# Survival Analysis (AIIMS Nov’08)

Several Statistical methods are available for analysis of survival data, of which three important ones are:**Survival rate**– mostly used in relation to*oncology to estimate 3- and 5-year*survival with various types of cancer put on different treatment regimen.^{Q}**Actuarial or Life table**method: also sometimes referred to in medical literature as Cutler – Ederer method. To illustrate the calculations involved in actuarial analysis, the length of time the patients have been in studyin frequency table. To determine the survival curve, the probability of death or outcome for each time period is calculated.^{Q}

The second step is to compute the probability for survival. This probability is sometimes also called the hazard functionor hazard rate.^{Q}

Results from life table analysis are usually presented in a survival curve rather than a table.**The Kaplan Meier**method of estimating survival is similar to actuarial analysis except that time since entry in the study is not divided into intervals for analysis. For this reason, it is especially appropriate in studies involving a small number of patients.^{ Q}

The Kaplan Meier product limit method involves fewer calculations than the actuarial method, primarily because survival is estimated each time a patient dies, so withdrawals are ignored.^{Q}

It gives exact survival proportionsbecause it uses exact survival times & hence the survival curve is step like rather than being a curve.^{ Q}

**Extra Edge:**

__Probability sampling__- Uses random method.
- It ensures all units in population
**have equal chance**of being represented **Probability****Simple random sampling****Stratified Random Sampling****Systematic Random Sampling****Cluster Sampling****Multistage Sampling**

**Non****Probability sampling :**- Non random selection
- May or
**may not represent population** - Used when no sampling frame available
- examples:
**Voluntary Sampling****Purposive Sampling****Quota Sampling****Snowball Sampling**

**DIAGNOSIS OF NON-NORMALITY**- Plot frequency curve and observe
- Mean vs. SD: If SD>30% of Mean: Non normal or highly dispersed

**Flow charts:**

Categorical = Qualitative

Related means “the same group” Unrelated refers to “different groups”

**How to solve various questions on tests of significance.**

- Look for “
**Type of Data"**in the question.