Survival analysis from cancer registries, explained
25 September 2026 · Cohira team
Population cancer registries follow patients for decades. That follow-up is the raw material of survival analysis.
The Kaplan-Meier curve
The share of patients still alive at each time point, drawn as a step function, with censoring handled honestly. Stratify by stage and the clinical story is usually visible before any model runs.
Competing risks
An elderly cohort dies of many things. When death from other causes competes with the cancer, cause-specific survival needs competing-risks methods, or the estimates flatter the disease.
From model to bedside
A Cox or Fine-Gray model becomes a nomogram: a one-page tool that turns a patient's characteristics into a predicted probability. Validated and reported to TRIPOD, it is among the most cited outputs a registry study can produce.