Understanding Risk in Epidemiology: Cumulative Incidence, Ratios, and Differences Explained

Mayta
Apr 30
2 min read

Updated: May 7

🔍 What Is Risk in Epidemiology?

📌 Fundamental Definition

Risk is a type of probability. It answers:👉 “Among those initially at risk, what proportion experienced the outcome over a defined period?”

Formally:

This is also called cumulative incidence, because it accumulates cases over time in a closed cohort (no one enters/leaves except through the event).

🧠 Derive the Formula from the Concept

Instead of memorizing, let’s reason it out:

🧮 Cumulative Incidence (a.k.a. Risk)

Imagine you’re tracking 100 patients for 1 year. If 10 develop hypertension:

The key is this:

Numerator = count of new events.
Denominator = count of people who could have the event at baseline.

So we define:

Where:

E = number of new events (e.g., myocardial infarctions),
N = people at risk at the beginning.

⚖️ Risk Ratio (Relative Risk)

When comparing two groups (e.g., smokers vs. non-smokers):

Where RR is the risk.Think of this as asking: “How many times more likely is the event in the exposed group?”

From the slide:

➕ Risk Difference (Absolute Risk Reduction)

Instead of asking how many times, we ask how many more or fewer events per person:

Clinical interpretation:

A positive RD → increased risk with exposure (e.g., smoking).
A negative RD → protective effect (e.g., vaccine).

🔧 Not All Ratios Are Risks

The slides clarify:

Prevalence is cross-sectional, a snapshot:

No time element; not a measure of incidence.
Odds and Hazards are other forms of risk-like concepts, but distinct. (We’ll get there!)

💡 Key Takeaways

Risk = cumulative incidence, interpreted as the probability of developing an outcome over time. Risk = Incidence = Probability
It's only meaningful among people initially at risk (e.g., someone without the disease at baseline).
Risk Ratio (RR) shows how many times more likely an outcome is in one group vs. another.
Risk Difference (RD) shows the absolute difference in risk, useful for estimating public health impact or treatment benefit.