Preventing coronary heart diseaseBMJ 2006; 332 doi: https://doi.org/10.1136/bmj.332.7542.617 (Published 16 March 2006) Cite this as: BMJ 2006;332:617
- Rod Jackson, professor (email@example.com),
- John Lynch, professor,
- Sam Harper, research fellow
- Section of Epidemiology and Biostatistics, School of Population Health, Faculty of Medical and Health Sciences, University of Auckland, Private Bag 92019, Auckland, New Zealand
- Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, QC, Canada H3A 1A2
In this issue (p 659) Manuel and colleagues report how they estimated the effectiveness of three strategies to lower blood cholesterol concentrations in Canadians adults.1 A “population” strategy assumed that blood cholesterol could be lowered by 2% in the whole population and deaths from coronary heart disease by 2.7%. The two other strategies were patient based, assuming that prescribing statins to subgroups of people at high risk of coronary heart disease would reduce their risk by 27%. A “single risk factor” strategy targeted patients with blood cholesterol levels greater than 6.2 mmol/l and a “baseline risk” strategy targeted those with a baseline risk of cardiovascular disease greater than 15% over five years, irrespective of their blood cholesterol levels.
Surprisingly, the population strategy, based on the axiom by British epidemiologist Geoffrey Rose that “a large number of people at small risk may give rise to more cases than a small number of people at high risk,”2 was the least effective strategy. The single risk factor strategy, prescribing statins to the 1.53 million Canadians with a blood cholesterol concentration greater than 6.2 mmol/l, was calculated to prevent three times as many coronary deaths as lowering cholesterol by 2% in all 12 million Canadian adults.
By far the most effective strategy, which is estimated to prevent seven times more coronary deaths than the population strategy, was (ironically) based on another of Rose's axioms, that “all policy decisions should be based …