Home > Medical Information > Disease & Transplant Outcomes > Understanding Survival Curves

Survival curves (sometimes called Kaplan-Meier curves) are a way of showing patients' survival rates in a chart. The survival rates shown on the National Marrow Donor Program^® (NMDP) Web site come from data the NMDP Research Program collects on blood stem cell transplant recipients. The data are collected to help doctors and patients make better treatment decisions and to improve transplant recipients' outcomes. The graphs are useful because they show several curves at once to compare different groups of patients.

In this example, three curves show survival rates for three groups of patients with chronic myelogenous leukemia (CML):

• Patients transplanted in chronic phase
• Patients transplanted in the accelerated phase
• Patients transplanted in the blast phase

You can use the curves to see patients' survival rate over time after the transplant.

• The vertical axis (the numbers going up the side) shows the survival rates (percent of recipients who survived).
• The horizontal axis (the numbers along the bottom) shows time measured from the day of transplant.
• The height of the curve at any point shows the survival rate at that point in time.

In this example, you can see that the survival rate at two years is about:

• 52% for patients transplanted in chronic phase.
• 30% for patients transplanted in accelerated phase.
• 15% for patients transplanted in blast phase.

The "p-value" (shown as P < 0.0001 on the example) is a statistical measure to help compare the groups shown in the different curves. The smaller the number, the stronger the evidence that survival rates are different for the different groups of patients. Often, a p-value less than 0.05 is considered good evidence, or "statistically significant." This means that we have enough evidence to conclude that the true survival rates are different among the groups shown. A p-value larger than 0.05 (not statistically significant) means the differences are not clear. It might mean that survival rates are the same, or it might mean that we do not have enough data to detect differences.

In this example, the p-value is very small (less than 0.0001), which means there is strong evidence to say that the survival rate is different depending on the patient's phase of CML at the time of transplant.