4 August 2016

Survival analysis (in R)

Nadeem cartoon - drowning

Course Description

This comprehensive course is tailored for medical professionals, researchers, and Clinical Research Organizations (CROs) seeking to master statistical methods for analyzing medical survival data. Survival data analysis is a crucial tool in medical research, tracking time-to-event outcomes such as mortality, disease progression, or treatment efficacy. Our course provides a robust foundation in R programming and advanced statistical techniques, empowering you to conduct sophisticated survival analyses that can potentially revolutionize patient care and medical research.

Course Content:

  1. Introduction to R: Your Statistical Swiss Army Knife
    • Navigate the R environment with confidence
    • Master data importation techniques (CSV, Excel, SPSS)
    • Perform essential data manipulation using dplyr and tidyr
    • Create publication-quality graphics with ggplot2
    • Example: Import a clinical trial dataset and create a basic visualization of patient demographics.
  2. Data Visualization: Bringing Your Data to Life
    • Explore advanced plotting techniques for survival data
    • Create informative and aesthetically pleasing survival curves
    • Utilize interactive visualizations for dynamic data exploration
    • Example: Develop a multi-layer plot comparing survival curves across different treatment groups.
  3. Life Tables: The Building Blocks of Survival Analysis
    • Construct and interpret life tables
    • Calculate survival probabilities and hazard rates
    • Understand the concept of censoring in survival data
    • Example: Create a life table for a cohort of cancer patients undergoing a novel immunotherapy treatment.
  4. Kaplan-Meier Survival Curves: The Gold Standard of Survival Analysis
    • Estimate survival functions using the Kaplan-Meier method
    • Compare survival curves across multiple groups
    • Conduct log-rank tests for statistical significance
    • Example: Analyze the efficacy of a new cardiovascular drug by comparing Kaplan-Meier curves of treated vs. control groups.
  5. Cox Regression Models: Unraveling the Complexity of Survival
    • Implement and interpret Cox proportional hazards models
    • Assess the impact of multiple covariates on survival
    • Test and handle violations of the proportional hazards assumption
    • Example: Evaluate the influence of age, gender, and biomarkers on survival rates in a lung cancer study.
  6. Parametric Survival Models: When Assumptions Are Your Friends
    • Explore Weibull, exponential, and log-normal models
    • Understand when and how to apply parametric models
    • Compare parametric and semi-parametric approaches
    • Example: Model the time to cardiac event in patients with coronary artery disease using a Weibull distribution.
  7. Non-parametric Methods: Breaking Free from Distributional Shackles
    • Implement non-parametric tests for survival data
    • Understand the advantages and limitations of distribution-free methods
    • Apply smoothing techniques to survival curves
    • Example: Use a non-parametric approach to analyze survival in a heterogeneous population of stroke patients.
  8. Competing-Risk Models: When One Risk Isn’t Enough
    • Understand the concept of competing risks in survival analysis
    • Implement Fine-Gray models for competing risk data
    • Interpret cumulative incidence functions
    • Example: Analyze time to either cancer recurrence or death from other causes in a long-term follow-up study.

Throughout the course, you’ll work with real-world medical datasets, tackling complex survival analysis problems that mirror the challenges faced in clinical research. By the end, you’ll be equipped with a powerful statistical toolkit to advance medical knowledge and potentially improve patient outcomes.

Disclaimer: This course significantly reduces the risk of p-value misinterpretation and hazard ratio confusion, but may increase your tendency to view life events in terms of survival probabilities. Side effects may include dreaming in Kaplan-Meier curves and an irresistible urge to calculate the median survival time of your houseplants. Consult your statistician if symptoms persist:)

en_USEnglish