07/22/2009
Swine Flu: knowns, unknowns, calculation and speculation
The importance of filtering the reasoned calculation from the increasingly wild speculation - and the importance for the media to report on the basis of robust evidence and credible opinon - will become ever more central as the pandemic evolves. This week I wrote of the need for careful control of advice issued to the public to ensure that, even when working with the many unknowns of this virus and its spread - it is as clear and consistence as possible. This was the communication banana skin encountered as three different maternity organisations quickly offered guidance in the wake of the death of a mother who was infected with swine and had given birth prematurely. Such occasions will arise again - the public clamour for answers will only intensify - but public bodies must tread carefully as to how and when important information is channelled.
Likewise official figures need to carry all their necessary caveats, be they presented in the House of Commons (Andy Burnham's "100,000 new cases a day by August" claim remains methodologically questionable) or in the media. This was highlighted in the way Sir Liam Donaldson's carefully presented planning advice for the NHS (July 16) was reported: that the NHS should have services capable of coping with a worst case scenario of 65,000 deaths in the first pandemic wave. This was absolutely not a mortality prediction (and could not justify reports such as the "Swine Flu will kill 350 people every day" of the Daily Express). Similarly Sir Liam's announcement that the death toll had risen from 17 to 29 did not represent 12 deaths since the Department's last announcement (the "Swine Flu kills 12 in 4 days" alert of The Sun and Evening Standard)...
It was actually nine new deaths in England, and one in Scotland (the 17 was only an England tally, the 29 a UK figure). These were new because they had not been disclosed before - but actually had most likely occurred during the course of the preceding fortnight or so (and confirmed as linked to swine flu more recently).
The latter point is particularly important when gauging the spread and impact of the virus, and one that the Government would do well to invest in and investigate further. The difference between when a death occurs and when it is reported (and likewise when someone is hospitalised, as opposed to when a hospital is phoned by data collectors and reports that a person is in its wards suffering from swine flu) is vital to tracking the evolution of H1N1 as accurately as possible. This is information that should be available, and is causing understandable frustration among statisticians and epidemiologists.
Another issue worth reiterating is the difference between a death from swine flu and the death of a person who had contracted the virus. The former sounds like a causative link, but to date has been inaccurate in most cases (a man from Basildon is the only person confirmed to have died after contracting swine flu with no other underlying health problems). The latter has become journalistic shorthand for a death accompanied by swine flu, but not definitely caused by it. This difference is vital, and not made explicit often enough.
The newspaper reporting of July 14 (the Daily Telegraph's frontpage headline "Swine Flu kills healthy girl, 6" and Daily Mail's opening paragraph on page 1 "Parts of Britain were in the grip of a swine flu epidemic last night as the first healthy child died of the virus") gave the very strong impression that swine flu was the only cause before anything was known (interestingly their online reports were slightly more tempered).
As official agencies now conclude, 6-year old Chloe Buckley's death was the result of septic shock following a bout of tonsilitis. Parents at the school gate - and the borough's head of education - may have said that she was "perfectly healthy" when speaking hours after her death was announced (and the comments, particularly from the official, were rightly reported as such), but they must carry the necessary caveats that cater for the many unknowns then still yet to be clarified. Meanwhile speculation about care failings by her GP was without any foundation, as her parents' statement shows. Now with the post-mortem and pathologists’ conclusions before us, questions can be raised about the likelihood of swine being a contributory factor in Chloe’s death and the wider context of its interplay with other infections (which will be much like other flus)...
The role of communicators in conveying a complex mix of fact and supposition is a tricky one. I would not dare suggest that The Times will not fall foul of the pitfalls - to replace an 'unknown' with a 'possibile' that appears probable (or worse definite) that then turns out to be wrong. The headlines above are simply chosen as some of the highest-profile front-page examples to date. (Indeed our headline writers wrestled long and hard with "UK prepares for 65,000 deaths from swine flu" of July 17, which was felt to accurately convey the fact that Sir Liam's statement was a planning assumption, and carried the topline explainer that this was the NHS preparing for a worst case scenario, as well as an analysis explaining the methodology). The point is this: as the pandemic develops the onus is on officialdom to flag fact, prediction and possibility accordingly, and the media, charities, professional bodies, scientists and academics to keep these distinctions as sharp as possible at all times.
Posts
