Personal risk
How do you make the right decision about preventive treatment when n = 1 and you happen to be the ‘n ’ concerned? I recently had to do this for myself, and it has made me reflect on how I approach this with my patients, too.
My own dilemma related to taking warfarin. Many years ago I had a series of unexplained pulmonary micro-emboli. I was screened for all the known genetic factors, but no particular cause was found. However, my mother died suddenly in her early sixties from a massive pulmonary embolus, so a genetic influence seemed likely. My haematologist suggested that I should take long-term warfarin, and I agreed. A year or so later I had a serious gastro-intestinal bleed, so my target INR was lowered to from 3 to 2.5. Since then I have been careful to have this monitored regularly, to make sure that it stays around this level. I have been stable now for over a decade. I have read reviews of the latest research on pulmonary emboli, and they seem to confirm that I am probably in a group of people that should take anticoagulants for ever, but who knows? Metanalysis is fine if you are a population, but if you are making a decision for yourself as an individual, the certainties start to crumble.
I am getting older. As each year goes by I face an increasing chance of getting a stroke as an adverse consequence of taking warfarin. Statistically the chance is not very high, but I am particularly fearful of strokes. I could of course have a minor stroke and recover completely, but I suspect that people on warfarin don't get minor strokes, and don't recover so easily either. Possibly I could find some research giving the figures for this as well, but I doubt if they would give the precise odds for a 57-year-old white male professional, married with two children and a Norfolk terrier. You can see the problem. If the treatment was for symptoms, I could test it out by doing a trial of n = 1 myself, but with long-term preventive treatment this simply isn't possible. It is all down to guesswork, hunches and choice.
What would you do in my shoes? My own way out of this dilemma is that I have decided to stay on warfarin, but to reduce my target INR to 2. This is less than any of the guidelines recommend, but it balances the different anxieties nicely for me. I have booked an appointment with the haematologist to check this decision out, partly through courtesy and partly through superstition, although I suspect she will agree it is sensible.
This process has made me re-examine the way that I offer information to patients, particularly in relation to preventive treatment. Naturally, I try to tell people about the clinical evidence wherever possible, explaining the statistical risk of any particular problem, and letting them know how far any treatment is likely to reduce that risk. However, as my own experience has helped me to realize, such a conversation may barely scratch the surface in terms of promoting real choice. It may fail to engage with important personal variables, or to spell out fully the scale of the imponderables. More significantly, it doesn't really engage at all with people's subjective assessments of their own risks, or their fears and fantasies about these. In a nutshell, it substitutes an impersonal discussion about NNTs (or ‘numbers needed to treat’) for a personal one about n = 1.
I have now discovered that there is some useful research that can help us to build bridges from impersonal facts to personal decisions about illness prevention. Rather than just looking at statistical vulnerability, it examines the way that patients perceive their own vulnerability to particular conditions. One particularly helpful study shows how patients with a family history of hypercholesterolaemia negotiate their own personal sense of vulnerability to coronary heart disease.1 The researchers found that people's estimations of their cardiac risk were grounded in a fairly reasonable understanding of genetics. However, these estimations also took account of personal factors such as lifestyle, and drew on specific comparisons with other family members. One typical patient summed up his own individual calculation as follows: ‘My uncle is about 60 years old. He got his first heart attack when he was at my age, 40 years. The difference between him and me is that I don't smoke, my lifestyle is healthy, and I take exercise ... so I think I have an advantage there ...’
Each person's sense of vulnerability was also likely to shift over time, perhaps as a result of changes such as cardiac events in the family, or becoming a parent: ‘Two of my cousins were diagnosed with heart disease before they were 40' said one participant in the research. ‘Then I thought 'We have to do something ...’. Another participant explained how and why she decided to start on a statin: ‘It's something to do with having children, and all of a sudden growing up, having to take responsibility ... If I were to die, it would be a crisis for my daughter.’
The authors of this research argue that doctors should be sensitive to their patients’ individual preferences over time, in order to assess their readiness for preventive behaviour. I would go further. What the research suggests to me is that a personal understanding of risk can actually build on clinical evidence, rather than acting as some kind of folk substitute for it, as we might sometimes assume. Indeed, it points to the possibility that personal narratives may enrich the evidence for individual choices, by adding individual factors into the calculation in a way that no population-based research or metanalysis could ever do. Looked at in this light, patients’ own narratives about risk are not mere background information that we listen to simply in order to be polite, or to encourage compliance with our advice. They may be essential for accurate decisions to be made in relation to such things as starting treatment, or titrating dosages.
It would be fascinating to find out by further research whether listening to narratives of risk, and allowing these to form the basis of shared decision-making, actually improves the results for preventive medicine. I would not be surprised if encouraging people to share their intuition with us is the best way of achieving better outcomes. Paying more attention to n = 1 may even transform our NNTs. I hope this turns out to be the case, for both personal and professional reasons.
Reference
1. Frich JC, Ose L, Malterud K, Fugelli P. Perceived vulnerability to heart disease in patients with familial hypercholesterolaemia: a qualitative interview study. Ann Fam Med (2006) 4:198–204.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||