Unfortunately, you can't really rely on old data for things like that. Lots of people die from misdiagnosis today, or die with a undiagnosed condition. Prior to all the developments in medical science the situation would have been much worse and the data wholly unreliable. Plus, people tended to die much younger which further distorts comparisons (for example, if you were going to die at 60 from cancer, but died from a mill accident at 45, the cancer rate appears lower).
Similarly, analysing a patient population and finding that one group (for example welders) is present more often in a statistically significant proportion tells you absolutely nothing about whether or not the disease is caused by welding. This is because even 'random' data has patterns in it if you look hard enough. This is also why they define target populations before doing drug tests, because you can, for example, find that 'all left handed tuba players with advanced lung cancer benefitted from the drug'.
Perhaps other studies which could find that exposure to manganese in other contexts (various other industries, ground water, and so on). If they also show an association, then you * may * be on to something.
In order to prove causation, you need a model and experimental (or other) proof that the model is valid. For example, you say manganese causes Parkinsons because ..., and you feed a bunch of mice or whatever, manganese and you see what happens.
Of course, none of this has anything to do with the question at hand. Class action suits are not about what caused the problem (or most significantly, in this case assuming there is a cause/effect, should or could the welding industry have known this).
Now, everything I have ever read about MIG/TIG or stick welding (including books written 50 years ago) say: wear gloves, wear a full face mask, and work in a well ventilated environment. Not that that should count for anything ...