Skip navigation

Making tough decisions

It can't happen to me

All of us make decisions about things we do which have risk attached to them. We might choose to dive, or go hang-gliding, or ride a motorcycle, or smoke. Yet we are certainly aware that there are risks attached to all of these activities. If we were rational beings we might go through a process which included:

But usually we don't. People show a consistent tendency to claim they are less likely than their peers to suffer harm, which is perhaps why apparently sensible people go diving and hang gliding, drive motorcycles, and smoke.

A review [1] of the literature on perceived risks is an enlightening, if none too easy read. It identifies unrealistic optimism as being a major contributor to why people do not use precautionary behaviours - with an underestimate of one's own risk and an overestimate of the risk of others as being contributory to this.

It argues that providing risk information is generally not sufficient to change behaviour. Other factors, such as the efficacy and costs of preventative behaviour, social pressure and perceived self-efficacy play a major role in helping people to change their behaviour. It is just not enough to give people the facts.

Giving the facts

But the way in which facts are given can make a big difference to an individual's choice of medical treatments [2]. A group of 100 outpatients were given information about cholesterol lowering and hypertension treatment (Helsinki heart study), and told that the medicine was free of side effects and that the treatment would have no cost to them.

Information was given in different formats, in the form of easily understood written statements. How this was done (it is given in an Appendix) was interesting in itself. But the formats were equivalent to relative risk reduction, absolute risk reduction, number needed to treat, average gain in disease free years in an average or a stratified format. The NNT format they chose was:

"Studies of a cholesterol-lowering pill showed that if 71 people took it for an average of just over 5 years, the medicine would prevent one of the 71 from having a heart attack. There is no way of knowing in advance which person that might be. Two people of the 71 would have heart attacks anyway, even though they took the pill."

and the relative risk reduction format was:

"A cholesterol-lowering pill was studied to see how it worked in reducing coronary artery disease (the disease that causes heart attacks). Persons in the group treated with this medicine had 34% fewer heart attacks than the non-treated group."

Giving the same information in a number of easily understood ways led to dramatic differences in the response of participants. For instance, for cholesterol lowering, 83% of participants would have the drug when data were presented as relative risk reduction, but only 31% when it was presented as number needed to treat.

Format % Agreeing
Relative risk reduction 88
Absolute risk reduction 42
Number needed to treat 31
Gain in disease free years 40
Stratified gain 56


Bandolier is continually surprised that more has not been done to understand what underlies the effective communication of information. Why bother to get information if we can't use it properly? These are important tools in helping people make decisions about therapy or lifestyle changes. Particularly where the NNTs are high, as in lowering of blood pressure, or cholesterol, where one might consider that we are treating a population as much as a single person, it is important to recognise the effects that data presentation has. Whether people are being realistic or optimistic, they have views too. Perhaps we should do more studies of data presentation as well as of drugs and other interventions.


  1. J van der Pligt. Perceived risk and vulnerability as predictors of precautionary behaviour. British Journal of Health Psychology 1998 3: 1-14.
  2. JE Hux, CD Naylor. Communicating the benefits of chronic preventive therapy: does the format of efficacy data determine patients' acceptance of treatment? Medical Decision Making 1995 15:152-7.
previous story in this issue