Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Si Hui Low has a total of 49 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Accuracy and impact of an AI safety system for missed specialist referrals in primary care

    by Wayne Han Lee, Lawrence Teck Meng Lam, Yiyang Liow, Ming Yi Lai, Binglong Lee, Si Hui Low, Justin Wee-Min Chong, Anandan Gerard Thiagarajah, Andrew Makmur, Kee Yuan Ngiam, et al.

  • Open Access

    Psychological Distress Among Primary Care Professionals in Singapore: Repeated-Measures Analyses of Two Independent Observation Windows (HADS and GHQ-12)

    by Sky Wei Chee Koh, Yixuan Chuah, Si Hui Low, Alicia Ying Ying Boo, Thomas Soo, Yii Jen Lew, Howard Bauchner, Lynette Mei Lim Goh

    Background: Healthcare professionals in Singapore’s primary care face sustained workload pressures and increasing complexity. Grounded in the Job Demands–Resources model, this study identifies demographic, clinical, and lifestyle factors associated with psychological distress among staff across two independent windows spanning the pre- and post-pandemic periods. Methods: The study was conducted among staff in National University Polyclinics (NUP) Singapore. De-identified health screening data included demographics, occupation, clinical measures, medical history, and lifestyle behaviours. Primary outcome was psychological distress, assessed across 2 distinct observation windows: the Hospital Anxiety and Depression Scale (HADS) from 2018–2022 (abnormal ≥ 8) and the General Health Questionnaire-12 (GHQ-12) from 2023–2025 (abnormal ≥ 2). Multivariate mixed-effects logistic regression identified factors associated with psychological distress while accounting for within-person clustering. Results: We analysed 3307 screening visits from 1260 staff. Within the HADS observation window (2018–2022), abnormal scores were significantly higher in 2022 compared with 2018 (OR 2.28, 95% CI 1.64–3.17). Within the GHQ-12 observation window (2023–2025), a non-significant decline in abnormal scores was observed. Older age (2018–2022: OR 0.967, 95% CI 0.948–0.987; 2023–2025: OR 0.955, 95% CI 0.928–0.983) and regular exercise (2018–2022: OR 0.724, 95% CI 0.529–0.991; 2023–2025: OR 0.357, 95% CI 0.231–0.551) were consistently associated with lower odds of distress. Medical staff had significantly lower odds of distress compared to ancillary staff in 2018–2022 (OR 0.391, 95% CI 0.207–0.739). Regular alcohol use was associated with higher distress in 2018–2022 (OR 1.56, 95% CI 1.05–2.31), while active smoking (OR 5.79, 95% CI 1.75–19.10) and family history of chronic disease (OR 2.22, 95% CI 1.28–3.87) were significant in 2023–2025. Conclusions: Older age and regular exercise were consistently associated with lower odds of psychological distress across both windows, while ancillary staff and alcohol use (HADS window), and smoking and family history of chronic disease (GHQ-12 window) were period-specific determinants. These findings support the implementation of proactive, targeted wellness initiatives alongside the routine mental health surveillance among primary care staff.

  • Open Access

    Diagnostic labelling, antibiotic appropriateness and watch use for upper respiratory infections in Singapore primary care: A multicentre retrospective study

    by Sky Wei Chee Koh, Shipra Lather, Si Hui Low, Jun Cong Goh, Pey Gein Wong, Cynthia Kum Wai Cheong, Justin Wee-Min Chong, Meena Sundram, Li Yang Hsu


  • Open Access

    Factors and determinants of primary care to tertiary care referrals in Singapore: A multi-centre analysis using artificial intelligence-powered large language models

    by Sky Wei Chee Koh, Jasper Yi Xuan Huang, Joelle Lam, Si Hui Low, Jun Cong Goh, Ge Ji, Xin Jin, Howard Bauchner, Yii Jen Lew

    Background: As Singapore adopts a population health approach under Healthier Singapore (Healthier SG), optimizing healthcare resources is crucial. We examined referral reasons (using large language models [LLM]), wait times, and analyse factors affecting referrals from primary to tertiary care. Methods: In 2023, 1,063,646 patient visits from seven primary care clinics in Singapore were analysed. Patient demographics, clinic, physician characteristics, referral volumes and wait times were extracted. LLM Claude 3.5 Sonnet was utilized to identify and classify top referral reasons within the most frequently referred specialties based on referral notes. Chi-square tests identified differences in referral rates among categorical variables, while a generalised linear model (GLM) with an identity link (normal distribution) determined factors influencing referrals by physicians. Findings: Around 1 in 5 visits resulted in a referral (n = 210,839, 19.8%), achieving 76.0% attendance rate. Referrals peaked among patients aged 60–70 years. Male (Odds ratio [OR] 0.88, 95% Confidence interval [CI] 0.87–0.89) and Malay (OR 0.71, 95% CI 0.70–0.72, compared with Chinese) patients were less likely to be referred. Significant variations were observed among clinics (p < 0.001). Ophthalmology (11.1%), orthopaedic surgery (10.3%), and emergency (10.0%) were the most referred specialties, with blurred vision (n = 7,461), abnormal diabetic retinopathy screening (n = 5,266) and pregnancy and antenatal care (n = 3,959) being the top referral reasons. 51.5% were routine referrals. Wait time averaged 52.7 days with 48.9% meeting targets, with long wait times for Gastroenterology & Hepatology, and Endocrinology. On average, each additional year of physician experience was associated with a reduction of 4.45 referrals per physician (95% CI: 1.40–7.58, p = 0.005). Interpretation: Our study highlighted disparities in referrals rates, patterns, and wait times. Continuing education and support for primary care is paramount. Resource allocation should be tailored to meet the population needs, with further research needed to ensure timely and appropriate referrals.

  • Open Access

    Classifying Patient Complaints Using Artificial Intelligence–Powered Large Language Models: Cross-Sectional Study

    by Sky Wei Chee Koh, Eunice Rui Ning Wong, John Chong Min Tan, Stephanie C C van der Lubbe, Jun Cong Goh, Ethan Sheng Yong Ching, Ian Wen Yih Chia, Si Hui Low, Ping Young Ang, Queenie Quek, et al.

    Background: Patient complaints provide valuable insights into the performance of health care systems, highlighting potential risks not apparent to staff. Patient complaints can drive systemic changes that enhance patient safety. However, manual categorization and analysis pose a huge logistical challenge, hindering the ability to harness the potential of these data. Objective: This study aims to evaluate the accuracy of artificial intelligence (AI)–powered categorization of patient complaints in primary care based on the Healthcare Complaint Analysis Tool (HCAT) General Practice (GP) taxonomy and assess the importance of advanced large language models (LLMs) in complaint categorization. Methods: This cross-sectional study analyzed 1816 anonymous patient complaints from 7 public primary care clinics in Singapore. Complaints were first coded by trained human coders using the HCAT (GP) taxonomy through a rigorous process involving independent assessment and consensus discussions. LLMs (GPT-3.5 turbo, GPT-4o mini, and Claude 3.5 Sonnet) were used to validate manual classification. Claude 3.5 Sonnet was further used to identify complaint themes. LLM classifications were assessed for accuracy and consistency with human coding using accuracy and F 1 -score. Cohen κ and McNemar test evaluated AI-human agreement and compared AI models’ concordance, respectively. Results: The majority of complaints fell under the HCAT (GP) domain of management (1079/1816, 59.4%), specifically relating to institutional processes (830/1816, 45.7%). Most complaints were of medium severity (994/1816, 54.7%), occurred within the practice (627/1816, 34.5%), and resulted in minimal harm (75.4%). LLMs achieved moderate to good accuracy (58.4%‐95.5%) in HCAT (GP) field classifications, with GPT-4o mini generally outperforming GPT-3.5 turbo, except in severity classification. All 3 LLMs demonstrated moderate concordance rates (average 61.9%‐68.8%) in complaints classification with varying levels of agreement (κ=0.114‐0.623). GPT-4o mini and Claude 3.5 significantly outperformed GPT-3.5 turbo in several fields ( P <.05), such as domain and stage of care classification. Thematic analysis using Claude 3.5 identified long wait times (393/1816, 21.6%), staff attitudes (287/1816, 15.8%), and appointment booking issues (191/1816, 10.5%) as the top concerns, which accounted for nearly half of all complaints. Conclusions: Our study highlighted the potential of LLMs in classifying patient complaints in primary care using HCAT (GP) taxonomy. While GPT-4o and Claude 3.5 demonstrated promising results, further fine-tuning and model training are required to improve accuracy. Integrating AI into complaint analysis can facilitate proactive identification of systemic issues, ultimately enhancing quality improvement and patient safety. By leveraging LLMs, health care organizations can prioritize complaints and escalate high-risk issues more effectively. Theoretically, this could lead to improved patient care and experience; further research is needed to confirm this potential benefit.


  • Open Access

    Increase in Antibiotic Utilisation in Primary Care Post COVID-19 Pandemic

    by Sky Wei Chee Koh, Si Hui Low, Jun Cong Goh, Li Yang Hsu

    Introduction: The COVID-19 pandemic has disrupted antibiotic use; easing public health measures may alter infection presentations and antibiotic prescribing in primary care. The study investigated post-pandemic antibiotic utilisation trends in primary care. Methods: A multi-centre, retrospective cohort study was conducted across seven public primary care clinics in Western Singapore, which included all patients prescribed oral antibiotics between 2022 and 2023. Descriptive statistics were used to visualise the prevalence and conditions of the prescribed antibiotics. Antibiotic quality was evaluated using the WHO’s AWaRe (access, watch, reserve) classification. Antibiotic use was quantified using the number of items dispensed per 1000 inhabitants (NTI), defined daily doses (DDD) per 1000 inhabitants per day (DID), and DDD per 100 visits. Segmented regression analysis was applied to monthly prescriptions to assess the utilisation trends. Results: Antibiotic prescription rates increased significantly, from 3.5% in 2022 to 4.0% in 2023 (p = 0.001), with a 9.5% relative increase (38,920 prescriptions for 1,112,574 visits to 42,613 prescriptions for 1,063,646 visits). Respiratory conditions drove the increase in antibiotics use, with a 68.3% rise in prescriptions, with upper respiratory tract infections being the most common diagnosis for antibiotic prescriptions (n = 9296 prescriptions in 2023), with a steady monthly upward trend. Access group antibiotics accounted for >90% of prescriptions. The most antibiotics were prescribed for acne, with 36,304 DDD per 100 visits in 2023. Both NTI and DID significantly increased in 2022, largely contributed by a >100% increase in Watch group antibiotic use. Total antibiotic NTI dipped slightly in 2023, with a stable trend in both NTI and DID for all antibiotics. Conclusions: The post-COVID-19 pandemic surge in the antibiotic prescription rate for respiratory conditions and Watch group antibiotic use highlight the need for targeted stewardship interventions. Optimising acne treatment and diagnosis coding are key strategies to further reduce unnecessary prescriptions.


  • Open Access

    Prescribing Antibiotics for Children with Acute Conditions in Public Primary Care Clinics in Singapore: A Retrospective Cohort Database Study

    by Vivien Min Er Lee, Si Hui Low, Sky Wei Chee Koh, Anna Szuecs, Victor Weng Keong Loh, Meena Sundram, José M. Valderas, Li Yang Hsu

    Data on primary care antibiotic prescription practices for children in Singapore, which are essential for health care policy, are lacking. We aimed to address this gap and to benchmark prescription practices against international standards. A retrospective cohort database study on antibiotic prescriptions for children (aged < 18 years) who visited six public primary care clinics in Singapore between 2018 and 2021 was conducted. Data were categorised according to the World Health Organization’s Access, Watch, Reserve (WHO AWaRe) classification. Quality indicators from the European Surveillance of Antimicrobial Consumption Network (ESAC-Net) and the National Institute for Health and Care Excellence (NICE) guidelines were used as a measure of appropriateness of antibiotic prescribing at the individual and overall patient level. In 831,669 polyclinic visits by children between 2018 and 2021, there was a significant reduction in mean antibiotics prescribed per month during pandemic years (2020–2021) compared to pre-pandemic (2018–2019) (MD 458.3, 95% CI 365.9–550.7). Most prescriptions (95.8%) for acute conditions fell within the WHO AWaRe “Access” group. Antibiotic prescription significantly exceeded (55.2%) the relevant quality indicator for otitis media (0–20%). The proportion of children receiving appropriate antibiotics for acute respiratory infections (n = 4506, 51.3%) and otitis media (n = 174, 49.4%) was low compared to the quality indicator (80–100%). There is a need to develop local evidence-based primary care antibiotic guidelines, as well as to support the development of stewardship programmes.


  • Open Access

    Prescribing Antibiotics in Public Primary Care Clinics in Singapore: A Retrospective Cohort Study

    by Sky Wei Chee Koh, Vivien Min Er Lee, Si Hui Low, Wei Zhi Tan, José María Valderas, Victor Weng Keong Loh, Meena Sundram, Li Yang Hsu

    Background: Antibiotic prescription practices in primary care in Singapore have received little scholarly attention. In this study, we ascertained prescription prevalence and identified care gaps and predisposing factors. Methods: A retrospective study was conducted on adults (>21 years old) at six public primary care clinics in Singapore. Prescriptions >14 days were excluded. Descriptive statistics were used to showcase the prevalence data. We used chi-square and logistic regression analyses to identify the factors affecting care gaps. Results: A total of 141,944 (4.33%) oral and 108,357 (3.31%) topical antibiotics were prescribed for 3,278,562 visits from 2018 to 2021. There was a significant reduction in prescriptions (p < 0.01) before and after the pandemic, which was attributed to the 84% reduction in prescriptions for respiratory conditions. In 2020 to 2021, oral antibiotics were most prescribed for skin (37.7%), genitourinary (20.2%), and respiratory conditions (10.8%). Antibiotic use in the “Access” group (WHO AWaRe classification) improved from 85.6% (2018) to 92.1% (2021). Areas of improvement included a lack of documentation of reasons for antibiotic use, as well as inappropriate antibiotic prescription for skin conditions. Conclusion: There was a marked reduction in antibiotic prescriptions associated with the onset of the COVID-19 pandemic. Further studies could address the gaps identified here and evaluate private-sector primary care to inform antibiotic guidelines and the local development of stewardship programs.