Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

On Risk

Risk perception and presentation
Results
Frequency presentation
Examples of risk presentation
Acceptable baseline risk
Perspective
Comments

Bandolier has visited the issue of risk on several occasions over the years. Risk is very important, and very serious, as much as anything because communicating risk is a difficult business. We showed in Bandolier 128 how patients were unable to appreciate risk communicated to them in words and numbers, and consistently over-estimated risk, by upwards of 400 times for rare risks. A few more thoughts on risk, then.

Risk perception and presentation


A study [1] was carried out on two groups, 38 graduate students and 47 healthcare professionals. A hypothetical situation about adverse events of an influenza vaccine was presented to them in either a probability format (5%), or a frequency format (1 in 20). Randomisation was by alternation in questionnaire handouts.

The questionnaire asked whether they would be prepared to receive a vaccine if the risk of fever and headache within seven days was either 5% (one group) or 1 in 20 (the other). A second question asked participants to match frequency with one of six phrases, from very common to very rare.

Results


There was no difference between occupation, age, or sex of the groups receiving information as probability or frequency. About 60% of participants would have elected to have the influenza vaccine, without any significant difference between a probability format (67% electing to receive it) and the frequency format (55%).

There were differences between the way in which the risk was matched to phrases (Figure 1). In both presentations, the same risk was labelled as very common, through to rare. Presentation as frequency (1 in 20) resulted in much greater consensus, with 84% happy that this could be called common or occasional, and only 9% considering it either rare or very common.


Figure 1: How graduates and professionals label the same adverse event presented in different ways






Frequency presentation


Using frequency presentations seems to make sense as a method that is intuitively sensible to most of us. Gigerenzer and Edwards [2] emphasise that frequency statements should always state a reference class (patients like us, inhabitants of Wales, people with a previous heart attack). They also give a number of examples of where patients and professionals can be misled by being given information as percentages, or those outputs of research so beloved by statisticians (relative risk, sensitivity, specificity). That which we cannot immediately understand and use is likely to be a cause of great mistakes.

John Paling [3] has come up with a series of useful ways of helping to communicate risk to patients, which is now expanded into a terrific book [4]. The article, and especially the book, are full of superb examples of various risks, and how to display them using visual presentations of natural frequencies. The book is well thought out, and comes with seven simple strategies for successful communication, and is packed with good sense. No brain ache, and lots of humour.

Examples of risk presentation


A few examples of how to present risk are always handy. Taking Gigerenzer & Edwards to heart, perhaps it is helpful to begin with some background information about risks. Table 1 therefore compiles US data from 2002 on the risk of dying, from two good websites [5,6].


Table 1: Annual and lifetime risk of death from different causes in the USA in 2002



Annual
Lifetime
Top 15 causes of death in USA in 2002
Diseases of the heart
415
5
Malignant neoplasm
515
7
Cerebrovascular disease
1786
23
Chronic lower respiratory disease
2273
29
Accidents
2703
35
Diabetes mellitus
4000
52
Influenza and pneumonia
4348
56
Alzhemier's disease
5000
65
Nephriotis, nephrotic syndome, nephrosis
7143
92
Septicaemia
8333
108
Suicide
9091
118
Chronic liver disease, cirrhosis
11111
144
Primary hypertension, hypertensive renal disease
14286
185
Parkinson's disease
16667
216
Pneumonitis due to solids or liquids
16667
216
External causes of death in USA in 2002
Any transport accident
5953
77
Car occupant
17625
228
Air and space
440951
5704
Any fall
17712
229
Fall from chair or bed
366804
4745
Fall on steps or stairs
180188
2331
Firearms discharge
377876
4888
Accidental drowning
83534
1081
Smoke, fire, flames
91149
1179
Lightning
4362746
56439
Cataclysmic storm
4570496
59127
Accidental poisoning
16407
212



Given that there is going to be a lot of lumping going on, it does give some useful ideas about what is important. For instance, it was interesting, given discussions in the UK right now about nuclear power, to discover that the number of deaths from radiation in the USA in 2002 was zero. Nor is the list necessarily fair because it is all about death, and does not reflect other issues. Chronic pain has the largest negative impact of quality of life, but because it does not kill people directly (Bandolier 83; it is a symptom, not a diagnosis), it does not figure in the list. There is more to life than death.
Accidents, for which a few figures out of many are given, are also interesting. It is more risky travelling by car than by air, but which gets the headlines?

Acceptable baseline risk


Actually, road traffic deaths are important. In the USA total road traffic accidents killed about 35,000 people in 2002 (compared with about 3,500 in the UK). While as a society and as individuals we strive to reduce this (breath tests, seat belts, air bags, road design), we accept it. We have to, otherwise we wouldn't travel at all. The figure of about 1 in 17,000 for death seems to be a limit of acceptability. We become uncomfortable with more likely risks.

And while we concentrate on death, we forget that for every death, there are 10 people who are seriously injured, and 77 others who have some lesser injury. So of the 60 million people in the UK, 300,000 have some injury travelling on the roads. That is 1 in 200 of us, every year; over a lifetime of 70 years, that makes the risk about 1 in 3. So at another level, travelling on the roads in the UK is an incredibly risky business. It should be banned. Back to Gigerenzer, and stressing the importance of setting.

Perspective


For some perspective on all of this, it is time to visit the Paling Perspective Scales again for some examples. Bandolier is grateful to John Paling for permission to reproduce the scales. Three examples follow:

Using information from John Paling's book, we reproduce information for the risk of a Down's syndrome baby at term according to maternal age. This is done in two ways, using the straight perspective scale with its logarithmic scale and words (Figure 2), and using the 1000-woman palette (Figure 3), showing, for each age, how many of those 1000 women might have a child affected by Down's syndrome.


Figure 2: Risk of Down's syndrome in a term infant for mothers of different age, using the Paling Perspective Scale


Here the risks are presented visually using a logarithmic scale, associated with verbal descriptors, and with numbers presented as natural frequencies, so that three different presentations are shown in the one graph. No contextual examples are included here, but might be in other presentations.






Figure 3: Risk of Down's syndrome in a term infant for mothers of different age, using the one thousand woman palette


This example has the same information as in Figure 2, but presented differently. There are 1000 small representations of women on the palette (others are available for men or families), and the number who might be affected are coloured with a pen or highlighter.

In Figure 3, as a demonstration, maternal ages are shown from 30 years and below (one woman or fewer affected), to 49 years where 91 women would be affected. Whether using all these ages, or only the one relevant for an individual is better is not known (to Bandolier, anyway), nor is it known whether this representation would be more or less acceptable than that in Figure 2.

The 1000-person palette may be useful in demonstrating the balance between benefits and harms of treatment. Figure 4 shows the example of low dose aspirin in people with a previous heart attack, and Figure 5 that of anticoagulant versus aspirin for AF in people at high risk of stroke. Benefits (heart attack or stroke avoided) and risk (bleeding events) have been annualised, and the results presented in words as well.





Low dose aspirin has benefits and harms. The benefits include reducing heart attacks, but the harm involves gastrointestinal bleeding, which can be serious and result in death in 1 in 10 patients who suffer it. Bandolier used information from systematic reviews relating to people with a previous heart attack, and annualised the data. The 1000-person palette is used to show benefit and risk (Figure 4).


Figure 4: Low dose aspirin after a heart attack






Warfarin or aspirin for patients with nonvalvular atrial fibrillation at moderate or high risk of stroke? Annualised data from a systematic review (Bandolier 108) is used on the 1000-person palette to show benefits and harms differences, in terms of heart attacks and strokes avoided with anticoagulant, but at the price of some more bleeds (Figure 5).


Figure 5: Anticoagulant versus aspirin for AF






Comments


These pictorial representations lose something reduced in print in a single colour. On screen they look great, and should be useful tools. The problem we have though, despite our enthusiasm, is that we are all of us beginners in the business of describing risk and knowing whether different formats make a difference to how patients and professionals perceive and react to risk information. The literature is silent, or at best very, very, quiet. We need to get some good information together, prepare some presentations, and trial it.

In the meantime, suggestions for areas where these types of presentations might be helpful would be welcomed. Bandolier will try and find some data.

References:



  1. SB Tan et al. Risk perception is affected by modes of risk presentations among Singaporeans. Annals Academy of Medicine of Singapore 2005 34: 184-187.
  2. G Gigerenzer, A Edwards. Simple tools for understanding risks: from innumeracy to insight. BMJ 2003 327: 741-744.
  3. J Paling. Strategies to help patients understand risks. BMJ 2003 327: 745-748.
  4. J Paling. Helping Patients understand risks. (ISBN 0-9642236-0-0), available from riskcomm.com
  5. National Vital Statistics Reports 2005 53: 15 (www.cdc.gov/nchs)
  6. What are the odds of dying? National Safety Council (www.nsc.org/lrs/statinfo/odds.htm)

previous or next story