Tuen Mun Hospital is a public hospital, with a 24-hour accident and emergency department, in Tuen Mun, New Territories, Hong Kong. Its construction began in 1979, and it was inaugurated on 8 March 1990. It is governed by the Hospital Authority, and is designated under New Territories West Cluster..
Copeptin is an emerging biomarker for polyuria-polydipsia syndrome. Postoperative copeptin levels may predict acquired arginine vasopressin deficiency (AVP-D). This prospective study evaluated the diagnostic performance of serum copeptin for predicting postoperative AVP-D after transsphenoidal surgery (TSS). Copeptin was measured preoperatively, immediately after surgery (periextubation, T0), upon first ward arrival (T1), daily until discharge, and at week 2. Delta copeptin was calculated as the percentage change from T0 to T1. Of 53 patients, 23 (43.4
Background Surgical outcomes monitoring has evolved from periodic retrospective audits to risk-adjusted benchmarking systems that support governance, quality improvement (QI) and research. The Surgical Outcomes Monitoring & Improvement Programme (SOMIP), launched in 2008 across all public hospitals under the Hospital Authority in Hong Kong, provides annual risk-adjusted outcome benchmarking and an infrastructure for local improvement initiatives.Methods This meeting report synthesizes the content of three presentations delivered at the RCSEd/CSHK Conjoint Scientific Congress session on 30 November 2025. Source materials comprised the speakers' slide sets and the session video-recording (excluding the Q&A session).Results Speakers reviewed the professional and historical foundations of outcomes monitoring from Codman's 'end result' philosophy to contemporary risk-adjusted surgical report cards, emphasizing the Donabedian structure-process-outcome framework. SOMIP was described as a risk-adjusted benchmarking programme adapted from the ACS NSQIP model, providing department-level signals to trigger system diagnosis and QI. Potential unwanted effects-league table thinking, morale harm, risk aversion and gaming-were highlighted, with the session advocating that outcome monitoring becomes an ally when embedded in a just culture and paired with actionable process improvement. A cluster-level case study from the New Territories West Cluster (NTWC) illustrated how early 'poor outlier' signals in 30-day mortality catalysed multi-layered interventions, including strengthened senior presence, structured mortality review, multidisciplinary quality circles, ICU liaison pathways, sepsis control and pneumonia prevention workstreams, and workflow integration of a risk calculator to support perioperative triage and optimization.Conclusions The session positioned SOMIP as a mature surgical outcome research and audit platform that can strengthen surgical governance and drive improvement where data credibility, fair risk adjustment and supportive implementation are prioritized. Future opportunities include deeper integration into clinical management systems and the development of surgical data science-enabled decision support, coupled with safeguards to mitigate unintended consequences.
Background: Most of the public hospitals in Hong Kong provide a regular colorectal endoscopic submucosal dissection (ESD) service. The current retrospective study aims to review the long- and short-term outcomes of these services, so as to identify areas for improvement. Method and Results: From January 2017 to March 2020, 634 lesions in 623 patients were removed by an ESD technique in seven endoscopic units. The mean lesion size was 31 mm (SD 13 mm, range 10–95 mm), and the mean procedure time was 121 min (SD 67 min). En bloc resection and R0 resection could be achieved in 91.3% and 79.3% of the lesions, respectively. The intra-procedure perforation rate was 12.3%. The delayed bleeding rate was 2.1%, and the delayed perforation rate was 0.8%. Only 0.9% (6/634) of the procedures needed emergency surgical salvage due to complications. Most of the lesions were adenomas (564/634), and 55 of them were adenocarcinomas. The cumulative local recurrence rate was 4.0% at a mean follow-up of 34 months. In multivariate analysis, longer procedure time, submucosal fibrosis, hybrid ESD and piecemeal removal were associated with intra-procedure perforation. Risk factors for failed en bloc resection included non-granular and polypoid morphology, colonic location, longer procedure time and low centre volume. Malignant pathology without salvage surgery was the only independent risk factor for local recurrence. Conclusions: Colorectal ESD has been carried out in Hong Kong with acceptable short-and long-term outcomes despite the technique still being in the learning phase in some centres. Potential areas for improvement should include targeted training to speed up the procedure and enable better handling of difficult cases, aiming to decrease the perforation rate and local recurrence.
This study aims to develop a machine learning (ML)-based pipeline to predict intensive care unit (ICU) mortality and length of stay (LOS). A dataset including 140,904 ICU admissions was collected from 15 public hospitals in Hong Kong over an 11-year period. The proposed pipeline deployed a suite of ML models to predict mortality and LOS. The performance of ML models was compared with the Acute Physiology and Chronic Health Evaluation (APACHE) systems on the collected dataset using five-fold cross-validation. Among all involved models, the Gradient Boosting with Categorical Features (CatBoost) achieved the highest area under the receiver operating characteristic curve (AUROC) of 0.9070 as well as the lowest Brier score of 0.0827 for mortality prediction and the lowest Mean Absolute Error (MAE) of 2.6364 for LOS prediction. The SHapley Additive exPlanations (SHAP) analysis conducted on CatBoost revealed that age, Glasgow Coma Scale (GCS) and urine output were the top-three important features for mortality prediction, whereas the top-three important features for LOS prediction were creatinine level, and the indicators for whether the lowest and highest respiratory rates were ventilator-measured. We further performed temporal validation and an in-depth analysis of CatBoost’s predictive performance across subsets grouped by age and hospital. Our results demonstrate that the proposed pipeline mitigates the overestimation of mortality predictions from APACHE systems in Hong Kong. Besides, the proposed predictive ML-based pipeline offers a transferable framework for researchers to develop models tailored to their local medical environments.
