Accuracy of ganga hospital open injury score in predicting amputation of severe open lower limb injuries among adults seen at Mulago Hospital

Abstract

Background: With the recent advent of more motorization in Uganda, severe open lower limb injuries are pretty common. Hence, a frequent ethical and medical-legal dilemma in decision-making by frontline clinicians often occurs on whether to amputate a limb to prevent mortality. There are a variety of injury scores to help guide decisions; unfortunately, most fall short of accuracy. There is a need for a precise score to assist clinicians in providing a tiebreaker to salvage or amputate with the most accuracy. Objective: To assess the accuracy of GHOISS in the clinical prediction of amputation among adult patients with severe open lower limb injuries at Mulago National Referral Hospital. Methods: A three-month hospital-based prospective cohort study was done at Mulago National Referral Hospital’s accident and emergency ward among 92 adult patients with severe open lower limb injuries using the Buderer sample technique. An interviewer-guided questionnaire was used to collect participants' socio-demographic and clinical characteristics. The patients with severe open lower limb injuries were evaluated and scored using GHOISS and MESS scores. Participants were thereafter followed up for up to 5 days to see if any subsequently experienced a delayed amputation. Results: The study included 92 participants, predominantly male (85.87%), with a mean age of 34 years (±11.47). The most common cause of injury was road traffic accidents (90.22%), and the tibia was the most frequently injured site (80.43%). Among the study population, 2.17% underwent primary amputation, 7.61% had delayed amputation, and 90.22% had limb salvage. The predictive accuracy analysis of GHOISS compared to MESS demonstrated superior performance for GHOISS. For primary amputation, both scores had AUC = 1.0, but GHOISS had slightly higher specificity (91.11% vs. 87.78%). For delayed amputation, GHOISS showed AUC = 0.95, sensitivity = 71.43%, specificity = 94.12%, outperforming MESS (AUC = 0.72, sensitivity = 42.86%, specificity = 88.24%). Conclusion: The findings indicate that GHOISS is a highly accurate tool for predicting primary and delayed amputations, significantly outperforming MESS in sensitivity and specificity.