AOAriel Yuhan Ong

Ariel Yuhan Ong

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Ariel Yuhan Ong has a total of 296 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    The clinician–artificial intelligence scientist: a proposed career pathway in medicine

    by Yilan Wu, Ariel Yuhan Ong, Haichao Chen, Gabriel Dawei Yang, Yih Chung Tham, Pearse A Keane, Tien Yin Wong

  • Open Access

    Association of Reticular Pseudodrusen Burden and Dark Adaptation in Nonadvanced Age-Related Macular Degeneration

    by Ariel Yuhan Ong, Robbert R. Struvyen, David A. Merle, Kevin Mendez, Helia Ashourizadeh, Souvick Mukherjee, Afonso Lima-Cabrita, Jooyoung Chang, Dylan Wu, Shrinivas Pundlik, et al.

  • Open Access

    Developing a scalable pipeline for data extraction from clinical letters through resource-efficient prompt engineering

    by Ariel Yuhan Ong, Quang Nguyen, Ishani Barai, Justin Engelmann, Fares Antaki, Mertcan Sevgi, David A. Merle, Lie Ju, Eliot Dow, Yukun Zhou, et al.

    Free-text clinical records represent an untapped wealth of data for secondary use, but realising their potential is limited by resource demands necessary for accurate information extraction at scale. We introduce a scalable, resource-efficient, and high-performance information extraction pipeline that leverages large language models (LLMs) to address these challenges. Our pipeline was developed and tested using real-world dual specialist-annotated ophthalmic clinical letters, and achieved strong performance with a proprietary model in development, yielding a maximum micro-averaged F1 score of 0.954 (95% CI 0.941–0.967) for diagnosis across nine conditions through iterative prompt refinement alone, also demonstrating strong generalisability (micro-F1 0.945–0.980) in temporal validation. This approach was extended to other models in the same family and 17 LLMs from seven open-weight LLM families. Beyond performance, we develop a multi-dimensional assessment for deployment in data extraction tasks, including an error taxonomy and Pareto frontier analyses to systematically map the operational trade-offs across different LLM configurations. A robust approach to operationalisation in real-world workflows at scale may help lay the foundation for next-generation data pipelines that accelerate scientific discovery and power continuous learning health systems.

  • Open Access

    Multilingual Voice AI for Postoperative Cataract Follow-Up in Turkish Speaking Patients in the United Kingdom: Patient and Public Involvement Focus Group Study

    by Mertcan Sevgi, Ariel Yuhan Ong, Katie Lean, Sian Rees, Ernest Lim, David Adrian Merle, Alexander C Day, Badrul Hussain, Pearse A Keane, Aisling Higham, et al.

    Background: Conversational voice AI assistants can automate postoperative follow-up calls in high-volume, low-complexity pathways such as cataract surgery but may widen health inequalities if language access and inclusive design are not built in. This patient and public involvement focus group was conducted to inform the Turkish-language adaptation of Dora ahead of a forthcoming multilingual clinical trial at Moorfields Eye Hospital. Objective: This study aims to inform the Turkish-language adaptation of Dora by gathering input from Turkish speaking community contributors about their experiences with UK ophthalmic care, language-related barriers, and design requirements for an equitable voice AI. Methods: We conducted a 1-time, 2-hour patient and public involvement focus group with 7 Turkish speaking adults recruited via the Derman community charity. The session ran in 2 phases: contributors first discussed their experiences with UK ophthalmic care, then evaluated a prerecorded Turkish-language telephone call from a voice AI to a Turkish speaking volunteer. The session was delivered bilingually, recorded with consent, and synthesized using an approach informed by the principles of reflexive thematic analysis. The voice AI uses automatic speech recognition and neural text-to-speech, with a large language model–based dialog manager for open-ended conversation within a postoperative review protocol. Results: Contributors described how pathway delays and limited language support shape their care, including reliance on family members for translation and concerns about privacy and autonomy. A language-concordant voice AI was conditionally acceptable for standardized postoperative follow-up, provided specific safeguards were met. Priorities included advance notice of calls, caller verification, privacy assurances, a clear standard Turkish accent at a slower pace, tolerance for regional dialects, interpersonal warmth, interactivity, accessibility for low vision and low literacy, and clinician escalation for complex issues. These priorities were synthesized into a 10-point checklist: preparation, verification, confidentiality, clarity and pace, voice, empathy, interactivity, dialect handling, accessibility, and efficiency. Conclusions: For patients facing language barriers, conversational voice AI may complement existing services when implemented with clear verification, privacy protections, and a defined scope under clinician oversight. The 10-item checklist will guide the Turkish-language adaptation of Dora and will be tested alongside similar consultations with other language communities in the forthcoming multilingual cataract follow-up trial.

  • Open Access

    How well do we write for patients? Longitudinal analysis of the readability and complexity of 4.6 million ophthalmic clinical letters

    by Ariel Yuhan Ong, Christine A Kiire, Siegfried K Wagner, Quang Nguyen, Peter B M Thomas, Pearse A Keane

    Objectives: Effective clinician–patient communication is a fundamental pillar of high-quality clinical care and can influence patient engagement, treatment adherence and clinical outcomes. However, the extent to which clinical letters meet recommended readability standards remains largely unquantified. We aimed to benchmark this at scale in an ophthalmology setting—the highest-volume outpatient specialty in the United Kingdom (UK) healthcare service. Methods: Over 4.6 million outpatient clinic letters written for 804 986 patients across 17 subspecialty services were included. Standard readability scores (including Flesch Reading Ease (FRE), SMOG Index, Automated Readability Index) were calculated for letters written between 2013 and 2025 at Moorfields Eye Hospital Foundation Trust, UK. Complex linguistic features such as lexical density and type-token ratio were also evaluated, and rule-based measures were applied to identify letters addressed directly to patients. Using the FRE as the primary outcome, readability was compared across subspecialties, and temporal trends were evaluated. Results: The median FRE was 51.1 (IQR 41.9–58.5), indicating college-level readability which was substantially poorer than recommended for patient-facing materials. Overall, 95.3% of letters were less readable than recommended, and these patterns extended across all readability metrics evaluated. Readability varied across subspecialty services (p<0.001), with limited association with sociodemographic covariates such as age and ethnicity. There was a significant decline in FRE over the study period (−0.64 points/annum, p<0.001). Rate of change varied across subspecialty services and was more pronounced in some such as neuro-ophthalmology (−1.24 points/annum, p0.05). The use of second person terms was very low (median 0 per 100 words, IQR 0.0–1.1). There was a small annual increase in the likelihood of letters being addressed directly to patients (OR 1.123, 95% CI 1.121 to 1.25), although the proportion remained low overall (2.25%). Discussion: Clinical correspondence across ophthalmology services remained well above recommended readability thresholds and has become more complex over time, with substantial variability across subspecialty services. These findings will provide an empirical basis for local quality improvement initiatives. We hope that the methodological framework will also serve as a valuable resource for other specialties. Conclusion: Our findings highlight a persistent gap between policy recommendations and real-world practice, underscoring the need for targeted interventions and innovative approaches to support clearer patient-centred communication to optimise patient outcomes.

  • Open Access

    Dropless after cataract surgery (DACS): concepts, evidence and clinical potential

    by Johannes Birtel, Ariel Yuhan Ong, Karolin Deichl, Guy Mole, Sher A. Aslam, Peter Charbel Issa

    Cataract surgery is one of the most commonly performed surgeries worldwide, and is traditionally followed by a postoperative regimen of eyedrops to reduce the risk of infection and inflammation. However, variable adherence to treatment and the burden of administering drops pose significant challenges. “Dropless after cataract surgery” (DACS) has emerged as an approach where intraoperative delivery may eliminate the need for postoperative drops. DACS offers a simplified alternative to eyedrop therapy, with the potential to improve patient-related quality of life, lower healthcare costs, and promote sustainable practices by reducing medical waste. Multiple studies over the past few decades have demonstrated the effectiveness of intraoperative intracameral antibiotics alone in reducing endophthalmitis rates, thus eliminating the need for topical antibiotic drops. Sustained-release steroid formulations have demonstrated reliable control of postoperative inflammation, similar to those achieved with standard topical regimens, but their adoption remains limited. Realising the full benefits of DACS will require systemic barriers to be addressed, and dropless regimens should be considered within a holistic value-based healthcare approach. As the global burden of cataract surgery continues to grow, embracing dropless techniques may represent a forward-looking step in modernising postoperative management to optimise patient-centred care.


  • Open Access

    Reconciling how clinical reasoning is learned in the age of artificial intelligence

    by Ariel Yuhan Ong, Margaret Sui, Kyra L. Rosen, Joseph C. Kvedar

    Two recent studies examine the role of AI in supporting medical students’ learning – a longitudinal study of self-reported outcomes from supervised real-world AI use, and a simulated experiment comparing the benefits of AI explanations against the risks of plausible misinformation. Collectively, their findings demonstrate that the effects of AI may depend on the individual and educational environment, which has implications for how we design medical curricula and define clinical competence.


  • Open Access

    Rerouting Eye Care: How AI and Telemedicine Are Reshaping Ophthalmology Patient Journeys in the United Kingdom and Germany

    by Broder Poschkamp, Ariel Yuhan Ong, Justin Engelmann, Michael Grün, Andreas Stahl, Pearse Andrew Keane, David Adrian Merle

    Artificial intelligence (AI) as a medical device is now progressively entering routine ophthalmic care, yet its impact is still mostly evaluated in terms of diagnostic performance rather than how it reshapes patient care pathways. This viewpoint argues that careful pathway design is crucial to implementing AI in ophthalmology so that it translates into real, practical benefits for patients and services. We propose a framework that classifies AI- and telemedicine-enabled eye care pathways by the initial point of contact and the role of AI: from direct human assessment to grader- or telemedicine-based use to AI-first contact and fully autonomous AI gatekeeping. We apply this framework to 2 contrasting health systems, the United Kingdom and Germany, focusing on common retinal diseases and cardiovascular risk assessment from retinal images. In the United Kingdom, AI is being introduced mainly as a modular upgrade to standardized programs layered onto a gatekeeper model centered on community optometrists and general practitioners. In Germany, direct access to office-based ophthalmologists, opportunistic screening, and commercial retail offerings are producing more fragmented clinician- and market-driven AI adoption. Our comparison shows that the same AI technologies can generate very different patient journeys across health systems. Therefore, their value depends not only on diagnostic performance but also on intentional pathway design, including clear escalation rules and structured information transfer. This includes appropriate task shifting from ophthalmologists to trained staff or new roles, such as image grading, tele-ophthalmology triage, or pathway management. Without such design, AI risks duplicate testing, incomplete or poorly coordinated referrals, and further fragmentation of care.


  • Open Access

    Rapid Access Macular Screening and Evaluation: A Datasheet for a Dataset to Diagnose and Triage Macular Diseases

    by Ariel Yuhan Ong, Reena Chopra, Henry David Jeffry Hogg, Alastair K. Denniston, Pearse A. Keane

  • Open Access

    Co-intelligence: a proposal for human–artificial intelligence collaboration for large language models in medical research

    by Ariel Yuhan Ong, David A Merle, Nigam H Shah, Yih Chung Tham, Tien Yin Wong, Pearse A Keane