Interpreting health data and risk perception
Studying this course? Keep your progress in one place.
Key facts
- Morbidity is illness; mortality is death; incidence is new cases; prevalence is all cases.
- Compare rates (such as per 100 000) rather than raw numbers when populations differ in size.
- Correlation does not prove causation; look for confounding variables, a mechanism and consistent evidence.
- Evaluate studies for sample selection and size, validity, reliability, duration and a control group.
- Perceived risk often differs from actual risk.
- Rate per 100 000
Illness and mortality data
Use rates, such as cases per 100 000, so that populations of different sizes can be compared fairly.
Morbidity is illness; mortality is death. Incidence is the number of new cases in a given time; prevalence is the total number of cases at a point in time. Compare rates, not raw numbers, because populations differ in size.
- Rate per 100 000
Worked example
There are 900 deaths in a population of 3 000 000. The earlier rate was 50 per 100 000. Find the new rate and the percentage change.
- 1
Rate = 900 ÷ 3 000 000 × 100 000 = 30 per 100 000.
- 2
Change = 30 − 50 = −20.
- 3
Percentage change = −20 ÷ 50 × 100.
Why compare rates rather than numbers of cases?
Correlation and causation
A correlation is only a relationship; a confounding variable may affect both factors, so causation needs stronger evidence.
A correlation is a relationship between two variables. It does not prove causation, because a confounding variable such as age or smoking may affect both.
Causation is more likely if there is a known mechanism (LDL in plaques), a dose-response relationship, consistent results from several studies, and controlled trials showing a change. Conflicting evidence arises from differences in sample, method, control of confounders and bias; weigh many studies, preferably with peer review.
- 1
Confounding variable
For example age or smoking.
- 2
Affects both factors
It influences factor X and outcome Y.
- 3
X and Y are correlated
They change together.
- 4
No proof of causation
Look for a mechanism, dose-response and controlled trials.
Correlation
- A relationship between two variables
- Does not show one causes the other
Evidence for causation
- Known mechanism
- Dose-response relationship
- Consistent studies and controlled trials
A study finds a correlation between coffee drinking and CVD. What should you say?
Evaluating study design
Check the sample, sample size, validity, reliability, duration and whether there is a control group.
A sample should be random and representative (age, gender, ethnicity) to avoid bias. A larger sample reduces the effect of chance. Validity means the study measures what it claims to, with confounders controlled and accurate data (self-reported diets may not be). Reliability means similar results when repeated. Long enough duration and a control group allow comparison.
Feature
- Sample selection
- Sample size
- Validity
- Reliability
- Duration and control group
What to check
- Random and representative
- Larger reduces the effect of chance and anomalies
- Measures what it claims to; confounders controlled
- Similar results when repeated
- Long enough, with a group for comparison
A study gives the same result each time it is repeated. Which quality is this?
Perceived and actual risk
Actual risk is calculated from data, but people judge risk by media coverage, familiarity, control and personal experience.
Actual risk is calculated from data on incidence and mortality. Perceived risk is a person's own judgement of how likely harm is.
The media give most attention to rare, dramatic events. People underestimate risks that are familiar or that they feel they control (diet, smoking) and overestimate involuntary or frightening ones. Personal experience, such as a relative with CVD, changes perception. Campaigns must give clear, accurate data.
- 1
Actual risk
Calculated from incidence and mortality data.
- 2
Influences on perception
Media coverage, familiarity, feeling of control, personal experience.
- 3
Perceived risk
A personal judgement, over- or underestimating the true risk.
- 4
Campaigns
Must give clear, accurate data.
Actual risk
- Calculated from incidence and mortality data
- The true likelihood of harm
Perceived risk
- A person's own judgement
- Shaped by media, familiarity, control and experience
Why might people underestimate the risk from smoking?
Try an exam question
A study finds a correlation between salt intake and CVD. Explain why this does not prove that salt causes CVD, and how the evidence could be strengthened.
[4 marks]
- [1]A correlation is only a relationship between two variables and does not prove causation.
- [1]Confounding variables such as age or smoking may affect both salt intake and CVD.
- [1]Evidence is stronger with a known mechanism, for example salt raising blood pressure.
- [1]Evidence is stronger with consistent results from several studies or controlled trials.
That's the notes covered.
Carry on to the next subtopic.
Try a question
What is meant by the term mortality?
1 mark, marked against the mark scheme
What is meant by the term mortality?
1 mark, marked against the mark scheme