Moorhens and mallards |
||
|
Bandolier is often written overlooking a pond which is home to a pair of Moorhens (at least we think they are Moorhens; we are told that Moorhens have yellow beaks and Coots white). Occasionally the pond is visited by some Mallards. Moorhens are aggressive little chaps, and the Mallards don't stand a chance - they get seen off even though the pond could easily support a pair or two. There are a lot of papers written about evidence-based medicine, NNTs, health economics, and ways to treat patients, or run the NHS. Many of them seem to take a Moorhen approach to the problem - that of trying to see off all opposition. This is a pity, because life isn't simple, the NHS isn't simple, and decision-making with individual patients isn't simple. Tools and toolboxesThe various methods we use, statistical, representational, or economic, all have their strengths and weaknesses. They are tools we have at our disposal to use when and where appropriate. Just as we would not try to take a wheel off our car with a screwdriver, there will be situations when statistical or economic tools are inappropriate.Ninety-eight percent of health economic evaluations are bunk. Now that is a strong statement that will make a few readers sit up. One might try and justify such a statement by saying that most economic evaluations are based on single trials, and that single trials, if small, are unrepresentative and may not give the "true" size of an effect. Economists might respond by arguing that careful accumulation of cost data best informs economic analysis. Baying for the moonBut the point, surely, is that what we have is what we have. We must strive for the best while accepting that we may not have it right now, but get on and do the best we can anyhow.Thoughtful wordsAlan Maynard's 1997 essay [1] sought to argue that using best evidence for the treatment of individual patients may not be the best value for society from a fixed-budget health service. It seeks to (and does) pour some cold water of reality on what some people consider an over-idealised concept.Jack Dowie makes a plea for decision analysis based medical decision making [2]. Bandolier finds his thesis not unattractive, and decision analysis can shed an interesting light on problems because, at least to some extent, it takes a broader look at costs and consequences. But Dowie gives NNTs a rough ride - "an intrusion of population-based reasoning into clinical decision-making" [3]. Some interesting arguments, best read in conjunction with workers who reach a very nearly opposite conclusion about the use of numbers needed to treat and harm in cost effectiveness studies [4]. Real worldBandolier has tried to avoid being a Moorhen here. These papers have good and bad points - the latter reflecting mainly the propensity to see problems everywhere. Sure there are problems, there always will be. The point is to get on and solve them, rather than moaning that they exist. And the real world changes. When the NNT is the outcome measure that GPs feel most able to understand and explain to others [5], we probably should work on reinforcing success. We need to use all the tools in our toolbox to fill the knowledge gap (Figure), and not be afraid of allowing a few Mallards to share our pond. |
|
References:
|
previous story in this issue