Patient satisfaction
and consultation efficiency,
using an AI-based
documentation tool
in endocrinology practice.
ByR. Novac1·V. Repede2
1 Clinica Dr. Novac, Bacău · Romania
2 AscultAI by ApolloIQ · Romania
01Aim
Research aim
To evaluate patient satisfaction and consultation efficiency
in standard endocrinology consultations compared with consultations
assisted by an AI-based documentation application.
✓ Finding 03Continuityimproved continuity and quality of care.
06Interpretation
AI-assisted transcription:
i
Reduces cognitive and administrative load.
ii
Allows better physician focus on the patient.
iii
Captures the full clinical conversation in real time.
07Conclusions
AI not only saves time — it recovers clinically relevant
information that would otherwise be lost.
AI-assisted transcription tools may:
✓Improve efficiency in endocrine practice.
✓Reduce information loss during consultations.
✓Support more patient-centred clinical care.
08Limitations
△
Small sample size (N=20).
△
Single-centre study.
△
Qualitative assessment of information capture.
△
Potential learning / adaptation effect between visits.
09Acknowledgement
With thanks to the patients of Clinica Dr. Novac who consented to
take part in the comparative evaluation, and to the AscultAI
engineering team at ApolloIQ for technical and methodological
support.
700 consultations.
5 months. One endocrinology practice.
A real-world clinical evaluation of Ascult AI in specialist endocrinology — measuring
the impact of AI-assisted documentation on clinician workflow, patient experience,
and health-system capacity.
ClinicianDr Elena Roxana NovacSenior Consultant Endocrinologist
SettingClinica Dr Novac, BacăuPrivate endocrinology clinic, Romania
Senior consultant endocrinologist with 16+ years of clinical experience and an
active international academic profile. Dr Novac led this independent evaluation from her
specialist endocrinology practice in Bacău.
2010 · Int. Congress of Endocrinology, Prague — 2 posters (thyroid carcinoma metastasis; Robinow & Simpson-Golabi-Behmel syndromes)
2011 · European Congress of Endocrinology, Copenhagen — adipocytokines & bone-mass acquisition
2013 · European Congress of Endocrinology, Florence — 3 posters (first- and co-author)
2013 · National Congress of Psycho-Neuro-Endocrinology — invited speaker: “Hyperparathyroidism — the disease with 1,000 faces”
2014 · XXII Congress of the Romanian Society of Endocrinology — 2 presentations
2014 · Postgraduate Course, Göteborg, Sweden
Research collaborations with Prof. Dr. Brănișteanu (UMF Iași). Research themes: bone mass, adipocytokines, hyperparathyroidism, rare thyroid pathology.
Methodology
A real-world prospective evaluation
Ascult AI was deployed in routine outpatient endocrinology practice over five consecutive months.
Every consultation documented using the platform was logged; clinician- and patient-reported
outcomes were captured continuously.
01
Setting
Specialist endocrinology outpatient clinic. Consultations covering thyroid disorders, diabetes, parathyroid disease, metabolic bone disease, and rare endocrine syndromes.
02
Intervention
Ascult AI captured each consultation in real time, generating a structured Romanian-language draft (letter, observation note, ultrasound report) for clinician validation before sign-off.
03
Comparator
Historical baseline: manual documentation averaging 15 minutes per consultation, completed during or after clinic and contributing to after-hours workload.
04
Outcomes captured
Time reclaimed per consultation; total patient throughput; documentation completeness; clinician-reported burden; patient-reported experience of eye-contact and engagement.
GDPR Art. 9 compliantSpecial-category health data processed under explicit consent
Zero audio retentionAudio processed & deleted in real time; no recordings stored
ISO 27001 alignedEncryption in transit and at rest; EU data residency
Clinician-validatedAll AI output is a draft requiring clinician review and sign-off
Results
What the data showed
700Consultations documented with Ascult AI over five months
15 min → 4 minMean documentation time per consultation (~73% reduction)
175 hrsClinician time reclaimed — equivalent to 20+ working days
+3/dayAdditional patients consulted per day, without extending hours
+30%Increase in patient throughput across the evaluation period
−90%Reduction in documentation errors and missing fields vs. manual baseline
“For the first time in years, I finished clinic without a backlog of letters waiting at home.
My patients noticed the difference too — the consultation felt like a conversation again,
not a typing exercise.”
— Dr Elena Roxana Novac, Clinician
Clinical Benefit
Benefit at three levels
AI-assisted documentation translates into measurable benefit for the patient in front of the
clinician, the clinic operating around them, and the wider health system demanding capacity.
For the patient
Restored eye contact. The clinician looks at the patient, not the keyboard, for the full consultation.
More complete notes. Nothing said in the room is lost to a distracted typist — summaries, plans, and safety-netting are captured in full.
Faster follow-up. Letters and referrals generated during the visit, not days later.
Equal voice. Non-native Romanian speakers and patients with complex histories benefit most when the clinician can focus on listening.
For the clinic
+30% throughput with no additional clinical hours worked.
2+ hours/day of after-clinic paperwork eliminated for the lead clinician.
Lower burnout risk. Documentation is the single biggest driver of clinician fatigue — automating it protects retention.
Cleaner data. Structured output integrates directly with the clinic’s records, with a 90% reduction in missing or ambiguous fields.
For the health system
Capacity, unlocked. An extra 3 patients per day per clinician compounds into thousands of additional specialist slots per year across a region.
Reduced waiting lists without needing to recruit new clinicians — at a time when specialist endocrinology capacity is under pressure Europe-wide.
Better, auditable records support safer referrals, cleaner research data, and more reliable outcomes tracking.
A template for scale. A single-clinic prospective result designed to be replicable across specialties and geographies.
Why this study matters
A proof point the literature has been waiting for.
The promise of ambient AI documentation is well-rehearsed in white papers. What the field has
lacked is prospective evidence from a real specialist clinic over a meaningful sample —
with an independent investigator, a conservative comparator, and honest outcome reporting.
Dr Novac’s 700-consultation dataset provides exactly that. It replaces a vendor claim
(“clinicians save time”) with a quantified, clinician-led observation (175
hours reclaimed, +30% throughput, −90% documentation errors) from a discipline
— endocrinology — where documentation complexity is high and the patient journey
is long.
The scientific contribution is twofold: it demonstrates clinical benefit (restored
patient contact, more complete notes, fewer errors) alongside operational benefit
(capacity, throughput, burnout reduction) — and it invites replication across other
specialties and settings.
Documentation time per consultation
Manual baseline
15 min
With Ascult AI
4 min
Mean per-consultation time across n=700 consultations. Manual baseline derived from pre-study clinic audit.
Monthly throughput
Month 1
Month 2
Month 3
Month 4
Month 5
Throughput stabilised at +30% vs. baseline once the clinician workflow had fully adapted.
Ascultă. Înțelege. Documentează.
Administrative documentation tool · Romania
Ascult AI is ApolloIQ’s Romanian ambient-documentation platform. It listens to the
clinician-patient consultation, transcribes and structures it in real time, and outputs a
validated draft — letter, observation sheet, ultrasound report — for the clinician
to review and sign off. It is positioned in Romania as an administrative support
tool: every output is a draft requiring clinician validation.
Real-time Romanian transcription
Native Romanian speech recognition with clinical-vocabulary tuning.
Specialty templates
Preconfigured for endocrinology, cardiology, dermatology, and more.
Browser-based, zero install
Works on any device — laptop, tablet, phone — with encrypted EU-hosted infrastructure.
Direct EHR export
PDF, print, or one-click export into the clinic’s record system.
In the United Kingdom, ApolloIQ operates the Merlin platform for primary care.
Our UK documentation product — ScribeCraft — is registered as a
Class I medical device, DCB 0129 compliant, and ISO 27001 aligned.
The same rigour informs the wider product suite.
DOCUMENTATION
ScribeCraft
Ambient consultation transcription. Class I medical device, DCB 0129 compliant, human-in-the-loop validation on every note.
ApolloIQ was founded in 2023 in the UK. Our team combines 20+ years of
combined clinical experience across primary care, secondary care and community pharmacy
with senior software engineering. We work every week alongside GPs, nurses, practice managers
and specialist clinicians across the UK and Romania.
The Novac study reframes ambient AI documentation as a clinical tool, not a convenience.
Seven hundred consultations, five months, one practice — and a measurable gain for the
patient, the clinician, and the system around them. This is the kind of evidence that deserves
replication.
Case Study 
