The Weibull distribution models time-to-failure. In reliability engineering, that means component lifetimes. In medicine, it means survival times. I have been working with Weibull models for my thesis on series system reliability. Then I got diagnosed with cancer, and now every time I work with survival curves, I am looking at mathematical abstractions of something very concrete: how long until failure?
The Mathematics
The Weibull CDF:
F(t) = 1 - exp(-(t/λ)^k)
Two parameters:
- λ: scale (characteristic lifetime)
- k: shape (how failure rate changes over time)
The shape parameter k tells you the whole story:
k < 1: Decreasing hazard. If you survive early on, your risk goes down. This is the infant mortality pattern.
k = 1: Constant hazard. Memoryless. This is just the exponential distribution.
k > 1: Increasing hazard. Things wear out.
The Hazard Function
The hazard function is what makes Weibull useful for survival analysis:
h(t) = (k/λ)(t/λ)^(k-1)
This is the instantaneous failure rate: given that you have survived to time t, what is the probability you fail in the next instant?
For cancer, this is the number that matters. Some cancers have increasing hazard (the longer you have it, the worse things get). Others have decreasing hazard after initial treatment, meaning if you make it past the critical period, prognosis improves. Knowing which pattern applies to your disease changes how you think about time.
Personal Context
When you study survival analysis academically, it is abstract. When you are living it, every curve is personal.
I look at Kaplan-Meier plots and see myself somewhere on that curve. I work with hazard functions and think: is my k > 1 or k < 1? Am I in the wearing-out regime or the if-you-make-it-past-this-it-gets-better regime?
The math does not change. But the meaning does.
The Irony
I chose reliability engineering for my thesis before the cancer diagnosis. I was studying component failures in series systems, where if any one part fails, the whole system fails.
Then I became a series system. Organs, treatment response, immune function. All have to work. Failure of any one is catastrophic.
The mathematics I was studying abstractly became uncomfortably literal.
Why Weibull Matters
In engineering, Weibull helps you predict when systems will fail, schedule preventive maintenance, design redundancy, optimize replacement schedules.
In medicine, it helps you estimate survival probabilities, compare treatment efficacy, identify risk factors, plan clinical trials.
Same math. Different substrates.
Censored Data
Both domains deal with censored data: observations where you know someone (or something) survived to time t, but not when failure actually occurs.
In engineering, tests end before all components fail. In medicine, studies end before all patients die.
The mathematical treatment is identical. Maximum likelihood estimation with likelihood contributions from:
- Failures: contribute density f(t)
- Censored observations: contribute survival S(t)
The abstract machinery handles both cases without caring which one you are talking about.
The Weibull Assumptions
Weibull assumes a specific relationship between time and failure risk. It is flexible, but not universal. Some cancers do not follow Weibull patterns. Some failure mechanisms are more complex. The model is useful but approximate.
Living with cancer while studying survival analysis teaches you something the textbooks skip: models are maps, not territory. The math gives you probabilistic guidance. Reality gives you a sample of size one.
Connection to My Thesis
My thesis work uses Weibull models for series systems with masked failure data. You know the system failed, but not which component caused it.
Cancer sometimes feels like this. Something failed, but identifying the exact mechanism is hard. Genetic? Environmental? Random? The cause is masked.
The EM algorithm I use for maximum likelihood estimation handles this ambiguity mathematically. It does not remove uncertainty. It quantifies it.
Why I Keep Working
People ask why I am continuing with the thesis while dealing with cancer.
Because the work is meaningful regardless. Mathematics is reliable when biology is not. Contributing to survival analysis feels appropriate. The distraction helps. And if I am going to be a data point in someone's survival curve, I might as well understand the mathematics.
The Broader Lesson
Weibull distributions teach you that failure is statistical, not deterministic. Populations have distributions; individuals have outcomes. Hazard rates change over time. Survival to time t changes your conditional probabilities. Mathematics can describe but not control fate.
These lessons apply to components and to people.
