The article "Competing Risks in Epidemiology: Possibilities and Pitfalls" by Per Kragh Andersen, Ronald B Geskus, Theo de Witte, and Hein Putter discusses the challenges and implications of analyzing competing risks in epidemiological studies. The authors begin by defining risk and rate in epidemiology, emphasizing the distinction between the probability of an event occurring (risk) and the frequency of its occurrence (rate). They introduce survival analysis, which deals with censored data, and explain how the Kaplan-Meier estimator and Nelson-Aalen estimator are used to estimate survival functions and cumulative hazards, respectively. The article then delves into the concept of competing risks, where the event of interest (e.g., disease onset) is subject to other competing events (e.g., death without the disease). This setting requires a different approach from standard survival analysis because the cumulative incidence of the event of interest depends on both the cause-specific hazard of the event and the hazard of the competing event. The authors illustrate this with an example using data from the European Group for Blood and Marrow Transplantation (EBMT), where they analyze the risk of relapse and non-relapse mortality (NRM) in patients with chronic myeloid leukemia (CML) after transplantation. Key points include: 1. **Naive Estimation**: Naive estimation methods, such as using the Kaplan-Meier estimator for relapse alone, can lead to biased estimates of the cumulative incidence. 2. **Modeling Competing Risks**: The authors discuss the use of Cox regression models for cause-specific hazards and the Fine-Gray model for cumulative incidences. While Cox models are straightforward to fit, they may not capture the complex relationships between covariates and cumulative incidences. 3. ** Independence Assumption**: The concept of "independent" competing risks, where the hazards of the competing events are independent, is often violated in practice and should be interpreted with caution. 4. **Interpretation**: The article highlights that the interpretation of parameters in competing risk models can be more complex compared to standard survival analysis, as the relationship between covariates and cumulative incidences may differ from their relationship with cause-specific hazards. The authors conclude by emphasizing the importance of understanding the limitations and appropriate methods for analyzing competing risks in epidemiological studies, particularly when dealing with complex data structures and competing events.The article "Competing Risks in Epidemiology: Possibilities and Pitfalls" by Per Kragh Andersen, Ronald B Geskus, Theo de Witte, and Hein Putter discusses the challenges and implications of analyzing competing risks in epidemiological studies. The authors begin by defining risk and rate in epidemiology, emphasizing the distinction between the probability of an event occurring (risk) and the frequency of its occurrence (rate). They introduce survival analysis, which deals with censored data, and explain how the Kaplan-Meier estimator and Nelson-Aalen estimator are used to estimate survival functions and cumulative hazards, respectively. The article then delves into the concept of competing risks, where the event of interest (e.g., disease onset) is subject to other competing events (e.g., death without the disease). This setting requires a different approach from standard survival analysis because the cumulative incidence of the event of interest depends on both the cause-specific hazard of the event and the hazard of the competing event. The authors illustrate this with an example using data from the European Group for Blood and Marrow Transplantation (EBMT), where they analyze the risk of relapse and non-relapse mortality (NRM) in patients with chronic myeloid leukemia (CML) after transplantation. Key points include: 1. **Naive Estimation**: Naive estimation methods, such as using the Kaplan-Meier estimator for relapse alone, can lead to biased estimates of the cumulative incidence. 2. **Modeling Competing Risks**: The authors discuss the use of Cox regression models for cause-specific hazards and the Fine-Gray model for cumulative incidences. While Cox models are straightforward to fit, they may not capture the complex relationships between covariates and cumulative incidences. 3. ** Independence Assumption**: The concept of "independent" competing risks, where the hazards of the competing events are independent, is often violated in practice and should be interpreted with caution. 4. **Interpretation**: The article highlights that the interpretation of parameters in competing risk models can be more complex compared to standard survival analysis, as the relationship between covariates and cumulative incidences may differ from their relationship with cause-specific hazards. The authors conclude by emphasizing the importance of understanding the limitations and appropriate methods for analyzing competing risks in epidemiological studies, particularly when dealing with complex data structures and competing events.