Is it possible that the fact of receiving a diagnosis is more strongly correlated with the likelihood of getting help than you accounted for? Also, what is "diagnosable" – that sounds suspiciously different than "diagnosed." That doesn't even sound like a statistic that tracks against individuals, so I don't know how you could make an inference like that.
Even if we saw the statistical effect you predicted (eg. diagnosed mental health issues correlated with a lower rate of suicide), you should consider that the stigma of getting a diagnosis of mental illness, is likely to cause some people who may even know they have a problem, to intentionally avoid getting diagnosed.
Also, these: "I can't seek help because if I do have a mental health issue then my insurance is going to skyrocket / I don't have any health insurance / my insurance doesn't cover mental health."