringe and a printable form.

Why audit at all?

Even care homes that “serve IDDSI” rarely produce food that passes the actual fork pressure, fork drip, spoon tilt and flow tests published by the International Dysphagia Diet Standardisation Initiative (IDDSI, 2019 Framework 2.0). The gap between “we purée the meat” and “this meal is truly IDDSI Level 4” is the gap in which aspiration pneumonia, choking, malnutrition and dehydration happen. The purpose of a self-audit is to surface those gaps before a resident, family member, dietitian, or regulator does.

Regulators already expect it. In Hong Kong, the HKCSS Care Food Directory 2023 and HKCSS 照護食標準指引 frame IDDSI as the reference standard for the sector. In the Greater Bay Area, T/SATA 084-2025 (《適老易食食品》, effective 7 June 2025) codifies IDDSI testing methods as mandatory for pre-packed care food, including fork, spoon tilt, and syringe flow tests (Appendix A) and texture-profile analysis (Appendix B). In Taiwan, 長照 2.0 professional services require nutrition and swallowing-feeding training to be delivered by registered dietitians and speech-language pathologists under the Ministry of Health and Welfare’s long-term care framework.

None of these frameworks police individual meals. That is the job of the care home itself.

What to audit — the seven domains

A complete IDDSI compliance audit covers:

1. Food texture — does each produced level actually pass the IDDSI tests?
2. Fluid thickness — does each thickened drink sit at the correct syringe flow range?
3. Diet order chain — does the resident receive the level the clinician actually prescribed?
4. Staff training — do kitchen, care, and nursing staff know the tests and the rules?
5. Mealtime practice — positioning, supervision, pacing, pre-thickening of drinks?
6. Equipment — is the testing equipment itself valid?
7. Corrective action — when something fails, is there a closed loop that fixes it?

The rest of this article walks through each domain with a practical check that a care-home manager, clinical lead, or kitchen supervisor can actually run.

1. Food texture — the on-plate verification

Pick three meals at random from today’s production. Do not tell the kitchen in advance. For each meal, run the test that applies to that level. All tests are performed at the intended serving temperature.

Level 4 Pureed. Scoop a spoonful and tilt the spoon sideways. The puree should hold shape, fall off in a single plop, leave only a thin film on the spoon, and sit on the plate as a mound that may spread slightly but should not flow. Press the fork tines onto a flat portion: the tines should leave a clear pattern, and the puree should not drip continuously through the prongs. It should not be sticky or firm, and no liquid should separate from the solid.

Level 5 Minced & Moist. Press vertically with a fork. Small lumps should separate and pass easily between the prongs (prong gap ≈ 4 mm on a standard dinner fork — this is why 4 mm is the paediatric and adult particle-width limit). Tilt the spoon 45°: the portion should slide off without sticking. Lumps must be easy to squash with little pressure, and there must be no thin liquid separating from the food on the plate.

Level 6 Soft & Bite-Sized. Select a single 1.5 × 1.5 cm piece (adult). Press with the side of a fork onto the piece until your thumbnail blanches on the fork’s flat surface. The piece should squash, break apart, and fail to return to its original shape. Confirm that no piece in the portion exceeds 15 mm (adults) or 8 mm (paediatric). Bones, gristle, fibrous parts, and hard skins are disqualifying.

Level 7EC Easy to Chew. Same thumbnail-blanch fork pressure test as Level 6 — the piece must not return to shape — but size is unrestricted. There must be no hard, tough, chewy, fibrous, crunchy, crumbly, pip-bearing, or skin-on components.

If a portion fails, record it. Do not re-serve. Ask the kitchen what step caused the drift — usually under-cooking, under-blending, or a recipe that did not account for the starch continuing to thicken during hot holding.

2. Fluid thickness — the syringe flow test

Every thickened fluid served on the menu must be sampled. Use a 10 ml BD-style slip-tip syringe with the barrel length from the 10 ml mark to the 0 ml mark equal to 61.5 mm. Measure it once with a ruler — if the syringe is longer or shorter, the test is invalid (IDDSI Flow Test, 2019). Before sampling, confirm the thickened drink has rested long enough for the thickener to hydrate (usually 1–3 minutes for xanthan-gum powders, longer for starch). Cap the tip with a finger, fill to 10 ml, lift the finger, and time exactly 10 seconds.

• Level 0 (Thin): < 1 ml remaining
• Level 1 (Slightly Thick): 1–4 ml remaining
• Level 2 (Mildly Thick): 4–8 ml remaining
• Level 3 (Moderately Thick / Liquidised): > 8 ml remaining
• Level 4: does not flow — use fork/spoon tests, not syringe

Three common failures: (a) thickener added to a drink that is too hot and drifts thinner as it cools, (b) pre-mixed jugs of thickened drinks that thicken further over the shift because starch-based thickeners keep gelling, (c) staff guessing the level by appearance instead of using the syringe. A jug labelled “Level 2” that is actually Level 3 by syringe is a genuine aspiration risk — the resident was prescribed Level 2 for a reason.

3. Diet order chain

For a stratified sample of six residents (two on Level 4, two on Level 5, two on Level 6), trace the paper or electronic trail from the SLP or physician’s written diet order to the tray that arrived at the bedside. Ask five questions:

• Does the prescribed IDDSI level (food) and thickness level (fluid) appear on the resident’s care plan, kitchen ticket, and tray card?
• Is the prescribed level the same across all three documents?
• Was the order reviewed after the last swallowing reassessment? (Date?)
• Is the resident allowed level-7 items alongside their prescribed level (e.g. a finger food), and is this explicitly documented?
• When the resident is transferred to hospital or home, does the discharge/transfer paperwork carry the IDDSI level in plain IDDSI language?

Any break in this chain is an incident. Residents have died from the wrong tray arriving at the wrong bed, sometimes because a new kitchen staffer assumed “soft” meant Level 5 when the order actually said Level 6. Use the IDDSI-standard language and colour codes (T/SATA 084-2025 Appendix C: Level 4 = green, Level 5 = orange, Level 6 = blue, Levels 7/7EC = black).

4. Staff training

Pick three staff at random from different roles: one kitchen chef, one care assistant, one registered nurse or HCA at the bedside. Ask each, without prompting:

• Name the 8 IDDSI levels.
• Show me how to do the fork drip test for Level 4.
• Show me how to do the syringe flow test for Level 2 or 3.
• What is the paediatric particle size limit for Level 5? (2 mm width × 8 mm length.)
• What is the adult bite-size limit for Level 6? (15 mm × 15 mm.)
• Which foods on today’s menu should never be served to a resident on Level 6? (Foods listed in the IDDSI Choking Risk table: nuts, raw carrot, stringy green beans, sticky rice cake, grapes whole, sausages, hot dogs, floppy lettuce, watermelon if juice separates, 乳豬脆皮, 小籠包, bubble tea, and so on.)

Training records should match the answers. If 70% of the sampled staff cannot demonstrate at least the fork drip and syringe tests, the training programme itself needs to be audited, not the staff.

5. Mealtime practice

Sit through one full meal service. Observe without intervening. Record:

• Is the resident’s chair angle ≥ 60° to upright? Head slightly flexed (chin-tuck position) where prescribed?
• Is there a supervising staff member within line of sight for every resident at medium or high aspiration risk?
• Are thickened drinks pre-thickened before leaving the kitchen, not at the table by untrained staff?
• Is the pace of feeding respectful — one spoonful, swallow, pause, second spoonful — or is the resident being rushed?
• Does any resident cough, pocket food in the cheek, show wet/gurgly voice after swallowing, or refuse food without explanation?
• Is there a mixed-consistency food on the menu today that should have been flagged (e.g. cereal with milk, soup with chunks, congee with meat floss)?

Every “no” or “yes, and staff missed it” is an audit finding. The single most common finding in our informal observation of Hong Kong and Taiwan care homes is thickened drinks being made at the bedside by a care assistant who eyeballs the thickener with a teaspoon instead of measuring it. Pre-thicken in the kitchen, label the jug with level and time of preparation, and discard after four hours.

6. Equipment

Five items, five checks, takes ten minutes:

• Dinner fork. Measure: width ≈ 15 mm, prong spacing ≈ 4 mm. Forks with wider gaps invalidate the Level 5 particle-size reference.
• 10 ml slip-tip syringe. Length from 10 ml to 0 ml mark must be 61.5 mm. Replace after visible wear.
• Kitchen scales. Can you weigh 1 gram accurately? Thickener dosing of 1.5 g per 100 ml (the rate used by The Project Futurus in Hong Kong, for reference) cannot be done by eye.
• Thermometer. Because all IDDSI tests are at intended serving temperature. A puree that is Level 4 at 65 °C may be thinner or thicker at room temperature.
• IDDSI colour labels. In use on trays, menus, and packaging? If not, transfer errors multiply.

7. Corrective action

Audits that do not close the loop are theatre. Every finding needs: a written note of what failed, the level it was meant to be at, the level it actually tested at, who was responsible for preparation, what the immediate correction was (most often: do not serve, re-prepare), and what the structural change is (recipe revision, training top-up, supplier change, equipment replacement). Track recurrence over quarters — if Level 5 minced meat fails the fork-drip test three quarters in a row, the recipe is wrong, not the chef.

Keep a single-page finding-and-action log per audit. At the next audit, start with the previous log and verify each item was closed.

Audit frequency

A workable default for a 50–100 bed residential home:

• Daily at the tray-assembly station: visual check that tray cards and menu match, syringe test on the batch jug of thickened drinks.
• Weekly full texture test on one randomly selected meal per produced level.
• Monthly full seven-domain audit of one meal service, including observation.
• Quarterly training refresh and documentation-chain audit.
• After any incident (choking, suspected aspiration pneumonia, family complaint): targeted audit of the resident’s specific diet chain and the shift that served that meal.

Common failures observed in Asia-Pacific care homes

In our review of HKCSS Care Food Directory operators and Taiwan 長照 facility reports, the recurring audit failures cluster around a handful of issues. Meat at Level 5 is often not finely enough minced (particles exceed 4 mm width or separate from sauce). Level 4 purees drift toward Level 3 at hot-holding temperature, especially starch-thickened purees. Thickened drinks made at the bedside with domestic teaspoons deliver inconsistent doses. Rice at Level 5 or 6 is often served glutinous or sticky (a choking risk). Congee-type breakfasts frequently contain a mixed thin–thick consistency (thin rice water plus solid grains) that is unsafe at Levels 4 and 5 without further modification. Oral care before and after meals — the single most evidence-based non-texture intervention to reduce aspiration pneumonia (Yoneyama 2002 and follow-up 2024 chlorhexidine evidence) — is frequently absent from the mealtime checklist.

Common mistakes

• Treating the audit as a one-person exercise. A good audit needs a chef, a nurse or SLP, and a manager — three lenses on the same meal.
• Announcing the audit in advance. You are auditing normal operation, not a showcase.
• Auditing only the kitchen. The most consequential failures happen between the kitchen and the mouth — in the diet order chain and at the bedside.
• Not running the actual tests. A visual glance at a tray is not an audit. Use the fork, the syringe, and the scales every time.
• No written log. If a finding cannot be traced in a quarter’s time, it did not happen.

Citations and sources

• Cichero JAY, Lam P, Steele CM, et al. “Development of International Terminology and Definitions for Texture-Modified Foods and Thickened Fluids Used in Dysphagia Management: The IDDSI Framework.” Dysphagia 32:293–314 (2017).
• International Dysphagia Diet Standardisation Initiative. IDDSI Framework 2.0, complete detailed definitions (2019).
• International Dysphagia Diet Standardisation Initiative. IDDSI Audit Tools. https://www.iddsi.org/Resources/Audit-Tools.
• 深圳市分析測試協會. T/SATA 084—2025 《適老易食食品（適老照護食）》. Effective 2025-06-07. Proposed by 香港中華廠商聯合會 + 香港社會服務聯會.
• 香港社會服務聯會. HKCSS 照護食標準指引 (2023). Care Food Directory Sections A–E.
• 衛生福利部. 長期照顧十年計畫 2.0（台灣）. Nutrition and swallowing-feeding training requirements under professional services framework.
• Yoneyama T, Yoshida M, Ohrui T, et al. “Oral Care Reduces Pneumonia in Older Patients in Nursing Homes.” Journal of the American Geriatrics Society 50:430–433 (2002).

This article paraphrases publicly available IDDSI, HKCSS, T/SATA, and Taiwan MOHW guidance. For clinical practice, refer to the current official documentation. This page is not medical advice.

---

Last updated: 2026-04-18 · License: CC BY 4.0 · Maintained by Editorial Team — a Hong Kong social enterprise producing IDDSI-compliant care food for people living with dysphagia. Training and audit consultation for care homes: [email protected]. This page is educational only; see About for our clinical partners and social mission.