Towards more efficient burn care: Identifying factors associated with good quality of life post-burn

• The probability of achieving a high BSHS-B score six months from injury was estimated.
• Gender, surgery, upper limb burn and BSHS-B score within six weeks were significant predictors.
• ROC analysis found 90% accuracy overall and 14% specificity at 80% probability of good outcome.
• ROC analysis in an independent sample confirmed high accuracy of the predictive model.
• Missing data analysis found that younger patients with smaller TBSA were lost to follow-up.

BackgroundAs minor burn patients constitute the vast majority of a developed nation case-mix, streamlining care for this group can promote efficiency from a service-wide perspective. This study tested the hypothesis that a predictive nomogram model that estimates likelihood of good long-term quality of life (QoL) post-burn is a valid way to optimise patient selection and risk management when applying a streamlined model of care.MethodA sample of 224 burn patients managed by the Burn Service of Western Australia who provided both short and long-term outcomes was used to estimate the probability of achieving a good QoL defined as 150 out of a possible 160 points on the Burn Specific Health Scale-Brief (BSHS-B) at least six months from injury. A multivariate logistic regression analysis produced a predictive model provisioned as a nomogram for clinical application. A second, independent cohort of consecutive patients (n = 106) was used to validate the predictive merit of the nomogram.Results and discussionMale gender (p = 0.02), conservative management (p = 0.03), upper limb burn (p = 0.04) and high BSHS-B score within one month of burn (p < 0.001) were significant predictors of good outcome at six months and beyond. A Receiver Operating Curve (ROC) analysis demonstrated excellent (90%) accuracy overall. At 80% probability of good outcome, the false positive risk was 14%.The nomogram was validated by running a second ROC analysis of the model in an independent cohort. The analysis confirmed high (86%) overall accuracy of the model, the risk of false positive was reduced to 10% at a lower (70%) probability. This affirms the stability of the nomogram model in different patient groups over time.An investigation of the effect of missing data on sample selection determined that a greater proportion of younger patients with smaller TBSA burns were excluded due to loss to follow up.ConclusionFor clinicians managing comparable burn populations, the BSWA burns nomogram is an effective tool to assist the selection of patients to a streamlined care pathway with the aim of improving efficiency of service delivery.