H&SB4W123

Página 1

Putting a figure on risk

Risk = probability that an event will occur during a specified time (a quantified uncertainty)

Reasons to be interested in risk:

- There isn't much certainty about it in medicine.

- We generally believe that quantifying uncertainty is important.

- e.g. will this operation make me better / will I have a healthy baby.

Options for 'quantifying' risk:

Risk

1 in 85 (natural frequency)

1/85 = 0.012 (probability)

1.2% (percentage)

Odds

Ratio of probability that will to the probability that won't

1/85 / 85/85

= 1 to 84 or 1/84 or 0.012.

Hence, in a group of people:

Risk of thing = number who get thing / number in group In a specified time period.

How to compare and contrast risk

Study population who are 'free of disease'

Then measure and classify 'participants on the pill, participants not on the pill'

Then measure outcome/compare rates of cervical cancer on participants with or without pillage.

(Watch this lecture the bit on calculating risk ratio is important)

Risk difference tells us something about 'impact'

Rist ratio tells us something about 'strength' of relationship.

There are some circumstances where we can't calculate a risk ratio - in particular case-control study.

In this case we are able to use odds ratio.

Página 2

Relative and absolute

Risk ratios and odds ratios are relative measures.

The risk of cervical cancer amongst OCP users it twice that of non-users

The risk of cervical cancer amongst OCP users is infinitely higher than in male OCP users.

Be clear about what group is your reference group.

Risk and odds themselves and risk differences are absolute measures.

- They can be interpreted without reference to any other group

- They are measured directly on the 'scale' of risk from 0 to 1.

Estimated risk ratio for cervical cancer amongst women taking oral contraceptives for 10 year compared to non-takers = 2.2

But absolute risk of cervical cancer is low.

Página 3

Risk ratio is the same as 'relative risk'

Risk difference is sometimes the same as 'attributable risk'

Other terms out there:

- Rate ratio

- Numbers needed to treat

- Population attributable risk

Patients/People are more willing to accept risks that are

- Voluntary

- Controllable

- Familiar

- Non-catastrophic

- 'Fair' in the way benefit and risk distributed (i.e. more risk = more possible benefit)

And vice versa for the opposite.

Unknown:

- Unobservable

- Unknown

- New

- Delay before harm

Dread:

- Perceieved lack of control

- Dread

- Catastrophic potential

- Fatal consequences

- Inequitable distribution of risks and benefits

Communication

Language: try and avoid descriptive terms.

Presentation of quantified risk

Develop trust

- Risk of vaccine related illness may be low, but do you 'believe' what you're told.

- Decision aids

- Practice

Summary:

Some risks can be quantified.

Differences in risk between groups can be summarised by: risk difference, risk ratio (relative risk), odds ratio.

Risk can be expressed in relative and absolute terms.

We interpret risk differently.

We are working at ways of communicating risk more effectively.