The ground or the goal posts?BMJ 2006; 332 doi: http://dx.doi.org/10.1136/bmj.332.7542.0-f (Published 16 March 2006) Cite this as: BMJ 2006;332:0-f
- Fiona Godlee, editor ()
To anyone who has been a medical student, Geoffrey Rose is perhaps best known for saying that a small shift in risk across a whole population will prevent more deaths than treating only those at high risk. The question taxing some of our authors this week is whether this axiom still holds true, and specifically what it means for our efforts to prevent illness and death from coronary heart disease.
Using Canadian population data from the 1990s, Douglas Manuel and colleagues tested three strategies for reducing coronary heart disease (p 659): lowering cholesterol across the entire population; treating only those with raised cholesterol levels; and giving statins to people at high baseline risk regardless of their cholesterol level, which is what the New Zealand guidelines, and others, recommend.
They found that overwhelmingly the most effective strategy was to treat people at high baseline risk. Does this mean that Rose was wrong? Apparently not. Rose said that a population strategy is needed where risk is widely diffused through the whole population. As Manuel and colleagues explain, this is not the case in Canada, where over half the adult population is at very low risk of heart disease and those at high baseline risk account for a large proportion of the total population risk.
Hugh Tunstall-Pedoe and colleagues tackle a related question and come up with an unexpected answer (p 629). They seek to explain the falls in blood pressure observed in 21 countries involved in the MONICA project during the late ‘80s and early ‘90s. They conclude that, because blood pressures fell at all levels of blood pressure, antihypertensive treatment in those with high blood pressure is unlikely to have made a big contribution. They can't say exactly what caused the fall in blood pressure, but in passing they conjure the evocative image of the ground (the population) moving under the treatment goal posts.
Taking up the story, Rod Jackson and colleagues (p 617) conclude that population and baseline risk strategies should be followed in tandem, but that the balance will depend on the risk profile of a particular population and the resources available. Low to middle income countries have not yet picked the “low hanging fruit” of population strategies to improve nutrition and cut smoking, so these are likely to reap large benefits in reducing the burden of disease. But even so, they say, this should be combined with targeting people at high risk.
Guidelines are only as relevant as the data they are based on. Current guidelines for lowering blood pressure after stroke are based on data from the PROGRESS trial, a randomised controlled trial in patients recruited from hospital. How applicable are they to treating patients in primary care? Not very, say Jonathan Mant and colleagues (p 635). Stroke patients in their primary care trust were generally older and had had their stroke less recently than the patients in the trial. Building on this, Sharon Mickan and Deborah Askew ask what sort of evidence we need in primary care (p 619). The answer is all too obvious. Relevant evidence and plenty of it.