As fellow statisticians, we were deeply disappointed with David Spiegelhalter’s article (Coronavirus deaths: how does Britain compare with other countries?, 30 April). Where is the careful statistical modelling underpinning the piece? Only with such analysis can comparisons be made so as to compare the effectiveness of strategies that different countries take. Yes, there are inconsistencies, underreporting and heterogeneity within countries, but the policies adopted by different countries show very large differences in effects that would seem to dwarf such worries. Is Spiegelhalter suggesting comparisons cannot be made now or at any time?
He suggests that the high death rate in New York might be explained by high population density, but then what about South Korea? If the high death rate in Italy could be explained by an ageing population, what is the effect of average age on death rate? We are worried the article will deflect criticism of political handling of the crisis, as it has already. In the US, a recent tool to model the effect of demographic variables and strategies such as social distancing has been developed for predicting how death rates vary across states. Such modelling could be applied to country comparisons. Model-based decision support methods are also critical for modelling testing and tracing distributed down to the community level.
Prof Philip J Brown University of Kent, Prof James Q Smith University of Warwick, Prof Henry P Wynn, London School of Economics
• The graph displayed at the Downing Street briefings purporting to show comparisons between the numbers of deaths in different countries is meaningless. David Spiegelhalter outlines many of the reasons why this is the case. But it is our government itself that has been keen to compare our performance with that of other countries. Presumably this is because, in its arrogance, it assumed that the graph would demonstrate that we would be more successful in “sending coronavirus packing” (as our prime minister put it) than many of our European neighbours. But it now appears that our performance is likely to be much worse.
The government has been forced to admit to the large numbers of deaths in care homes. With these now included in the overall total, it can be clearly seen from the graph that the number of deaths in the UK is increasing at a greater rate than in the other European countries shown.
Suddenly, in the same week that the care home numbers are included, we are told by ministers and advisers that it is impossible to make meaningful comparisons between countries. While I agree that this is the case, I cannot believe that the sudden change is a coincidence. This is yet another example of the government’s manipulation of the interpretation of data to mislead the public.
Bridget Shield
Professor emerita, London South Bank University
• David Spiegelhalter’s conclusion that it is too early to compare Covid-19 deaths between countries may be scientifically accurate. However, while it’s true that excess deaths may ultimately provide the best international comparison, the current statistics are good enough to draw some general conclusions: the UK is clearly behind the curve in bringing down deaths and has one of the highest Covid-19 death rates in the world, even per capita. It is unlikely that new data at the end of the year will change the picture very much.
Scott Wilson
St Andrews, Fife
• Join the debate – email guardian.letters@theguardian.com
• Read more Guardian letters – click here to visit gu.com/letters