What they mean by evidence based medicine.

Apologies for these musings. Others probably pointed out the problems long ago.

A problem seems to be that the term "evidence based medicine" is made to suffice for both "the whole" and "the part".

As described by Sackett in the paper referred to above "Evidence based medicine" is a system in which medical practice with varying degrees of supporting evidence is ranked according to the general quality of evidence of a particular type.

Usually when "evidence based medicine " is used in general parlance the term is used to describe evidence belonging to one of the particular types.
Thus, for example, practice supported by meta- analysis is one type of evidence based medicine; that supported by anecdotal expert opinion is a different type.

It might be helpful if Sackett's use were referred to as Evidence Based Medicine, and the other type had an adjective attached to describe it as "best", "good", "moderate" or "poor". So one might say "moderate-evidence based medicine".

There is then a problem in that that description would only refer to general evidence of that class. It would not refer to the quality of the particular evidence supporting the practice. Thus one might have a poor example from a moderate class, but to suggest that people use the term "poor moderate-evidence based medicine" might be a step too far.

 

Yes, it is a question of generalisation. If it works for one person, will it work for others and if it works for others, will it work for a specific patient?

I agree with your point about the importance of assigning quality - unblinded clinical trials relying on subjective outcome measures, case study reports, cross-sectional epidemiological studies and so on are all "suggestive" quality evidence. (where the suggestion is further research...)

A real problem is where there are insufficient sample size for statistical comparison, in cases of rare diseases for example, where detailed mechanisms are generally required as high quality evidence.

 

The problem is all those terms are already in use, it would might just add to the confusion. The issue here seems to be that the term EBM is used as a shorthand, and the details that explain its relevance, or lack thereof, are generally skipped. "Evidence-based" can be everything from the best scientific evidence available to horse manure. Its the details that matter.

When someone says evidence-based to me in real life I am currently always wanting to ask "so what is the evidence in this case, what data is convincing?".

So very much in medicine is dubious, and the two worst fields are probably first, psychiatry, and second, nutrition. Like right now I am reminded of a line in a Woody Allen movie where, if I recall correctly, someone from the future points out we are all told to eat lard. The funny thing is ... that is coming true. Most nutritional fats are now thought protective, its the chemically modified fats that are a big problem. The other example from nutrition I am becoming aware of is there is growing evidence that almost the entirety of the nutrition advice we give to type 2 diabetics is seriously flawed, if not dangerous. A small clinical trial here of a low carb diet, with our CSIRO researchers, shows much better results than the low fat high carb diets we tell diabetics to eat. Indeed, its been hypothesised that insulin is the primary driving factor, and it might be more important to focus on keeping that low rather than blood glucose, and that the low glycemic sugar fructose might be more damaging than glucose overall. Psychiatry is of course only as reliable as its objective diagnostic tests ... . Now in all the above examples, there is still little first rank evidence that conclusively shows what is what.

Its even worse when people just accept RCTs or meta-analyses as though they have to be good. They don't. Details still matter. This is a quagmire of terminology, with many aspects of EBM are not well defined so far as I can see ... but some is well defined, though how many doctors know the details?

 

This is what I think about when thinking about N=1 data, or data from a specific patient or data point. The main things about EBM is its about reliability as generalisations, which is ranked in various ways. I am very concerned however that, as you point out @Snow Leopard , its not always reliable when you go from the general to specific instances. General rules are not typically universally correct rules.

 

I also worry about its cavalier misuse, and not-so-subtle deterioration into a marketing ploy, or worse, propaganda. Today, when I read or hear EBM, I sometimes wonder more about medical politics than proof.

 

I've tried it. It's quite a fun experiment. Sat in front of a specialist who was spouting pure tripe straight from The BPS Guide to the Galaxy with the phrase Evidence Based Medicine sprinkled on top. Let him go for a bit and when he paused asked -

"Whose evidence? Can you give me a list of trials? Just how high was the risk of bias? Did you look at the data yourself?"

Then I just sat and looked at him, waiting for a response. He wasn't happy. Patients aren't supposed to ask those kind of questions.

 

Yes, its now a political rhetorical phrase.

Yet, if you don't, you have no idea if its good advice or tripe. Yes, I found doctors hate being questioned, and hate even more being proven wrong. Very few doctors are comfortable when dealing with their limits to knowledge. Yet we all have limits to knowledge, how you deal with it is what is important, and denying it is a poor choice.

 

Just found this on Wikipedia: https://en.wikipedia.org/wiki/Evidence-based_medicine

I fear the harm done by the BPS crowd will last a generation, even if an objective test for ME was found tomorrow.

 

Sometimes they just never end. Not in the lifetime of those affected at least.