7up MAYBE causes diabetes

After a hard day at The Institute, I was presented last night, appropriately at dinner, with the fact that "an extra fizzy drink a day will increase your chance of diabetes by 20%" - as The Beloved heard it on the radio.  The ould skepticism alarm [DRRRRRANG!] went off immediately.  So it's taken me a little (company) time to track down the study which is is commendably available on line,(see the In The News section there about Consumption of sweet beverages and type 2 diabetes incidence in European adults: results from EPIC-InterAct by The InterAct consortium). As I suspected, it's a little less black and white when you read the research paper let alone with access to the actual data.  The study is noisy: it's a meta-analysis of 12 separate studies from 8 European countries; but it's also chunky: more than 15000 individuals have been included in the study.  As a tiny example of the issues of integrating such disparate data, the paper reports that all the study-centres measured the height and weight of their subjects except those in Oxford and France who just asked their people how much they weighed.

So the media-fact that has been abstracted from this huge study is a convenient soundbyte "The HR of diabetes per 12oz increase in sugar-sweetened soft-drink consumption was 1.22 (95% CI 1.09, 1.38)". Which I've put in bold so that reporters from Nauru (where the incidence of diabetes is a frightening +40%) don't need to read the small print.  HR is hazard ratio and CI is confidence interval: here 1.09-1.38. If the lower CI value is less than one ( <1.00) the association is not significant.  Now get this.  In the same paragraph as the soundbyte the paper reports that the "HR . . . for artificially sweetened soft drinks was 1.52 (95% CI 1.26, 1.83".  A much "better" result than a 22% increase, surely, except that when BMI (body mass index was calculated for everyone in the study along with a raft of other data as well as soda-consumption) was factored into the equation, the Hazard Ratio evaporated to zero.  And am I alone in my Eurochauvinism in objecting to "12oz" appearing in a scientific report compiled by a European consortium: jakers! even the Brits and Irish have embraced {ml, kg, km} these last 20 years and more.

Maybe soft-drinks contribute to diabetes, and maybe I'd like to see them banned (easy for us because we don't have 'em in the house), but I wish RTE would be a little more discursive and critical in their presentation of information that may impinge on a) dialysis rates in Galway b) peripheral blood circulation problems, gangrene and amputation in Donegal c) blindness in Wexford and d) jobs in Atlanta.  The way the report is received in an average math-anxious Irish kitchen where a family of 5 is washing down dinner with Fanta (and why shouldn't they?) is that one of them is done for (but that the rest will be all right).  
But if they are all within the normal range of waist, they're probably ALL going avoid diabetes.