Thursday, 25 August 2011

The camouflage of numbers

Often in my blog I ask a question about the way we do things in medicine, and reflect on how honest we are being with both ourselves and others. My blog is sometimes my way of asking myself how I really do things. But you knew that already, didn’t you? This week, I have been thinking about how we as doctors use data, both with patients and among ourselves.

For a little while, I have known that we are often guilty of the pitfalls of ecological fallacy. This is when we take data that we know applies across whole populations, and inappropriately apply it to individual circumstances; we are guilty of this when every time we resort to stereotypes (and we know how much all doctors like to stereotype.....), but we are also guilty of it in the way that we use data and information, and in the way that we relay that information to patients and colleagues.

When a patient asks me what the chances of recovery/complication/dying is, I can refer to the data available and say that there is a 75% of full recovery, a 10% chance they will develop a complication and a 5 year survival of 76%. I rarely do this, for a number of reasons, but there are two main ones. Firstly, even when such information is available to me, and it often is not, I do not really know how it applies to the particular patient in front of me. The epidemiological data we have is based on populations of people, and the figures we derive from such data are applicable to the average person within that population, and who really knows what average looks like? I'm pretty sure that whatever average is, it is rarely the elderly patient I am looking after – the elderly tend to be excluded from the major trials, so application of the evidence to the elderly usually requires extrapolation outside of the data set.

But although what I have just said about evidence in the elderly is entirely true, I have taken an intellectual shortcut, in the full knowledge that the sort of argument I have just given tends to be more acceptable than the argument that I am just about to give.

The use of clinical data in conversation is an obfuscation, driven not by a desire to inform the conversation, but rather driven by a desire to assert one's intellectual authority or to avoid the real question.

The epidemiology of disease rarely informs one's clinical practice, or what one actually does for an individual patient, but it can give the appearance of competence, and give weight to credibility. This often plays out between competing professionals, and in the interesting interpersonal dynamics in the work place.

Furthermore, blinding patients with numbers can be a way of putting up a barrier between you and them. One of the definitions of a profession is that it a trade with a syndicalist mode of representation, with a knowledge monopoly in a particular field, and using the kind of data that we have access to can be a way of emphasising that fact: I have the information, I am a professional, and therefore, my views and opinions are important.

I am not suggesting that that attitude is commonly expressed, but I think that it is subconsciously implied.

When a patient asks about their chances, are they really asking for the numbers, or are they asking you to help them know what to expect? The judicious use of relevant information can be a great way of making a point, but we should not forget what is often the subtext: being a patient can be daunting and bewildering, and it can be very difficult for patients to put this feeling into words. They therefore resort to socially validated ways of seeking the kind of information on to which they might be able to hang their hopes. 'Doctor, what are the risks of …...?' is often the way that patients say, 'Actually, doctor, I'm a bit bewildered, and confused, and actually, what I would really like is for someone to talk me again through what is happening.'

No comments:

Post a Comment