Comparing Survival Outcomes Between Groups: A Clinician's Guide
ClinicoPath Development Team
2025-10-09
Source:vignettes/14-comparing-survival-user-friendly.Rmd
14-comparing-survival-user-friendly.RmdIntroduction
In clinical practice, we often want to compare the outcomes of different groups of patients. For example, does a new treatment improve survival compared to the standard treatment? Does the presence of a specific biomarker affect patient prognosis? The Comparing Survival feature in the jSurvival module is designed to help you answer these types of questions.
Learning Objectives:
By the end of this guide, you will be able to:
- Compare survival outcomes between two or more groups using the jSurvival module.
- Interpret the results of a multi-group survival analysis.
- Understand the concept of “clinical significance” and how it differs from “statistical significance.”
- Create publication-ready survival plots from jamovi.
Comparing Two Groups
Let’s start with a simple example: comparing the survival of patients who received a new treatment to those who received a placebo.
Step 1: Set up Your Analysis
- Open your data in jamovi.
- Navigate to Analyses > jSurvival > Survival Analysis.
- Assign your Time Elapsed, Outcome, and Explanatory Variable (e.g., “Treatment Group”).
Step 2: Examine the Output
The output will look familiar to what you saw in the introductory guide. You will see a Kaplan-Meier plot with two survival curves (one for the treatment group and one for the placebo group) and a log-rank test with a p-value.
Interpreting the Results:
- The Plot: Visually inspect the Kaplan-Meier plot. Is there a clear separation between the two curves? If the curve for the treatment group is consistently above the curve for the placebo group, this suggests that the treatment is associated with better survival.
- The p-value: Look at the p-value from the log-rank test. If the p-value is less than 0.05, you can conclude that there is a statistically significant difference in survival between the two groups.
Comparing Multiple Groups
What if you have more than two groups to compare? For example, you might want to compare the survival of patients with different tumor grades (e.g., Grade 1, Grade 2, and Grade 3).
The process in jamovi is the same. Simply use your multi-level grouping variable (e.g., “Tumor Grade”) as the Explanatory Variable.
Interpreting the Results with Multiple Groups:
When you have more than two groups, the log-rank test will give you a single p-value. This p-value tells you if there is a significant difference somewhere among the groups, but it doesn’t tell you which specific groups are different from each other.
For example, if the overall p-value is 0.02, you know that there is a significant difference in survival between at least two of the tumor grades, but you don’t know if the difference is between Grade 1 and Grade 2, Grade 1 and Grade 3, or Grade 2 and Grade 3.
Pairwise Comparisons
To find out which specific groups are different, you can perform pairwise comparisons. In the jSurvival module, you can select the option for “Pairwise comparisons” to get p-values for each pair of groups.
Clinical Significance vs. Statistical Significance
It’s important to understand the difference between clinical significance and statistical significance.
- Statistical Significance: This is determined by the p-value. A statistically significant result means that the observed difference is unlikely to be due to random chance.
- Clinical Significance: This refers to the practical importance of the result. Is the difference in survival large enough to be meaningful to patients and clinicians? Will it change clinical practice?
A result can be statistically significant but not clinically significant. For example, a new treatment might result in a statistically significant increase in survival, but the actual increase might only be a few days. Whether this is clinically meaningful is a matter of clinical judgment.
How to Assess Clinical Significance:
- Look at the Median Survival: The jSurvival module will report the median survival time for each group. The median survival is the time at which 50% of the patients are still alive. Comparing the median survival times between groups can give you a sense of the magnitude of the difference.
- Consider the Context: The clinical context is crucial. A small improvement in survival might be very important for a disease with a very poor prognosis, while the same improvement might be less meaningful for a disease that is already highly curable.
Creating Publication-Ready Plots
The jSurvival module allows you to customize your survival plots for publication.
You can:
- Add a risk table to show the number of patients at risk at different time points.
- Display confidence intervals for the survival curves.
- Add the p-value from the log-rank test directly to the plot.
Conclusion
Comparing survival outcomes is a core task in clinical research. The jSurvival module in jamovi provides a user-friendly way to perform these comparisons and generate publication-quality plots.
Key Takeaways:
- You can compare survival between two or more groups using the same simple interface.
- For multi-group comparisons, use pairwise comparisons to identify which specific groups are different.
- Always consider both the statistical and clinical significance of your results.
Now that you know how to compare survival between groups, you are ready to explore more advanced features of the jSurvival module, such as Cox regression for multivariable analysis.
Common Pitfalls
- Multiple testing: Use adjusted p‑values when doing many pairwise comparisons.
- Follow-up imbalance: Large differences in follow-up time can bias comparisons; consider RMST or adjusted analyses.
- Sparse strata: Very small subgroups produce unstable estimates and misleading curves/tables.
- Clinical vs statistical significance: A small p‑value does not guarantee a meaningful effect size for patients.