I think we are at a point with diabetes where we have a statistical problem. There are so many people out there with some form of diabetes or another that almost any behaviour or food choice can be found to be linked to diabetes if you don't have a thorough grasp of statistical analysis.
Of course proper research does exist, where you compare an experimental group to a control and you see that certain behaviours have a better-than-chance likelihood of leading to certain outcomes, in those cases we can properly talk about some kind of link, potentially (even then, there are often so many confounding variables when it comes to health - was it because they ate processed burgers specifically, or was it because those burgers are often eaten with fries and sugary drinks in a real-life environment?). But it's unethical to conduct that kind of research to look for disease causes, because it would mean that if you are right, you are expecting that at least some of your participants will develop diabetes as a direct result of having volunteered for your experiment...not a nice thing to do to them.
Anyway my point is that depending on how the claims are made in the research and reported in the journalism, it can often amount to something akin to those jokes about how "100% of people who have ever consumed chicken in their lifetimes eventually died from cardiac arrest." (Because medically speaking we all die from a stopped heart, it's the reason why the heart stopped that's different!)