Nelson-Aalen Cumulative Hazard Estimator
Also known as: Nelson-Aalen cumulative hazard, Aalen estimator, empirical cumulative hazard, Nelson-Aalen kümülatif hazard tahmincisi
The Nelson-Aalen estimator is a non-parametric estimator of the cumulative hazard function from right-censored time-to-event data. Developed by Wayne Nelson for reliability hazard plotting in 1972 and placed on a rigorous counting-process foundation by Odd Aalen in 1978, it accumulates the ratio of observed events to the number at risk at each event time, providing the natural hazard-scale companion to the Kaplan-Meier survival curve.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use the Nelson-Aalen estimator when you want to describe accumulated risk on the hazard scale from right-censored survival data without assuming a parametric distribution. It is the standard way to obtain an empirical cumulative hazard for diagnostic plotting — for example, checking whether a hazard is roughly constant (a straight Nelson-Aalen line suggests an exponential model) or assessing proportional-hazards assumptions before fitting a Cox model. It assumes non-informative (independent) censoring and is most informative when the number at risk remains reasonably large across the follow-up. When the quantity of interest is the survival probability itself, report Kaplan-Meier; when it is the hazard or cumulative hazard, Nelson-Aalen is the natural estimator.
Strengths & limitations
- Non-parametric: makes no assumption about the shape of the underlying hazard or survival distribution.
- Provides the cumulative hazard directly, which is the natural scale for diagnostics and for checking parametric model fit.
- Handles right-censoring correctly through the at-risk set, using all observed follow-up time.
- Has a clean counting-process variance, giving pointwise confidence intervals and a basis for group comparisons.
- Only describes the data; it does not model the effect of covariates, for which a Cox or parametric model is needed.
- Relies on the assumption of non-informative censoring; dependent censoring biases the estimate.
- Becomes unstable in the tail when the at-risk set is small, so late-time estimates carry wide uncertainty.
- The exponentiated survival curve, while valid, is non-standard for reporting survival probabilities compared with Kaplan-Meier.
Frequently asked
How does Nelson-Aalen relate to Kaplan-Meier?
They are complementary non-parametric estimators from the same right-censored data. Kaplan-Meier estimates the survival probability S(t) via a product over event times; Nelson-Aalen estimates the cumulative hazard H(t) via a sum of dᵢ/nᵢ. They are linked by S(t) = exp(−H(t)), and the two survival curves are asymptotically equivalent.
When is Nelson-Aalen preferred over Kaplan-Meier?
When the cumulative hazard is the quantity of interest — for diagnostic hazard plots, checking proportional hazards, or reliability hazard analysis. The Nelson-Aalen-based survival curve can also be marginally more stable in very small samples, though Kaplan-Meier remains standard for reporting survival probabilities.
Does it require a distributional assumption?
No. It is fully non-parametric, making no assumption about the hazard or survival shape. That is precisely why a straight Nelson-Aalen plot is informative — it suggests a constant hazard (exponential model) rather than being imposed by one.
What does the slope of the cumulative hazard mean?
The slope of Ĥ(t) approximates the instantaneous hazard rate at that time. A steepening curve indicates rising hazard, a flattening curve indicates falling hazard, and a straight line indicates a roughly constant hazard.
Sources
- Nelson, W. (1972). Theory and applications of hazard plotting for censored failure data. Technometrics, 14(4), 945–966. DOI: 10.1080/00401706.1972.10488991 ↗
- Aalen, O. (1978). Nonparametric inference for a family of counting processes. Annals of Statistics, 6(4), 701–726. DOI: 10.1214/aos/1176344247 ↗
How to cite this page
ScholarGate. (2026, June 2). Nelson-Aalen Cumulative Hazard Estimator. ScholarGate. https://scholargate.app/en/survival/nelson-aalen
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Cox RegressionSurvival↔ compare
- Frailty ModelSurvival↔ compare
- Kaplan-MeierSurvival↔ compare
- Log-Rank TestSurvival↔ compare
- Weibull RegressionSurvival↔ compare