Development of a nomogram for triage admission prediction: A pilot study in Hong Kong

Abstract Background Emergency department (ED) overcrowding can have detrimental consequences for both patients and staff. Actively predicting admissions at triage is a potential solution to alleviate ED overcrowding. This study aims to develop a simple nomogram to estimate the probability of admission at the time of ED triage in Hong Kong. Methods A retrospective cross‐sectional study was conducted on consecutive 1‐week ED visits at a regional hospital in Hong Kong. A nomogram based on the logistic regression analysis of potential predictors collected at ED triage was developed to estimate the probability of admission. The nomogram's performance was preliminarily assessed using receiver operating characteristic curve analysis and goodness‐of‐fit tests. Results A total of 1457 unscheduled adult (≥18 years old) ED visits from 1 January to 7 January 2024 were included for analysis. The final model revealed seven variables showing clinically and statistically significant associations with admission: age, comorbidities, living in residential care home, triage category, arrival by ambulance, referral status, and national early warning score. The nomogram demonstrated an excellent admission/discharge discrimination power with an area under the curve of 0.886 (95% CI 0.869–0.903). An easily accessible online calculator was developed for the nomogram. Additionally, a higher total score calculated by the nomogram correlated with prolonged length of stay and increased in‐hospital mortality following admission. Conclusion We developed a simple nomogram that accurately estimates the probability of admission at the time of ED triage. Further validation, refinement, and implementation of this tool, in conjunction with available strategies, hold the potential to alleviate ED overcrowding in Hong Kong.