[Durgical](https://durgical.com/en/) › [DurgiPedia](https://durgical.com/en/durgipedia/) › [Epidemiology](https://durgical.com/en/epidemiology/) › [Bias, confounding and causation](https://durgical.com/en/epidemiology/areas/bias/) › Causal models: necessary, sufficient and component causes

# Causal models: necessary, sufficient and component causes

- Concept

Almost no disease has a single cause: several pieces act together. Removing one piece prevents every case that needed it.

Updated Nov 13, 2026

## Key points

- Necessary cause: no disease without it (the agent in an infection); it is almost never sufficient on its own.
- Sufficient cause: a complete set of components that produces the disease; one disease may have several.
- Removing a single component prevents every case from the sufficient causes that contain it.
- The preventable fractions of different components can add up to more than 100%: each case needs several components at once.
- In a DAG: confounder = common cause (adjust); mediator = step on the path; collider = common effect (do not adjust).

## What it is

A necessary cause must be present for the disease to occur; a sufficient cause is a set of factors that together inevitably produce it. The sufficient-component cause model (the ‘causal pies’) draws each sufficient cause as a pie made of several slices, and one disease can have several pies. Causal diagrams (directed acyclic graphs, DAGs) use arrows to show what causes what and help decide which variables to adjust for and which to leave alone.

## Worked example

- A town has 60 cases a year of an infection. Agent A is present in every case.
- 40 cases arise from sufficient cause I (A + malnutrition + crowding) and 20 from sufficient cause II (A + diabetes).
- Fixing crowding breaks pie I: 40 of 60 cases are prevented, 67%.
- Controlling diabetes breaks pie II: 20 of 60 are prevented, 33%.
- Removing A, the necessary cause, prevents all 60 cases: 100%.
- Malnutrition also prevents 40 (67%). The fractions add to 67 + 67 + 33 + 100 = 267%, more than 100%.
- Reading: each case has several causes at once, and acting on any slice of a pie prevents it.

## Combinations of necessary and sufficient cause

| Type | What it means | Example |
| --- | --- | --- |
| Necessary and sufficient | Always present and enough on its own | Very rare in medicine |
| Necessary, not sufficient | Always present, but needs other factors | The tubercle bacillus for tuberculosis |
| Sufficient, not necessary | Enough on its own, but other pathways exist | Rare; mainly a theoretical model |
| Neither necessary nor sufficient | One component among several | Smoking for lung cancer |

## Markers

- Necessary cause
- Sufficient cause
- DAG
- Collider

## In practice

- In a DAG every arrow runs from cause to effect and there are no loops; draw it before the analysis, from what is known about the topic.
- Adjusting for a confounder (a common cause of exposure and outcome) closes a false path; adjusting for a collider opens one.
- Adjusting for a mediator removes part of the effect being measured when the question is about the total effect.

## Common pitfalls

- Thinking a ‘weak’ cause (low RR) matters little: if it is a common component, its preventable fraction can be large.
- Adjusting for every available variable: some are mediators or colliders and bias the result.
- Confusing necessary cause with main cause: the infectious agent is necessary, but prevention can act on another component.

## Memory aid

A causal pie bakes only when every slice is in the tin: pull out any slice and that pie never comes out of the oven.

## Test yourself

Think of the answer, then tap the question to see it.

## What is a necessary cause?

A factor that must be present for the disease to occur; without it there are no cases, although on its own it is usually not enough.

## Why can the preventable fractions of several factors add up to more than 100%?

Because each case needs several components at once; removing any of them prevents the same case, which is counted in each fraction.

## In a DAG, what is a collider and what happens if you adjust for it?

A variable that arrows from both exposure and outcome point into (a common effect). Adjusting for it creates a spurious association.

## What is an example of a necessary but not sufficient cause?

The tubercle bacillus: there is no tuberculosis without it, but most infected people never become ill.

[Add to my cards (4)](https://durgical.app/estudiar/?add=/en/epidemiology/causal-models-necessary-sufficient-and-component-causes/)

## Connections

## Related 6

- [Bradford Hill criteria for causation](https://durgical.com/en/epidemiology/bradford-hill-criteria-for-causation/)
- [Chance, bias, confounding or cause](https://durgical.com/en/epidemiology/chance-bias-confounding-or-cause/)
- [Confounding](https://durgical.com/en/epidemiology/confounding/)
- [Controlling confounding](https://durgical.com/en/epidemiology/controlling-confounding/)
- [Additive and multiplicative interaction](https://durgical.com/en/epidemiology/additive-and-multiplicative-interaction/)
- [Population attributable risk](https://durgical.com/en/epidemiology/population-attributable-risk/)

[← All of epidemiology](https://durgical.com/en/epidemiology/)

Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.

Sources: [Principles of Epidemiology | Lesson 1 - Section 8](https://archive.cdc.gov/www_cdc_gov/csels/dsepd/ss1978/lesson1/section8.html) (Centers for Disease Control and Prevention, 2012); [Collider bias | Catalog of Bias](https://catalogofbias.org/biases/collider-bias/) (Catalogue of Bias Collaboration, 2019); [Confounding | Catalog of Bias](https://catalogofbias.org/biases/confounding/) (Catalogue of Bias Collaboration, 2018).

## **Spotted a mistake?** We are always improving for the community. Tell us what is wrong and our team will review it. Report a mistake

Representative model, for educational purposes only.

---
Source: https://durgical.com/en/epidemiology/causal-models-necessary-sufficient-and-component-causes/
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Cite as: DurgiPedia (durgical.com), https://durgical.com/en/epidemiology/causal-models-necessary-sufficient-and-component-causes/