Comparison of the rBaux Model with Harborview (HMC) and ABA-NBR Data for Standard Admissions.

Mortality generally declined over time in the Harborview standard admission data and in 2005–2009 mortality matched rBaux prediction save in Group 2, less than predicted, and 5, slightly more than predicted (bold). There was little change in the ABA-NBR standard admission mortality over time and mortality matched the rBaux predicted mortality in all groups. Since the ABA-NBR data from 2000–2007 was used to build the rBaux system, this is not surprising. However the ABA-NBR data from 1995–1999 also matched suggesting that rBaux scores are robust over time and that there has been little improvement in survival since 1995. (x%/y/z x = percent observed mortality, y = sample size and CC = the number placed on comfort care).

港景医疗中心（Harborview）标准入院数据集的死亡率随时间整体呈下降趋势；2005-2009年的研究数据中，除亚组2（实际死亡率低于预测值）与亚组5（略高于预测值，已加粗标注）外，其余亚组的死亡率均与rBaux预测值相符。ABA-NBR标准入院死亡率随时间几乎无明显变化，且所有亚组的实际死亡率均与rBaux预测值匹配。由于2000-2007年的ABA-NBR数据被用于构建rBaux模型，这一结果并不令人意外。然而1995-1999年的ABA-NBR数据同样呈现出匹配性，这表明rBaux评分具有良好的时间稳定性，且自1995年以来，患者生存率几乎未得到改善。（x%/y/z：x为观察死亡率百分比，y为样本量，CC为接受舒适护理的病例数）