Notes on Survival Time Analysis

Introduction

• Overview of survival time analysis.
• Key topics: censoring of data, Kaplan-Meier curve, log-rank test, Cox regression.
• This is an introductory video; detailed videos for each topic will follow.

What is Survival Time Analysis?

• A group of statistical methods analyzing the time until an event occurs.
• Key components:
  • Start time: Beginning of the observation period.
  • End time: Moment the event occurs.
  • Examples:
    • Time from drug rehabilitation to relapse.
    • Time until death after disease diagnosis.
    • Time to return to work post-burnout.
    • Duration of a component in engineering tests.

Censoring in Survival Analysis

• Censoring occurs when the event is not observed for some subjects.
• Reasons for Censoring:
  • Study ends before the event occurs.
  • Participants withdraw from the study.
  • Other events occur (e.g., death, loss of the object).
  • Events may go unnoticed until the next examination.
• Censored data must be handled properly for accurate survival analysis.

Methods in Survival Time Analysis

1. Kaplan-Meier Curve

• Graphical representation of survival rates over time.
• Axes:
  • X-axis: Time (days, weeks, months).
  • Y-axis: Survival rate.
• Example: Assessing the likelihood of a dental filling lasting longer than five years.
• Provides visual insight into survival probabilities.

2. Log-Rank Test

• Compares survival distributions between two or more independent groups.
• Example: Comparing survival times of different filling materials (Material A vs. Material B).
• Hypotheses:
  • Null Hypothesis: No difference in survival distributions.
  • Alternative Hypothesis: Groups have different survival distributions.
• Evaluates p-value to determine significance (common significance level: 0.05).

3. Cox Regression

• Assesses the impact of several variables on survival time.
• Example: Investigating if material type and age of participants affect survival time.

Practical Calculation of Methods

• Use of DataTab for online calculations:
  • Website: datadeb.net
  • Input data into specified columns (time, event status, material type, age).
  • Options for Kaplan-Meier curves, log-rank tests, and Cox regression.
• Interpretation of results is crucial. A summary option is available for clarity.

Conclusion

• Follow-up videos available for detailed explanations of Kaplan-Meier curves, log-rank tests, and Cox regression.