I simplified a bit. This was the finding of the 'thud' experiment in the US in the 70s. Diagnosis varied according to the amount of money the establishment charged for admittance, with the less stigmatised 'bipolar' far more common in the exclusive private clinics.
This is a review of studies in both the US and the UK from 1997 detailing differences in diagnosis and the effect of race and class. A section at the start of page 107 seems relevant to this discussion - a finding that alcohol abuse is less likely to be detected in better-off patients. The review does suggest that class bias is becoming less of a problem, but it also suggests that the clinicians' preconceptions about what a particular kind of patient looks like are still a significant factor in diagnosis. Tbh the most shocking differences are those relating to race, both here and in the US.