More seriously, that's a big and interesting problem that I'm thinking about from the evidence end. Personal experiences, anedotes, and whatnot are very low-ranked sources of evidence in e.g. GRADE and other evidence-based medicine hierarchies. Yet medicine as a clinical encounter starts with taking patient histories, and (at least as British Drs are taught - can't be sure what happens elsewhere) these personal experiences are absolutely the foundations of diagnosis and so on. So I think there's a good piece of investigation that needs doing urgently - otherwise why the doublethink?
My impression is each individual, regardless of the philosophy and science seem to have their own thresholds for the generalisation of knowledge (and implied directionalities of inferences; individual->group, group->individual). These thresholds most likely vary over time, based on prior beliefs and how much the new knowledge challenges those beliefs.
All this talk about what is or isn't a "scientific field" still suggests to me that the philosophy of science lags, rather than predicts the practise.
Scientists, given sufficient interest in the field are creative and will devise new approaches to demonstrate causation.
Attempting to reduce models of causation down to ideas like Russo-Williamson Thesis or variants will be problematic, namely the more specific you make the model, the more you reduce the sensitivity of the model and vice versa. My own view on the examples you have mentioned are as follows:
Discovery of McArdle’s syndrome. The initial discovery was insufficiently justified (to be generalised) based on statistics, it isn't until further observations are made before one should have high confidence in the findings (regardless of whether someone explicitly calculates the statistics or not). It isn't until the deeper genetic findings were confirmed that McArdle’s syndrome wasn't merely a placeholder.
The HSV link with cervical cancer likewise was a preliminary working hypothesis in the absence of stronger evidence. Albeit seemingly large amount of evidence to the naive observer. The problem is putting too much confidence in preliminary hypotheses (hence various debates about lack of replication in biomedical and psychologial sciences by Ioannidis and others). There are many potential examples across medicine which existing hypothetical models that are popular among physicians may turn out to be false, for the same reason that the HSV hypothesis turned out to be false. Scientists know that these hypothetical models are not strong theories, otherwise there would no longer be a reason to employ them. The biggest mistake is when journalists, doctors, political and corporate leaders treat these preliminary findings as confirmed facts, especially when there is substantial debate among scientists (as in ME & CFS). In practise it is
rational (according to some models) to form beliefs without applying any specific thresholds of demarcation for a "justified true belief", so long as one is open to change one's view when better evidence comes along. The lack of such strict demarcations means there is always going to be debate, so the practical question is what does this mean for the behaviour of people in such debates where the outcomes are specific policies...
ME & CFS are unique in that very little money (comparatively) has been spent doing novel in-depth research, hence the continued debate and lack of understanding how replicated findings like elevated TGF-Beta and certain autoantibodes could play a role in the condition.
Another example is that of scientific creativity is macroevolution. Scientists obviously cannot go back in time to observe, so they simply developed new methods that can be tested to confirm macroevolution, namely the propogation of SNPs in conserved genes across a phylogenetic tree. Notably, how this correlates with the propogation of SNPs of viral fossils in our genomes. This is a finding which can only be made using a theory of evolution and one that clearly contradicts intelligent design. Likewise, many proteins have multiple functions that seem strange and contradictory unless looked through the lens of evolution and how predictions these proteins are necessary for the function in organisms of varying complexity and function (and their position in the phylogenetic tree).
This is an example of creative scientists, after being told "it cannot be done", devising methods of solving the problem.