بسم الله الرحمن الرحیم. generally,survival analysis is a collection of...
TRANSCRIPT
Generally,survival analysis is a collection of statistical procedures for data analysis for which the outcome variable of interest is time until an event occurs.
The usual objective with this type of data is to determine the length of remission and survival and to compare the distributions of remission and survival time in each group.
Thirty melanoma patients (stages 2 to 4) were studied to compare the immunotherapies BCG (Bacillus Calmette-Guerin) and Corynebacterium parvum for their abilities to prolong remission duration and survival time. The age, gender, disease stage, treatment received, remission duration, and survival time are given in Table 3.1.
Comparison of Two Treatments
Comparison of Three Diets
A laboratory investigator interested in the relationship between diet and the development of tumors divided 90 rats into three groups and fed them low-fat, saturated fat, and unsaturated fat diets, respectively (King et al., 1979). The rats were of the same age and species and were in similar physical condition. An identical amount of tumor cells were injected into a foot pad of each rat. The rats were observed for 200 days.
h1(t): patients with acute leukemia who do not respond to treatment have an increasing hazard rate
h2(t): indicates the risk of soldiers wounded by bullets who undergo surgeryh3(t): is the risk of healthy persons between 18 and 40 years of age whose mainrisks of death are accidents.h4(t): describes the process of human lifeh5(t): patients with tuberculosis have risks that increase initially, then decrease after treatment
The possible confounding effect
In this case, we would say that the treatmenteffect is confounded by the effect of log WBC.
Need to adjust for imbalance in the distribution of log WBC
Interaction
What we mean by interaction is that the effect of the treatment may be different, depending on the level of log WBC.
There is strong treatment by log WBC interaction, and we would have to qualify the effect of the treatment as depending on the level of logWBC.
1) To stratify on log WBC and compare survival curves for different strata
or
2) To use mathematical modeling procedures such as the proportional hazards or other survival models
Introduction to Kaplan-Meier
Non-parametric estimate of the survival function:No math assumptions! (either about the underlying hazard function or about proportional hazards).Simply, the empirical probability of surviving past certain times in the sample (taking into account censoring).
Introduction to Kaplan-Meier
• Non-parametric estimate of the survival function.
• Commonly used to describe survivorship of study population/s.
• Commonly used to compare two study populations.
• Intuitive graphical presentation.
Beginning of study End of study Time in months
Subject B
Subject A
Subject C
Subject D
Subject E
Survival Data (right-censored)
1 .subject E dies at 4 months
X
100%
Time in months
Corresponding Kaplan-Meier Curve
Probability of surviving to 4
months is 100% = 5/5
Fraction surviving this
death = 4/5Subject E dies at 4
months
Beginning of study End of study Time in months
Subject B
Subject A
Subject C
Subject D
Subject E
Survival Data
2 .subject A drops out after
6 months
1 .subject E dies at 4 months
X
3 .subject C dies at 7 monthsX
100%
Time in months
Corresponding Kaplan-Meier Curve
subject C dies at 7 months
Fraction surviving this
death = 2/3
Beginning of study End of study Time in months
Subject B
Subject A
Subject C
Subject D
Subject E
Survival Data
2 .subject A drops out after
6 months
4 .Subjects B and D survive for the whole
year-long study period
1 .subject E dies at 4 months
X
3 .subject C dies at 7 monthsX
100%
Time in months
Corresponding Kaplan-Meier Curve
Rule from probability theory:
P(A&B)=P(A)*P(B) if A and B independent
In survival analysis: intervals are defined by failures (2 intervals leading to failures here).
P(surviving intervals 1 and 2)=P(surviving interval 1)*P(surviving interval 2)
Product limit estimate of survival = P(surviving interval 1/at-risk up to failure 1) * P(surviving interval 2/at-risk up to failure 2) = 4/5 * 2/3= .5333
The product limit estimate
• The probability of surviving in the entire year, taking into account censoring
• = (4/5) (2/3) = 53%
• NOTE: 40% (2/5) because the one drop-out survived at least a portion of the year.
• AND <60% (3/5) because we don’t know if the one drop-out would have survived until the end of the year.
n(f ): the number of subjects in the risk set at the start the interval
t(f): failure time
q(f): the number of censored subjects
m(f): the number of failures
how to evaluate whether or not KM curves for two or more groups are statistically equivalent?Themost popular testing method is called the log–rank test.
All the test results are highly significant yielding a similar conclusion to reject the null hypothesis.
Time-independent variable:Values for a given individualdo not change over time; e.g.,SEX and Smoking status(SMK).
Why the Cox PH Model Is Popular?1) Semiparametric property
2) Cox PH model is “robust”
the baseline hazard is not specified, reasonably good estimates of regression coefficients, hazard ratios of interest, and adjusted survival curves can be obtained for a wide variety of data situations.
We need are estimates of the b’s to assess the effect ofexplanatory variables of interest.
The measure of effect, which is called a hazard ratio (HR)
Maximum likelihood (ML) Estimation of the Cox PH Model
1) Test for treatment effect:Wald statistic: P <0.001 (highly significant)
Conclusion: treatment effect is significant
2) Point estimate:HR = 4.523
Conclusion: the hazard for the placebo group is 4.5 times the hazard for the treatment group
3) 95% confidence interval for the HR: (2.027,10.094)
the potential confounding effect
HR for model 1 (4.523) is higher than HR for model 2 (3.648)
Confounding: crude versus adjusted HR are meaningfully different.
Confounding due to log WBC must control for log WBC, i.e., prefer model 2 to model 1.
The Meaning of the PH Assumption
The PH assumption requires that the HR is constant over time
The hazard for one individual is proportional to the hazard for any other individual, where the proportionality constant is independent of time.
There are two types of graphical techniques available.
1) Comparing estimated –ln(–ln) survivor curves
2) Compare observed with predicted survivor curves.