Rangel-Frausto M S, Pittet D, Hwang T, Woolson R F, Wenzel R P
Department of Internal Medicine, University of Iowa College of Medicine, Iowa City, USA.
Clin Infect Dis. 1998 Jul;27(1):185-90. doi: 10.1086/514630.
We conducted a 9-month prospective cohort study of 2,527 patients with systemic inflammatory response syndrome in three intensive care units and three general wards in a tertiary health care institution. Markov models were developed to predict the probability of movement to and from more severe stages--sepsis, severe sepsis, or septic shock--at 1, 3, and 7 days. For patients with sepsis, severe sepsis, and septic shock, the probabilities of remaining in the same category after 1 day were .65, .68, and .61, respectively. The probability for progression after 1 day was .09 for sepsis to severe sepsis and .026 for severe sepsis to shock. The probability of patients with sepsis, severe sepsis, and septic shock dying after 1 day was .005, .009, and .079, respectively. The model can be used to predict the reduction in end organ dysfunction and mortality with use of increasingly effective antisepsis agents.
我们在一家三级医疗机构的三个重症监护病房和三个普通病房，对2527例全身炎症反应综合征患者进行了为期9个月的前瞻性队列研究。构建马尔可夫模型来预测在第1天、3天和7天进入和脱离更严重阶段（脓毒症、严重脓毒症或感染性休克）的概率。对于脓毒症、严重脓毒症和感染性休克患者，1天后仍处于同一类别的概率分别为0.65、0.68和0.61。脓毒症进展为严重脓毒症1天后的概率为0.09，严重脓毒症进展为休克1天后的概率为0.026。脓毒症、严重脓毒症和感染性休克患者1天后死亡的概率分别为0.005、0.009和0.079。该模型可用于预测使用越来越有效的抗菌药物后终末器官功能障碍和死亡率的降低情况。
