Cutaneous functional units (CFUs) versus total body surface area burned (TBSA) for predicting range of motion outcomes: A comparison of predictive models.

Abstract

Motion-limiting scar contractures are a common and functionally impactful complication after burn injury, but predicting their development remains difficult. While total body surface area (TBSA) burned is a standard metric for burn extent and severity, it does not consider burn location, an important factor impacting joint-level motion. Cutaneous functional units (CFUs) offer a more location-specific and functionally relevant assessment of burn injury characteristics, yet their predictive value for motion outcomes remains underexplored. This study validated clinical prediction models incorporating CFUs and TBSA, separately and together, to compare their predictive performance for motion outcomes after burn injury. Using data from two multi-center studies, we compared the performance of predictive models with CFU, TBSA, and CFU+TBSA. Fractional regression was used to model range of motion as a function of predictors. Models were internally validated using 10-fold cross-validation on a training dataset and externally validated with an independent testing dataset. Discrimination and calibration were evaluated. The CFU+TBSA model demonstrated the highest discriminatory power (0.7675; 95 % CI: 0.7497-0.79323), which was only slightly higher than the CFU model (0.7673; 95 % CI: 0.7497-0.7915) in the training dataset. Both the CFU and CFU+TBSA models statistically outperformed the TBSA model (0.7558; 95 % CI: 0.7366-0.7914). Validation in an external dataset revealed lower discrimination for all three models (C-statistic 0.6196-0.6322) and no difference between the models. The discriminatory power of all three models improved when analyzing only finger and thumb observations in the training dataset (C-statistic 0.7952-0.8087), with the CFU and CFU+TBSA models again significantly outperforming TBSA. Models using CFUs demonstrated better predictive performance than the TBSA-only model, indicating that CFUs enhance predictive value for normalized joint-level motion outcomes at hospital discharge. They appear to have particular utility with hand burns. CFUs offer a more granular, location-specific, and functionally relevant assessment than TBSA alone.