40% of Americans over age 20 (twenty!) are obese. Source: https://www.cdc.gov/nchs/fastats/obesity-overweight.htm
45% of American adults have hypertension. Source: https://www.cdc.gov/bloodpressure/facts.htm For the 55+ age group I'd bet money, without Googling, it's ~57%.
Those aren't co-morbidities, they just reflect the population. The diabetes prevalence seems maybe ~2x higher than would be expected, however, unless the 34% figure includes pre-diabetes and borderline cases. For age distributions see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3464829/
But for the age groups at issue the difference isn't half. For example, for the age group 65-74 31% are obese. See https://www1.nyc.gov/assets/doh/downloads/pdf/episrv/2019-ol..., p23. And if you adjust for race, poverty, etc, the gap probably closes even more.
I'm simply taking issue with the characterization of these supposed co-morbidities as representing "high risk" patients when in fact in the context of Americans generally (and to a significant degree at-risk NYC population), these conditions are typical, not atypical. I'm not sure what's to be gained, either in terms of our understanding of pathogenesis or in treatment, by distinguishing these conditions from a hypothetically healthier population. At the very least, we should be careful about imbuing the confluence of these conditions with too much meaning beyond the simple and obvious. Such conditions are co-morbidities for most causes death across the spectrum of illnesses, infections and otherwise.
The usage seems fairly obvious to me. The risk to young health people is somewhere in the range of 20-200x smaller than someone who is old and in poor health. It's not a surprising claim, but it's clearly a useful thing to know for medical planning. It's also useful at an individual level to get a sense of their personal risk level. Finally, it is useful in understanding past outcomes. If two different countries have different mortality rates, but also different demographics, and both of them had similar mortality rates normalized to those demographics... that's pretty useful to know.