Home About Us Contact Us Index Search Consumer Topics Back Issues

Prescribing by numbers

Eve Hurley, Senior Editor, Australian Medicines Handbook, Adelaide

(Aust Prescr 2000;23:38)

The results of clinical studies are often presented in terms of the relative risk reduction achieved with an active treatment over a control. The relative risk reduction is usually expressed as a percentage and can appear impressive but, as it is isolated from the underlying incidence of the event being prevented, it has little value in the clinical situation.

Absolute risk reduction is the difference in event rates between active and control groups, but it can be difficult to visualise its clinical relevance. The reciprocal of the absolute risk reduction gives the number of patients who need to be treated to prevent one event. This is the number needed to treat and is a more useful measure which can be used to compare a range of interventions.¹

Calculations

Event rate	=	events in group number of subjects in group
Relative risk reduction %	=
Absolute risk reduction	=	event rate control - event rate active
Number needed to treat to prevent one event	=	___________1_________ absolute risk reduction

The results of the Helsinki heart study² (see box) were generally presented as a reduction of 34% in the incidence of coronary heart disease with gemfibrozil treatment.

Expressing results as the number of patients who need to be treated to prevent one event (or for one patient to benefit) is much more meaningful. It can be useful when discussing treatment options with patients.

Example Helsinki heart study Subjects: 4081 asymptomatic men aged 40-55 with dyslipidaemia (total cholesterol minus HDL >5.2 mmol/L). Treatment: gemfibrozil 600 mg twice daily (2051 men) or matched placebo (2030 men) in a five year randomised double-blind study. Results: number of events (fatal, non-fatal myocardial infarction or cardiac death) gemfibrozil - 56 events, placebo - 84 events. Calculations
Event rate placebo =	_84_ 2030	= 0.041 (4.1%)
Event rate active =	_56_ 2051	= 0.027 (2.7%)
Relative risk reduction % =		= 34%
Absolute risk reduction =	0.041 - 0.027	= 0.014 (1.4%)
Number needed to treat for five years to prevent one event =	__1__ 0.014	= 71 men

R E F E R E N C E S

1. Cook RJ, Sackett DL. The number needed to treat: a clinically useful measure of treatment effect [published erratum appears in Br Med J 1995;310:1056]. Br Med J 1995;310:452-4.

2. Frick MH, Elo O, Haapa K, Heinonen OP, Heinsalmi P, Helo P, et al. Helsinki Heart Study: primary-prevention trial with gemfibrozil in middle-aged men with dyslipidemia. N Engl J Med 1987;317:1237-45.