When doctors ask AI the questions of clinical practice
DANADE 2026 featured 11 presentations directly addressing artificial intelligence, from eye clinics and thyroid ultrasound to intensive care. What stands out is not the number but the questions: what does AI do for patients, and how do we know it really helps?
The 3rd Da Nang Endocrinology, Diabetes and Metabolic Disorders Conference (DANADE 2026) took place in Da Nang on 21–22 November 2026, co-organised by the Vietnam Association of Diabetes and Endocrinology, the Da Nang Department of Health, the Da Nang Endocrinology and Diabetes Association and Vinmec Da Nang International Hospital. This year’s theme was “From medical knowledge to AI applications: a comprehensive shield against endocrine–metabolic disease and multi-organ complications”. Even in the theme, AI comes after medical knowledge, as an extension of it rather than a replacement.
AI across many sessions, not only the AI session
The dedicated session “Artificial intelligence in endocrinology, diabetes and its complications” had seven presentations, ranging from risk prediction and individualised treatment in type 2 diabetes, risk stratification from real-world data and the integration of AI with multi-omics and regenerative medicine, to diabetic retinopathy screening, diabetic macular oedema, the diagnosis of endocrine disease and AI-assisted thyroid nodule ultrasound for thyroid cancer screening in Vietnam.
Four further presentations brought AI into other clinical sessions:
- Geriatrics: an AI-enabled model for personalised health management at the Vinmec Da Nang Longevity Centre.
- Obesity: individualised multi-organ protection for people with obesity in the AI era.
- Critical care: AI in haemodynamic assessment of critically ill and surgical patients.
- Clinical nutrition: the role of AI, alongside a nutrition support team, in outcomes for mechanically ventilated ICU patients.
That AI appears even in sessions not named after it shows the topic has moved out of a technology corner and into the everyday professional conversation of many specialties.
Questions that start from real clinical work
Each specialty asks AI a different question:
- Ophthalmologists ask at which step of retinal screening AI can help, so that people who need specialist review are found early and referred to the right place.
- Endocrinologists ask whether clinical data can help decide who needs closer follow-up and who needs further assessment for complications.
- Sonographers ask how AI can support thyroid nodule classification on imaging data from Vietnamese patients.
- Intensivists ask whether AI can support timelier haemodynamic assessment and better nutritional care for ventilated patients.
None of these questions starts from “there is new technology, let’s try it”. Each starts from a real bottleneck in care. A screening tool is only valuable if the people it flags actually reach the next step of treatment. A risk model is only useful if its output changes the follow-up plan.
Room for scepticism
The session did not only present potential. A talk on diabetic macular oedema asked in its very title whether AI in treatment is “myth or hype?”. Another, on the diagnosis of endocrine disease, asked plainly: “where are we now?”.
Putting these critical questions in the main session shows a more mature discussion: doctors are interested in technology, and at the same time demand evidence of real value for patients.
Vinmec Da Nang and my own role at DANADE 2026
Vinmec Da Nang took part as a co-organiser, with 10 speakers and 12 scientific presentations, including talks in English and presentations by nurses. For me it was a conference of several roles: Co-Chair of the Organising Committee, Member of the Scientific Council, and speaker at the opening ceremony on behalf of the co-organiser and as Vice President of the Da Nang Endocrinology and Diabetes Association.
I co-chaired two sessions: Artificial intelligence in endocrinology, diabetes and complications, with Assoc. Prof. Phan Huong Duong, Deputy Director of the National Hospital of Endocrinology; and Diabetic foot and peripheral arterial and venous disease, with People’s Teacher Prof. Huynh Van Minh.
My two presentations on behalf of Vinmec Da Nang are two consecutive steps on the same path. The first, “AI in endocrine–metabolic disease: from real-world data to risk stratification and prevention of multi-organ complications”, drew on 86,808 visits by 38,199 patients at Vinmec Da Nang, 29.7% of them by people aged 40 and over. It proposed moving from “disease – treatment – complication” to “data – risk assessment – early prediction – prevention”, with a four-tier risk model matched to four intervention strategies. That is the step of identification: who is at high risk, and which problems come first.
The second introduced the AI-enabled model at the Vinmec Da Nang Longevity Centre, with five pillars (precision prevention, healthy ageing, chronic disease optimisation, AI and digital health, and international longevity) and a 2027–2030 roadmap from a Healthy Aging Clinic to the Longevity Centre. That is the next step: a risk profile must become a follow-up schedule, care goals and a prevention plan for each individual, aiming to extend healthy years, not just lifespan.
The hardest part is the transition between the two: when a number on a screen has to become a decision that doctor and patient both understand, agree on and carry out together. Slides of both talks are openly archived on Zenodo (DOI 10.5281/zenodo.24438096 and 10.5281/zenodo.24438098).
Understanding AI is becoming a professional competence
The programme shows interest and directions of application. The real-world effectiveness of each tool can only be established through local validation data and patient follow-up. Health professionals need not become data engineers, but increasingly they need to answer three questions about every AI tool:
- On which patient population was this tool built and validated, and does it resemble my patients?
- Where can it be wrong, and what happens when it is?
- How should its output be checked before it informs a care decision?
That level of understanding helps teams choose the right tasks for technology, and recognise when to stop and take a second look.
When doctors ask whether AI can help detect complications early, prioritise those at highest risk or assess a critically ill patient in time, technology has entered the conversation about quality of care. That is where the value of AI is measured: by outcomes that matter to patients.
This content is for information only and does not replace individual medical advice.
Further reading
Sources
This article reflects the author’s personal views and does not represent Vinmec Healthcare System.