Improving Appendicitis Prediction in Children Using the Paediatric Appendicitis Score (PAS): A Three-Year Retrospective Study

Introduction Acute appendicitis is a leading surgical emergency in children, yet diagnosis remains challenging. The Paediatric Appendicitis Score (PAS) was developed to standardise risk stratification using clinical and laboratory indicators. While its diagnostic accuracy has been validated in controlled settings, its real-world application remains variable. This paper aimed to evaluate the impact of implementing PAS into routine clinical assessment at an NHS district-general hospital. The objectives were to reduce negative appendicectomy rates and improve diagnostic accuracy. Materials and methods A two-cycle retrospective audit was conducted at Southend University Hospital. Cycle 1 (August 2021-August 2024) included 156 paediatric patients who underwent appendicectomy. In August 2024, an educational intervention was introduced. It involved departmental teaching and dissemination of a quick-reference guide aligned with the Getting It Right First Time (GIRFT) abdominal pain pathway. Cycle 2 (August 2024-March 2025) included 34 cases. Patients under 18 years undergoing appendicectomy were included. The data collected included demographics, PAS scores, and histological outcomes. Diagnostic performance of PAS was evaluated by plotting the PAS value against the final histological outcome, using sensitivity, specificity, predictive values, likelihood ratios, and area under the receiver operating characteristic curve (AUC-ROC). Results A total of 190 patients were analysed. The negative appendicectomy rate fell from 29.5% (Cycle 1) to 17.6% (Cycle 2) following the educational intervention, although this was not statistically significant (p=0.16). In the combined cohort, PAS ≥7 achieved 87% specificity and positive predictive value (PPV) of 0.89, but only 40% sensitivity. PAS ≥4 maximised sensitivity (96%) but reduced specificity (31%). Mean PAS was significantly higher in confirmed cases (6.18 vs 4.69; p < 0.001). The AUC-ROC was 0.73. Discussion Across the full cohort, PAS showed expected trade-offs: at a high threshold (≥7), the tool provided good specificity (87%) but reduced sensitivity (40%), whereas a low threshold (≥4) delivered high sensitivity (96%) but poor specificity (31%). These findings are lower than those reported in the initial derivation cohort but broadly consistent with subsequent external validation studies. The score's modest AUC-ROC (0.73) underlines the need for adjuncts such as imaging in equivocal cases. Post-intervention trends suggest improved adherence to PAS-based risk stratification and reduced avoidable surgeries, although small sample size possibly led to poor statistical power. Conclusions Integrating PAS into routine assessment improved risk stratification and reduced the negative appendicectomy rate, although without statistical significance. High scores (≥7) remained highly specific, low scores (≤3) safely ruled out disease, and mid-range scores benefited from adjuncts such as imaging or serial review. Continued use within a structured pathway is advisable, but larger multicentre studies, and the inclusion of imaging adjuncts, are recommended to refine the score and therefore improve diagnostic accuracy and reduce negative appendicectomy rates.