RCTs and human beings

Something went wrong when Professor David Sackett's original thoughts on "evidence based medicine" turned into exercises in statistical analyses of the things which can be measured at the expense of not thinking about the things which can't be measured.
What happened was an over-reliance on the "gold standard" of the double blind randomised controlled trial. If you don't know what that means, it is a research methodology which involves randomly allocating patients to different groups, giving the different groups either different treatments or giving one group a treatment and the other a placebo (we can talk about that later!), then measuring the differences in the outcomes between the groups. The "double blind" bit means neither the patients nor the doctors know which group is getting which treatment.
There have been many criticisms of this method and many people support it as a rational way to discover the true effects of a treatment. My BIG problem with this method is the basic assumption.
What's the basic assumption?
People don't matter.
What do I mean by that? Well, the idea of "controlling" is to try to set up a trial where you rule out the human factors by balancing them between the groups. That means that, once the study gets under way, then it doesn't matter who the person is who is receiving the treatment or who the person is who is giving the treatment.
Does that strike you as a good idea?
Firstly, is it likely that it doesn't matter who the people are? No. The answer usually given is that the researchers recognise that human factors influence the outcomes but that they "control" for it by "matching" the groups. But the matching is only done on a limited, a very limited, number of parameters. Human beings are incredibly complex and can't reasonably be reduced to a small number of measurable parameters.
However, what bothers me most is that when the health service tries to use this approach to produce real services for real patients then we find that we get services designed on the basis of statistical measures, not services based on the understanding that people matter.
Does it matter who the doctor is who you consult? Yes it does. We are all different and the doctor-patient relationship is just that - it's a relationship. The people who form that relationship are the most important factors in understanding that relationship.
If we were to base our services on the recognition that people matter most, then we would make greater efforts to understand the needs and wishes of the individuals. We would develop patient-centred services to use a fashionable phrase, but we would also give prominence to the what makes it easier for those who provide the treatments to work more effectively - that's about terms and conditions of service, it's about support and encouragement for continuing education and professional development and its about morale.
If I'm going to have my hope of being treated by someome who "frankly, gives a damn" realised then that person needs to be recognised and supported as a person too.
We need services based on the needs of patients being met by therapists whose professional and human needs are also being met, not services based on the assumption that people don't matter.