This is a good point but similar kinds of phenomena could lead to something more benign but equally problematic. For example, let's say someone inherits (genet

This is a good point but similar kinds of phenomena could lead to something more benign but equally problematic. For example, let's say someone inherits (genetically or environmentally) etiologic factors for psychosis and blindness. It's possible that whatever these factors are, they are severe enough to have widespread enough effects that the person no longer receives a psychosis diagnosis, but instead some other diagnostic label that's more comprehensive in nature. Schizophrenia is really just a label for a semi-heterogeneous set of pathologies that might be labeled something else. I'm not saying that these problems don't exist, just that how they are perceived and described might vary a bit from scenario to scenario (this is one of the main impetuses for DSM-III and later actually). So someone with one pathology might receive a diagnosis of schizophrenia, someone else, bipolar disorder, and yet someone else developmental disorder NEC/NOS. It's entirely possible that if there is a negative correlation, it's because one reflects some preventative factors for the other, or something like that, but it's also possible that whatever causes the conjoint presence of the etiologies is catastrophic, leading to death, like you suggest, or the need for some entirely different label.