Who's most at risk from swine flu?

A headline in the news section of this week’s BMJ reads: "people with asthma most at risk from H1N1 complications". But are they?

To support the BMJ’s headline, the evidence presented would have to compare the percentage of (H1N1-exposed) asthma-sufferers who develop complications of confirmed-H1N1 sufficient to require hospitalization with the corresponding percentage of, say, diabetes-sufferers in the same situation. 
 
However, the FLU-CIN data reproduced in the news report are quite-other, namely: the percentage of some 192 hospitalized confirmed-H1N1 cases in England for whom a particular complication, such as asthma, diabetes or pregnancy, is present.
 
Wrong denominator is the problem, as can be simply illustrated: Suppose that an H1N1-exposed population of 100,000 consists of 1,000 folk with disease A, 3,000 with disease D, and 96,000 with neither (group N).
 
Further suppose the confirmed H1N1 hospitalization rate is actually 1.5 in 1,000 for group N (so that 144 are hospitalized), 10 in 1,000 for those with disease D (so that 30 are hospitalized) and it is highest at 18 per 1,000 for those with disease A (of whom 18 are hospitalized).
 
Of my 192 hypothetical patients (144+30+18) hospitalized with confirmed H1N1, only 9 per cent (18/192) have disease A, which is uncommon but was associated with the highest H1N1-hospitalization-risk, but 16 per cent have disease D by dint of its being more common in the population despite having lower H1N1-hosptalization rate than disease A.
 
BMJ  pray tell: what are the relative prevalences of asthma, diabetes, pregnancy - three (only) of the highlighted co-morbidities in the 192 hospitalized H1N1 cases on which their headline teetered? The text of the article accurately reported that asthma is the commonest comorbidity found in those admitted to hospital: but this not the same as saying that people with asthma are most at risk of complications.
 
Muddled denominators matter, not just as an irritation, but because wrong sums may also seriously mis-direct public health attention to wrong targets.