Intended for healthcare professionals

Letters General health checks

Authors’ reply to Lauritzen and colleagues, Newton and colleagues, and Mangin

BMJ 2014; 349 doi: (Published 30 July 2014) Cite this as: BMJ 2014;349:g4790
  1. Peter C Gøtzsche, professor1,
  2. Karsten Juhl Jørgensen, doctor1,
  3. Lasse T Krogsbøll, doctor1
  1. 1Nordic Cochrane Centre, Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen, Denmark
  1. pcg{at}

As our editorial explained, criticism of a clear negative result cannot turn it positive.1 Lauritzen and colleagues say that the trials in our Cochrane review tested outdated screening tests and treatments.2 3 This is not correct—for example, probably all 14 trials measured blood pressure and six trials tested for diabetes. They also say the trials were old, but if we increase the number of deaths from 11 940 to 15 103 by including the results of the Inter99 trial,4 the risk ratio for total mortality remains at 0.99, with virtually the same confidence interval (0.95 to 1.02). How can they challenge such results by referring to a meta-analysis of surrogate outcomes and to retrospective non-randomised comparisons? They also argue that the Inter99 trial didn’t include pharmacological intervention, but when more diagnoses are made, more drugs are being used.

Newton and colleagues, who are employed by Public Health England, which is responsible for delivering the NHS Health Check programme, say that the Inter99 trial is not directly comparable with the programme.5 Of course not, but they ignore that to recommend screening we need randomised trial evidence that it works. The programme operates in direct contradiction to the NHS’s own screening criteria ( They say that their patients are older and worse off and that it may take more than 10 years to see an effect. However, all trials included high risk patients, and if there were an effect we would have seen it. Like Lauritzen and colleagues, they talk about modelling studies, which are highly dependent on assumptions, but seem to be the standard “rescue” when results from randomised trials go against popular beliefs.

What we need are not general health checks but structural changes, which should involve a much tighter grip on the food and drug industries. As Mangin points out, adverse drug events are now a leading cause of death in developed countries.6 It is therefore unconvincing when Newton and colleagues say that “substantial levels of undiagnosed treatable illness are being detected.” Treating more people through screening has the potential to increase mortality.1


Cite this as: BMJ 2014;349:g4790


  • Competing interests: None declared.


View Abstract

Log in

Log in through your institution


* For online subscription