A Model For Prediction Of Survival In Critically III Dogs

摘要: Summary A system was developed for objective prediction of survival, which could be applied to all critically ill patients, regardless disease, at an early stage hospitaluation. Such a would allow risk assessment groups experimental studies according probability survival and it might us avoid expenditure scarce clinical resources on animals with little chance survival. The prospective study included 200 dogs: 93 females 107 males, representing 62 breeds. With defined as alive 30 days after admisssion ICU the overall mortality rate 40.5% (81 dogs). Data collected signalment parameters that reflected vital organ function, severity physiologic derangement extent reserve. We recorded most abnormal value each parameter within 24 hours admission logistic regression analysis used analyze four different weighting systems. best model had concordance 0f 86.5% outcome, then re-evaluated determine whether individual variables eliminated without losing predictive accuracy. Fourvariables were eliminated, resulting in final 18 86.3% outcome. At 0.5 cut-off predicted risk, sensitivity Of 69%, specificity ficity 86% positive 77%. A Receiver Operating Curve constructed using serial cut-offs outcome from 0.1 0.9, area under curve 0.89. Thus, equation generated gives estimate