Health Risk Assessment and Causation

October 7 & 9 2025
Eric Delmelle

Chapter Overview

Defining Risk: Understanding risk as the probability of an event occurring, often implying harm.
Concepts of Causation: Reviewing causation from historical and contemporary perspectives.
Quantifying Association: Using ratio and difference measures to quantify associations.

Assessing Causal Effect: Applying attributable risk and its variants.
Competing Risks: Explaining how competing risks affect mortality patterns.
Risk Perception and Communication: Recognizing the critical role of risk perception and communication in public health.

Central Task: Assessing health risks core function in pop. health.
Broad Usage of “Risk”: The term describes the probability of adverse outcomes, risk factors, and health consequences.
Risk Factors vs. Risk Markers: Distinguishing between factors that cause disease and those that indicate its presence.
Health Risk Appraisal (HRA): Using quantitative procedures to estimate mortality and morbidity risks.
Health Risk Assessment: Evaluating scientific data on hazardous environmental agents ⇔ human exposure.

Hazard Identification: Identifying harmful agents and describing toxicity.
Dose-Response Assessment: Modeling exposure and health effect size variations.
Exposure Assessment: Identifying exposed populations, exposure routes, and dose characteristics.
Risk Characterization: Integrating data to estimate harm likelihood.

Flint resident lives in uncertainty after lead exposure youtube-link
Japan Declares Crisis As Fukushima Reactor Begins Falling Into Ocean And Radiation Levels Soar youtube-link

Dose-Response Assessment: Modeling exposure and health effect size variations.

Exposure Assessment: Identifying exposed populations, exposure routes, and dose characteristics.

Exposure Assessment: Identifying exposed populations, exposure routes, and dose characteristics.

Necessary Cause: A cause that must be present for an effect to occur.
Sufficient Cause: A cause that inevitably produces the effect.
Remote vs. Proximate Causes: Understanding causal chains and effect proximity.

Hume’s Criteria:
- Spatio-temporal Continuity – The cause and effect must occur close together in space and time.
- Temporal Priority – The cause must always come before the effect.
- Necessary Connection – There must be an expectation that the effect must follow the cause, though Hume argued this is something our minds impose rather than an inherent property of reality.

Mill’s Five Methods of Inductive Reasoning: Agreement, difference, joint method, residues, and concomitant variations.
- Agreement – If two or more cases of an event share only one common factor, then that factor is likely the cause of the event.
- Example: If multiple people who ate at a restaurant got food poisoning and the only dish they all ate was the salad, then the salad is likely the cause.
- Method of difference – If an event occurs in one case but not in another, and the only difference between them is one factor, then that factor is likely the cause.
- Example: If two people follow the same diet, but only one develops an allergy after eating a particular fruit, the fruit is likely the cause.

Mill’s Five Methods of Inductive Reasoning: Agreement, difference, joint method, residues, and concomitant variations.
- Joint method – A combination of the first two methods (1) Identify a common factor across cases where an event occurs, and (2) Ensure that when the factor is absent, the event does not occur..
- Example: If people who drink contaminated water get sick, and those who don’t drink it remain healthy, then the contaminated water is likely the cause.
- Method of Residues – If known causes are accounted for, and there is still an unexplained effect, the remaining factor(s) must be responsible for the residual effect.
- Example: If a medication has multiple ingredients and the effects of some ingredients are already known, but a new side effect appears, it is likely caused by the ingredient whose effects are still unknown.

Mill’s Five Methods of Inductive Reasoning: Agreement, difference, joint method, residues, and concomitant variations.
- Concomitant variations – If a factor varies in intensity along with an effect, there is likely a causal connection.
- Example: If increased smoking leads to increased lung cancer rates, and reduced smoking leads to lower rates, smoking is likely a cause of lung cancer.

Bradford Hill’s Criteria:
- 1 Strength & Consistency – A strong and repeated association across studies increases the likelihood of causation.
- 2 Specificity & Temporality – The cause should lead to a specific effect, and the effect must occur after the cause.
- 3 Biological Gradient & Plausibility – A dose-response relationship should exist, and the cause should make biological sense.
- 4 Coherence & Experiment – The evidence should align with existing knowledge, and experiments should support the causal link.
- 5 Analogy – If similar factors cause similar effects, the association is more credible.

Henle-Koch’s Postulates:
- 1. Agent Presence – The microorganism must be found in all cases of the disease.
- 1. Isolation – The microorganism must be isolated and grown in pure culture.
- 1. Disease Causation – Introducing the isolated microorganism into a healthy host should cause the disease.
- 1. Reisolation – The microorganism must be re-isolated from the newly infected host.
- 1. Consistency – These findings should be repeatable across multiple cases.

Ratio Measures: Relative risk, risk ratio, rate ratio, odds ratio.
Difference Measures: Risk difference or attributable risk.
Relative Risk (RR): \(RR = \frac{I_1}{I_0}\)
Odds Ratio (OR): \(OR = \frac{(a/c)}{(b/d)} = \frac{ad}{bc}\)
Risk Difference (RD): \(RD = I_1 - I_0\)
Attributable Risk Among the Exposed [AR(E)]: \(AR(E) = \frac{(I_1 - I_0)}{I_1} = \frac{(RR - 1)}{RR}\)
Population Attributable Risk (PAR): \(PAR = \frac{(I - I_0)}{I} = \frac{P(RR - 1)}{P(RR - 1) + 1}\)

Definition: Recognizes that individuals are exposed to multiple risks, where one event prevents another.
Importance: Essential for understanding mortality trends and disease impact.
Statistical Approaches: Methods to estimate death probability in competing risks.

Subjectivity of Risk: Public perception differs from expert assessments.
Social Construction of Risk: Influences include social, cultural, and political factors.
Factors Influencing Risk Acceptance: Fair distribution, benefits, alternatives, voluntary vs. imposed risks, and control.
Biases in Probability Assessment: Representativeness, anchoring, and availability.
Effective Risk Communication: Public involvement, planning, listening, honesty, coordination, media engagement, clarity, and compassion.