How do you document a diabetic foot assessment with an AI scribe?

Short answer: an AI scribe records what it hears, so a diabetic foot assessment documents well when you do four things. Say your neurological and vascular findings aloud. Use a template whose headings follow the screening the guidelines ask for. State the risk category yourself, and check it against the findings. Then draft the report to the GP from the same consult.

This article draws on the 2021 Australian evidence-based guidelines for diabetes-related foot disease, published by Diabetes Feet Australia. It's about documentation, not clinical advice, and the guideline itself is the authority. We build Hanah, an AI scribe for podiatrists, so read it with that in mind. Everything outside the Hanah-specific parts applies to any scribe.

What does a diabetic foot assessment need to record?

The Australian guideline on prevention of foot ulceration (Kaminski et al., 2022) sets out what the screening examination should include:

• Loss of protective sensation (LOPS): a 10 g Semmes-Weinstein monofilament, or the Ipswich Touch Test if a monofilament isn't available. If monofilament testing doesn't show a loss, vibration sense with a tuning fork or a biothesiometer/neurothesiometer.
• Peripheral artery disease (PAD): a cardiovascular history, palpation of foot pulses, pedal Doppler arterial waveforms and pressure measurements.
• For anyone at increased risk: a history of foot ulceration or lower-extremity amputation, end-stage renal disease, foot deformity (present or progressing), limited joint mobility, abundant callus, and any pre-ulcerative sign.

The findings then set a risk category, and the category sets when you screen again. The guideline uses the International Working Group on the Diabetic Foot (IWGDF) system:

The same guideline recommends structured foot self-care education for anyone at increased risk (risk 1 to 3): not walking barefoot, checking both feet and the inside of shoes daily, washing and carefully drying between the toes, using emollients for dry skin, cutting nails straight across, and not using chemical agents or plasters on callus or corns. It also recommends properly fitting medical grade footwear for people at moderate risk. The advice you gave belongs in the note as much as the findings do.

Why can't the scribe capture what I found?

Because it only hears the consult. Monofilament testing is silent. So is palpating a pulse. Unless you say the result, the transcript has nothing to work from, and a draft that fills the gap with a plausible "normal" is worse than one that leaves it blank.

A few habits make the difference:

• Say the side and the site every time. "Right, first metatarsal head: not felt" is unambiguous. "Not felt there" isn't.
• Say the negatives you want recorded. "No pre-ulcerative lesions, skin intact" should be on the record. If you don't say it, the note can't tell "normal" apart from "not checked".
• Say the category and the interval. "IWGDF risk 2, screen again in three months."

Calling findings aloud can double as the explanation the patient needs. Anything you'd rather not say in front of them, speak into the recording once they've left.

What should the template look like?

Use headings that mirror the screening, so a gap is easy to spot. For example:

• History: diabetes type and duration, previous ulcer or amputation, renal disease, current symptoms, footwear and self-care.
• Neurological: monofilament result by side and site; vibration if tested.
• Vascular: pulses, Doppler waveforms and pressures, by side.
• Skin and nails: callus, pre-ulcerative signs, nail pathology.
• Musculoskeletal: deformity and joint mobility.
• Risk: IWGDF category and next screening interval.
• Treatment and plan: what you did today, the education and footwear advice given, follow-up.

Add one rule to the instructions: omit anything not assessed. Hanah's built-in SOAP format already works this way: its objective section asks for one finding per line, with figures and units, and to omit anything that wasn't assessed. Put the same rule in your own template so "not tested" never turns into "normal".

In Hanah, a template is a set of drafting instructions: your headings, and a line under each saying what belongs there. You can write one from scratch, or upload the Word or PDF proforma you already use and have Hanah fill it from the consult.

Who decides the risk category?

You do. A scribe drafts what was said. Classifying risk is a clinical judgement, and under Ahpra's guidance on AI in healthcare you remain responsible for checking the accuracy of records an AI scribe creates.

A quick check before you save: read the neurological and vascular lines, then the category. If the draft records LOPS and a deformity but says risk 1, one of those lines is wrong, and the transcript will tell you which. The shared Code of conduct that applies to podiatrists expects records that are accurate, factual and objective, whoever typed them.

A worked example

The patient here is fictional. Margaret is 67 and has had type 2 diabetes for 14 years. Her GP has referred her under a chronic condition management plan. During the examination, the podiatrist says:

"Monofilament, left foot: felt at all sites. Right foot: not felt at the first and fifth metatarsal heads. Dorsalis pedis and posterior tibial pulses palpable on both feet, Doppler triphasic at both sites on both feet. Moderate hallux valgus on the right, callus under the second metatarsal head. Skin intact, no pre-ulcerative signs. So that's loss of protective sensation on the right with deformity: IWGDF risk 2, moderate. I've debrided the callus today. We'll see you again in three months."

From that, a template like the one above drafts:

• Neurological: 10 g monofilament. Left: felt at all sites tested. Right: not felt at 1st and 5th metatarsal heads.
• Vascular: DP and PT palpable bilaterally. Doppler triphasic DP and PT bilaterally.
• Skin and nails: Plantar callus, right 2nd metatarsal head. Skin intact; no pre-ulcerative signs.
• Musculoskeletal: Moderate hallux valgus, right.
• Risk: LOPS (right) with foot deformity. IWGDF risk 2 (moderate). Screen again every 3 to 6 months.
• Treatment and plan: Callus debrided, right. Education on daily foot checks and not walking barefoot; footwear to accommodate the hallux valgus. Review in 3 months.

Vibration isn't mentioned: the monofilament had already shown a loss, so it wasn't tested. That's the "omit anything not assessed" rule at work.

What goes in the report back to the GP?

In Australia, Medicare allied health services for chronic conditions now sit under a GP chronic condition management plan, which replaced GP management plans and team care arrangements from 1 July 2025. Plans made before then can be used until 30 June 2027. You must send the referring GP a written report after the first and last service, and more often if clinically needed. Services Australia says it should cover the investigations, tests and assessments carried out, the treatment provided, and recommendations for future management.

For a foot assessment, that means the neurological and vascular findings, the risk category and when you'll screen again, what you did today, and your recommendations, including anything you're asking of the GP. For Margaret, the key sentence might read: "Her screening today shows loss of protective sensation on the right with hallux valgus, placing her at moderate risk (IWGDF 2); I've debrided a plantar callus and will review her in three months."

For veterans, DVA's treatment cycle asks for an end of cycle report to the GP covering all the conditions you treated (DVA).

In Hanah, the referral letter drafts from the same session as the note, with sections for the reason for referral, the history, your findings and impression, the management started, and your requests and follow-up. You check it and send it as usual.

Is it different in the UK or New Zealand?

The documentation habits carry over. The risk systems don't. In the UK, NICE guideline NG19 has its own low, moderate and high categories, plus "active diabetic foot problem", with reassessment from annually at low risk to every 1 to 2 months at high risk. In New Zealand, follow the screening tool your local diabetes pathway uses.

Whichever system you use, name it in the note: NICE "moderate risk" and "IWGDF risk 2" aren't defined the same way.

FAQ

Do I need the patient's consent to record? Yes. Ahpra's guidance says an AI scribe generally needs the patient's informed consent, and that you should ideally note their response in the record.

What if I didn't test something? Leave it out, and make sure your template tells the scribe to leave it out too. A missing line is honest; an invented "normal" isn't.

How often should I screen again? Under the Australian guideline: annually at IWGDF risk 0, every 6 to 12 months at risk 1, every 3 to 6 months at risk 2, and every 1 to 3 months at risk 3.

Does Hanah send the GP report for me? No. Hanah drafts it from the consult, and nothing is filed or sent automatically. You review it and send it through your usual channel.

If you'd like to see this working on your own templates, see Hanah for podiatrists or book a demo below.