RESUMEN
Researchers in statistical shape analysis often analyze outlines of objects. Even though these contours are infinite-dimensional in theory, they must be discretized in practice. When discretizing, it is important to reduce the number of sampling points considerably to reduce computational costs, but to not use too few points so as to result in too much approximation error. Unfortunately, determining the minimum number of points needed to achieve sufficiently approximate the contours is computationally expensive. In this paper, we fit regression models to predict these lower bounds using characteristics of the contours that are computationally cheap as predictor variables. However, least squares regression is inadequate for this task because it treats overestimation and underestimation equally, but underestimation of lower bounds is far more serious. Instead, to fit the models, we use the LINEX loss function, which allows us to penalize underestimation at an exponential rate while penalizing overestimation only linearly. We present a novel approach to select the shape parameter of the loss function and tools for analyzing how well the model fits the data. Through validation methods, we show that the LINEX models work well for reducing the underestimation for the lower bounds.
RESUMEN
BACKGROUND: Our study sought to identify independent risk factors predisposing patients with necrotizing soft tissue infections (NSTIs) to mortality from among laboratory values, demographic data, and microbiologic findings in a small population. To this end, a retrospective review was conducted of the medical records of all patients with NSTI who had been treated at our institution from 2003 to 2012 (n=134). METHODS: Baseline demographics and comorbidities, clinical and laboratory values, hospital course, and the microbiologic characteristics of surgical incision cultures were recorded. Each variable was tested for association with survival status and all associated variables with p<0.15 were included in a logistic regression model to seek factors associated independently with mortality. RESULTS: Surprisingly, no demographic or pre-existing condition proved to be a predictor of mortality. Two laboratory values had an inverse correlation to mortality: High C-reactive protein (CRP) and highest recorded CRP. Of surgical incisions that grew bacteria in culture, 33.6% were polymicrobial. Mortality rates were highest with Enterococcus-containing polymicrobial infections (50%), followed by those containing Pseudomonas (40%), and Streptococcus spp. (27%). Understanding why so many studies across the literature, now including our own, find such disparate results for correlation of NSTI mortality with patient data may lie in the fundamentally dynamic nature of the organisms involved. CONCLUSIONS: This study suggests that no single factor present on admission is a robust predictor of outcome; it is likely that survival in NSTI is predicated upon a complex interaction of multiple host and microbial factors that do not lend themselves to reduction into a simple formula. It is also abundantly clear that the well-established principles of NSTI surgery should continue to be followed in all cases, with an emphasis on early debridement, irrespective of apparent severity of initial presentation.