Nottingham Hip Fracture Score Versus Surgical Outcome Risk Tool in Predicting 30-Day Mortality in Hip Fracture Patients in a District General Hospital in the United Kingdom.

Introduction and aims Neck of femur fractures (NOFFs) represent high costs to the National Health Service and the elderly population's quality of life. It is important to recognize vulnerable patients for early optimisation. This is a retrospective study of 157 NOFF patients to assess the 30-day, 60-day, six-month, and one-year mortality rates. The actual 30-day mortality was compared against the Nottingham Hip Fracture Score (NHFS) predicted 30-day mortality and the Surgical Outcome Risk Tool (SORT) version 2.0 30-day mortality. Methods This is a retrospective study conducted at York District General Hospital in the United Kingdom. The computerised patient data (CPD) records were accessed to find NOFF patients between 1/1/24 and 1/9/24 over a period of eight months. This allowed us to assess the 30-day, 60-day, six-month, and 12-month mortality rates. The NHFS and SORT version 2.0 (henceforth referred to as SORT) was calculated using the pre-defined parameters. The actual 30-day mortality was compared against the NHFS and SORT score, using the Z-test of proportions and the receiver operating characteristic to assess the model fit and statistical significance, with a p-value of 0.05 being the determining value. The Hosmer-Lemeshow (H-L) test was used to assess the goodness of fit for logistic regression models. Results The average age of the 157 patients was 83 years ±1.304 (±1.6%) (81.696-84.304; 95% CI intervals). The cohort was made up of 106 women and 51 men (106/157 women to 51/157 men). The mean NHFS was 5.0 - representing a 30-day mortality risk prediction of 5.99%. The mean SORT 30-day mortality was 8.28%. When compared to the actual 30-day mortality of 6.67%, the NHFS was not significantly different to this value (5.99% vs. 6.67%; p = 0.453; p > 0.05); however the SORT 30-day prediction of 8.28% was significantly higher than the actual mortality rate (8.28% vs. 5.99%; p < 0.00001); therefore, NHFS was significantly more accurate than the SORT score in predicting 30-day mortality for this particular cohort over this particular time period. Conclusion This paper suggests there is perhaps greater utility in predicting 30-day mortality through the use of the NHFS over the SORT tool; however, one must be cautious in determining clear superiority. The main rationale for this finding is that a dedicated tool, which incorporates measures relevant to the cohort most affected. Much work is yet to be done in fine-tuning predictor tools.