Survival analysis II: Cox regression

In contrast to the Kaplan-Meier method, Cox proportional hazards regression can provide an effect estimate by quantifying the difference in survival between patient groups and can adjust for confounding effects of other variables. The purpose of this article is to explain the basic concepts of the Cox regression method, and to provide some guidance regarding the presentation of the results.     

Survival analysis I: the Kaplan-Meier method

The Kaplan-Meier (KM) method is used to analyze 'time-to-event' data. The outcome in KM analysis often includes all-cause mortality, but could also include other outcomes such as the occurrence of a cardiovascular event. The purpose of this article is to explain the basic concepts of the KM method, to provide some guidance regarding the presentation of the KM results and to discuss some important limitations of this method. To do this, we use a clinical example derived from the nephrology literature.     

血液透析高龄患者生存状态调查及分析

目的调查本中心血液透析高龄患者生存状态,了解患者并发症情况及死亡因素。方法选取接受维持性血液透析患者52例,使用SF-36健康调查表调查患者生存质量,检测患者贫血指标、营养指标和炎症指标,统计患者血管通路、并发症情况和死亡情况及死亡因素。结果 23例患者出现贫血症状,45例患者出现营养不良现象,38例患者出现轻微炎症;14例患者选择内瘘,26例患者选择Cuff导管,3例患者选择临时血透导管,9例患者因功能障碍放弃内瘘改用Cuff导管;17例患者死亡,肺部感染和脑梗死是主要致死原因;患者生存质量评分为(85.4±16.5)分;经Cox回归模型分析发现合并糖尿病和透始血红蛋白是影响患者生存质量的主要因素。结论维持性血液透析能延长高龄患者生存时间,改善生存状态。     

支气管肺泡灌洗术治疗危重患者的生存分析

目的探讨支气管肺泡灌洗术(BAL)治疗危重患者的生存期差异。方法收集重症监护室149例使用呼吸机辅助呼吸的重症患者作为观察对象,分为BAL组及未使用BAL的对照组,对患者进行2年的随访,使用Kaplan-Meier法比较两组的生存期差异。结果治疗后两组的APACHE II及简化急性生理评分II评分均明显下降,且均以BAL组上述两项评分显著低于对照组（P < 0.05）。两组分别有5例及7例患者治疗1周内死亡,分别占6.94%及9.10%,第1周病死率无统计学差异（χ²=0.168,P=0.682）。BAL组最长生存期的患者大于124周,且该例患者删失;而对照组最长生存期为104周;BAL组的生存期的算术均数及中位数分别为43.5、30.0周,而对照组分别为33.2、25.0周,Kaplan-Meier生存分析显示BAL组的生存期显著长于对照组（χ²=4.419,P=0.036）。结论尽管使用与不使用BAL短期内对患者的生存率无明显差异,但使用BAL治疗的患者的生存期显著长于不使用BAL者,提示BAL可改善危重患者的远期预后。     

104例成人急性白血病生存分析

目的 总结成人急性白血病患者的预后因素。方法 对我院3年来104例治疗和随访的成人急性白血病患者，采用对数秩检验(Logrank)和比例危险率(COX)回归模型分析其预后因素。结果 104例患者中有15例达5年以上长期生存，总的5年长期生存率为14．4％，急性髓细胞白血病(AML)为18．6％，急性淋巴细胞白血病(ALL)为1.8％，确诊后第1年病死率高；急性早幼粒细胞白血病(M3)的长期生存率较其他类型高。第1次化疗诱导缓解、骨髓移植、乳酸脱氢酶(LDH)正常的患者生存时间较长，核型结构或数目异常组生存概率均低于正常棱型组。结论 成人急性白血病5年长期生存率为14．4％，M3预后较好，第1次诱导是否缓解及是否骨髓移植是影响长期生存的主要预后因素，LDH可考虑作为急性白血病(AL)疗效及预后判断的一项参考指标，核型分析在AL中具有独立于其他临床指标的预后意义。     

Survival from out-of-hospital cardiac arrest. A multivariate analysis

A recursive estimation model is used to investigate the roles of cardiopulmonary resuscitation (CPR) and advanced life support in improving survival from out-of-hospital cardiac arrest. The importance of life support measures is clearly evidenced in the analysis: Fewer elapsed minutes between the cardiac arrest and the start of CPR increase the probabilities of both a favorable cardiac rhythm and defibrillation and the probability of survival. Similarly, a shorter elapsed time between the start of CPR and defibrillation is significantly related to a higher probability of survival of the cardiac arrest. Personal characteristics also contribute to survival, but primarily via their association with a favorable initial postarrest cardiac rhythm and the probability of defibrillation. The initial postarrest cardiac rhythm is shown to be an indicator of the heart's condition, but when other factors associated with survival are included in the analysis, it does not independently influence an individual's probability of survival.     

老年非小细胞肺癌患者术后生存期的研究

目的探讨影响70岁以上非小细胞肺癌患者术后生存期的相关因素,为高龄肺癌患者的手术方式选择提供参考。方法回顾性分析手术治疗的70岁以上高龄非小细胞肺癌患者95例,利用Kaplan-Merier单因素和COX回归分析性别、年龄、症状、伴随疾病、并发症、手术方式、病理分期、体力状况(ECOG)评分、病理类型等9个因素对术后生存期的影响。结果术后随访截止2009年6月,随访率90.53%(失随访9例),术后总的1、3、5年生存率分别为73.5%、63.6%、43.1%,COX回归多因素分析显示,影响术后生存的独立因素为:病理分期和ECOG评分。结论70岁以上非小细胞肺癌患者,病理分期早和术后体力状态好的患者术后生存期长。     

Survival of nurse-managed centers: the importance of cost analysis

Despite the explosive growth in nurse-managed centers (NMC) in the past 20 years, most have been unable to achieve financial self-sufficiency, and many have closed. Combining costing techniques with outcome measures provides essential information needed by NMC for making operating decisions and for marketing NMC performance. These outcome data can be persuasive to policy makers and institutional decision makers and are crucial for NMC to improve their competitiveness in the health care market place.     

HMO enrollment and Medicaid. Survival analysis with a Weibull function

This paper analyzes the length of enrollment in Health Maintenance Organizations (HMOs) by performing secondary data analysis. It estimates a Weibull survival function using only simple regression. For middle-class enrollees in one HMO, the hazard rate (the disenrollment rate for the remaining members of a cohort) falls sharply with the length of enrollment. For medicaid beneficiaries allowed to remain in the fee-for-service sector, the hazard rate in another HMO is much greater but falls less sharply. Limiting the choice of Medicaid beneficiaries to HMOs would increase their length of enrollment, lowering a major barrier to HMOs' participation in Medicaid.     

乳腺癌根治术后568例患者生存分析

目的：探讨乳腺癌根治术后患者临床病理因素与生存率的分析。方法回顾性分析568例女性乳腺癌根治术后患者的临床病理学资料,以Kaplan-Meier法比较生存差异,Cox单因素和多因素方法分析根治性乳腺癌的无病生存率及总生存率的危险因素。结果中位随访73个月,随访期内,124例患者出现复发转移,74例患者死亡。全组患者的1年、3年及5年无病生存率分别为94.0%、81.5%和73.6%,1年、3年及5年总生存率分别为97.5%、92.1%和87.7%。单因素分析结果显示,年龄、T、N、放疗及化疗方案与乳腺癌根治术后肿瘤的无病生存时间相关（P均＜0.05）,而年龄、T、N、ER及有无放疗与乳腺癌根治术后肿瘤的总生存时间相关（P均＜0.05）。多因素分析结果显示,年龄、T、N、化疗方案是影响Ⅰ、Ⅱ、Ⅲ期乳腺癌根治术后无病生存率的独立危险因素（P均＜0.05）,而年龄、是否放疗是影响Ⅰ、Ⅱ、Ⅲ期乳腺癌根治术后总生存率的独立危险因素（P均＜0.05）。结论年龄、T、N、化疗方案是影响Ⅰ、Ⅱ、Ⅲ期乳腺癌根治术后无病生存率的独立险因素,而年龄、放疗与否是影响Ⅰ、Ⅱ、Ⅲ期乳腺癌根治术后总生存率的独立危险因素。     

Survival

A 500-bed hospital enlists more than one-third of its staff to prepare for a Joint Commission survey.     

Survival and cost analysis of fatalities of the kobe earthquake in Japan

Objectives

The authors investigated the dying patterns, and cause and preventability of deaths in a major earthquake disaster, and estimated the cost needed to enhance emergency medical services (EMS) response to prevent “unnecessary” deaths.

Methods

The authors reviewed autopsy data in the Hanshin-Awaji (Kobe) earthquake of 1995. A survival analysis was performed to determine the time course and pattern of dying of these deaths. A cost analysis to estimate acceptable cost for EMS to reduce fatalities was also performed. Potentially salvageable life-years based on expected life-years among fatalities were calculated and used to simulate an acceptable cost for an enhanced EMS disaster response.

Results

The authors analyzed 5,411 fatalities. More than 80% of these patients died within three hours. There were statistically significant differences in survival/dying patterns among causes of death. Thirteen percent of victims experienced a protracted death, which could have been prevented with earlier medical or surgical intervention. The monetary cost of these lost lives was estimated at approximately $600 million US.

Conclusions

Survival analysis revealed a significant population of potentially salvageable patients if more timely and appropriate medical intervention had been available immediately after the earthquake. Based on our cost analysis, and assuming a 1% annual probability of an earthquake and a 30% enhanced lifesaving capability of the EMS effort, approximately $2 million annually could be a reasonable expenditure to achieve the goal of reducing preventable deaths in disasters.     

Survival analysis of adverse effects data in the Beta-Blocker Heart Attack Trial

Monitoring for adverse effects in an integral part of controlled clinical trials. Traditionally the results of monitoring are reported as either cumulative percentages at the end of the study or cross-sectional percentages at a given time in the study. These results are likely to underestimate the true number of complaints because participants may be withdrawn (e.g., deaths, losses to follow-up, and refusals) before they ever complain of an adverse effect. However, survival analysis methods can be used to compare the distributions of "time to first complaint" in the active and placebo treatment groups, taking into account withdrawals. Participants in the Beta-Blocker Heart Attack Trial were monitored for possible adverse effects. On each follow-up visit they were asked whether they had had any of four conditions (blacking out, fatigue, depression, and bronchospasm) since their previous visit about 3 months earlier. The patients were followed for up to 30 months. For fatigue and bronchospasm, the complaint-free time was significantly longer in the placebo vs. active (propranolol) treatment group (P less than 0.005).     

艾滋病患者终止抗病毒治疗后的生存状况分析

目的:了解艾滋病患者终止抗病毒治疗后的生存状况及其影响因素。方法:采取回顾性队列研究方法,选择2007-2013年百色市终止抗病毒治疗的201例艾滋病患者作为研究对象,运用寿命表法估计生存率,Kaplan-Meier绘制生存曲线,COX回归比例风险模型进行生存时间影响因素分析。结果:研究对象男女比例为2.65∶1,平均年龄(43.35±14.05)岁,以壮族、初中以下文化、农民及农民工、已婚/同居为主,感染途径主要为异性性途径传播。艾滋病患者终止抗病毒治疗后第12、24、36个月的累积生存率为0.77、0.64、0.52。患者终止抗病毒治疗前最近一次CD4+T淋巴细胞计数水平≤100个/mm3组(A组)、100~200个/mm3组(B组)和>200个/mm3组(C组)的组间生存率比较差异有统计学意义(P<0.05),A组和B组患者相对于C组患者的死亡危险度分别为4.587(95%CI:2.389~8.809)和2.501(95%CI:1.308~4.784)。多因素COX比例风险模型分析结果显示:患者性别、治疗终止时是否有治疗药物毒副作用、终止治疗前最近一次CD4+T淋巴细胞计数水平与生存时间存在统计学关联,其中男性相对于女性的死亡危险度为2.846(95%CI:1.257~6.4 43),患者终止治疗时出现治疗药物毒副作用的死亡危险度为2.765(95%CI:1.446~5.285)。结论 :终止抗病毒治疗降低了艾滋病患者的生存时间,应加强抗病毒治疗患者的依从性教育及依从性支持服务,及时处置抗病毒治疗不良反应,减少停药及治疗失访的发生,对延长患者生存时间有重要意义。     

持续腹膜透析患者的生存率及多变量预后分析

目的分析腹膜透析患者的生存率及其独立的预后因素,观察腹膜转运特性及相关临床特点对患者生存的影响。方法选择北京协和医院肾内科随诊腹膜透析患者232例,采用Cox模型回归分析透析开始时各临床指标对生存时间的影响。结果CAPD患者1年、2年、3年及4年总体生存率分别为91.1%、77.7%、68.7%及55.8%;单因素COX模型回归分析表明：糖尿病患者、腹膜高转运状态、年龄增加、血浆白蛋白低于3.0的患者预期生存率均降低（P〈0.05）。232例患者中180（78%）位患者在开始透析的前6个月完成腹膜平衡实验（PET）,其中高转运状态与死亡显著相关（与低于平均转运患者相比,RR2.70;95% CI 1.03 to 7.05;P=0.043）。经多因素COX模型回归分析,糖尿病和年龄是尿毒症腹膜透析患者死亡的独立危险因素。糖尿病患者死亡的相对危险度为非糖尿病患者的2.96倍（95% CI 1.62 to 5.38;P〈0.0001）;年龄每增加10岁,相对危险度增加0.31（P=0.039）。结论根据透析前年龄、原发病、血清白蛋白和腹膜的转运状态可以对腹膜透析患者的预后进行初步判断。