Undark: The Value of One: What Can We Learn from Case Studies?

https://undark.org/2020/02/04/value-of-case-studies/

I'm not really convinced.

 

Works sometimes with autopsies, I think. They're not usually done on large scale, but one-offs can be compelling. Case studies are why I still eat abx from time to time, despite the havoc they render to my body.

 

Case in point: https://www.mdpi.com/2079-6382/8/4/183/htm

There are many of these in Lyme world. They speak to several different things that are not addressed in most guidelines.

I should think this utility of autopsy case studies can at times prove cross-elastic across several different medical communities.

 

Albeit some time ago, I worked for several years in neuropsychology research, parts of which rely almost totally on single case studies. My interest was acquired reading and writing disorders following brain damage. Single case studies are very useful in exploring ideas and developing theoretical approaches.

This is particularly so where you dealing with measuring change after the fact in a very diverse patient population. For example group studies of acquired spelling disorders are particularly problematic because you are unlikely to have quantified data on individuals premorbid abilities and strategies. You can not go back in time to assess individuals before their stroke or head injury.

However they can only ever be part of developing ideas, as given their very nature they can not be reliably applied to whole populations or clinical groups. Any functioning explanation must be able to cope with all specific individuals, but in contrast there is no logical justification for generalising from the specific to the universal.

Particularly in psychology and with behavioural interventions, most available experimental designs are potentially problematic. One way of dealing with this is to seek convergent evidence, to use very different approaches and experimental designs that hopefully come up with a coherent understanding. With my example of acquired spelling disorders, I was using single case studies with mainly a CVA patients to develop possible descriptions of how we generate written language, but then explored those ideas in relation to linguistic measures of large of corpuses of published written material and formal experiments looking at normal written language production both developmentally (ie with children of different ages) and in ‘normal’ adults.

Single case studies are a valuable option in developing a whole range of approaches to ultimately better understand an issue or clinical group, but they are only any use as part of a whole battery of diverse tools. To rely on single case studies alone is a disaster. I would argue that the disaster of the BPS approach to ME is because they have relied solely on just one very narrow experimental design, with significant limitations, which become very obvious when you look at other evidence sources.

 

How then comes that a specific and sadly not-generalized in an individual doesn´t disturb the universal in an individual?

 

Sorry I don’t think I understand what you are saying.

It is a fundamental principle of logic that you can not safely generalise from one example to every other situation. However, just one counter example can render an asserted generalisation invalid.

 

@Peter Trewhitt This is true, and therefore it is said that we never can be sure to have found the truth (in empirical matters, and the truth of this sentence then is a non-empirical or a logical one).

Nevertheless there are all these individuals, and we often quite aren´t able to "understand" what´s going on there. True, this doesn´t mean that we weren´t able to understand it in principle, instead we simply don´t look through the complexity (of, say, an individual patient or a new case). I guess this problem leads easily to attempts of generalisation that doesn´t succeed. So I don´t disagree with the assumption that case reports are of lowest or questionable evidence.

For joke I assumed that in an individual would be something that would not be a matter for generalisation, but nevertheless would be at work there. Then things that are able to become generalised or are already known, would have lost their value for application.

In fact, this even might be in a very very tiny small amount true in the universe, and the constants of nature might not be really constant but changing with time, or even with space coordinates. Then a theory of things in space and time would be true only for specific coordinates and every "case of" (using the generalised term) would be in charge only for one single thing.

Begging some pardon for thinking unnecessarily queerly - but this sort of thinking could be used as well as misused in medicine, when the relationship of patient-doctor is said to be important or even said to be a major player.