Intended for healthcare professionals

Rapid response to:


Excess deaths: government commissions review

BMJ 2018; 361 doi: (Published 26 June 2018) Cite this as: BMJ 2018;361:k2795

Rapid Response:

As I was saying...

I am surprised at Nicola Oliver [1]. She tells us that 15,969 people died of flu (in England) last year, although only 320 died in hospital, and 15,649 were apparently left to die without due medical attention at home. What she fails to note is that the 15,969 deaths were not recorded deaths but a projection derived from the Flumomo algorithm [2] for 'flu attributable deaths' based on all cause mortality [3], so it does not really get us anywhere (except that it is just kind of thing I am complaining about!)

The Flumomo modelling may be more sophisticated (and possibly more opaque) than the model described by Donaldson et al in 2009 for projecting flu mortality from years in which excess winter mortality was above average [4], but it seems to be another method for overcoming the problem that actual recorded flu deaths are far to few to justify a vaccination programme.

Donaldson et al wrote:

"The official estimate of influenza mortality is produced by the
Health Protection Agency. It is derived from excess all-cause death
registrations in the winter. When the number of all-cause death
registrations rises above an ‘expected’ level in a given week, this excess
is counted. The estimates for the last five years in England & Wales
are: 1965 (2004-05 winter season), 0 (2005-06), 0 (2006-07), 426 (2007-
08), and 10351 (2008-09) . The highest estimate in recent years was for
the 1999-2000 ‘flu season, at 21,497.

"This method has its limitations. It captures the numbers of deaths
‘above expected’, but does not examine causation directly. An observed
excess may therefore reflect deaths due to cold weather or other
circulating viruses. When the number of deaths is small, there may be no
significant rise observable above the baseline of ‘expected deaths’. This
leads to a zero estimate, as seen in 2005-6 and 2006-7...

"A second estimate of ‘flu deaths is found in the annual mortality
statistics produced by the Office for National Statistics. These
statistics record the underlying cause of death. They are based on all
registered deaths, based on the information on death certificates. The
number of deaths for England & Wales with an underlying cause of
influenza (ICD-10 code J10-J11) for the four recent calendar years are: 39
(2008), 31 (2007), 17 (2006) and 44 (2005). .."

It should not be necessary to project flu mortality, at all, since cause of death is recorded. Not only does the flu not account for 50,000 excess deaths, it does not account for 16,000 either, but only perhaps 300.

[1] Nicola J Oliver, 'Re: Government exaggerates flu mortality again: why cannot the public be trusted with the truth?', 24 May 2019,

[2] 'Surveillance of influenza and other respiratory viruses in the UK: Winter 2017 to 2018' , Public Health England May 2018, p.46-7,

[3] Nielsen et al, 'Influenza-associated mortality determined from all-cause mortality, Denmark 2010/11-2016/17: The FluMOMO model.', Influenza Other Respir Viruses. 2018 Apr 16. doi: 10.1111/irv.12564. [Epub ahead of print],

[4] Donaldson et al, 'Author's Reply', 24 December 2009,

Competing interests: No competing interests

24 May 2019
John Stone
UK Editor
London N22