Intended for healthcare professionals

Endgames Statistical Question

Interpreting hazard ratios

BMJ 2015; 351 doi: (Published 28 August 2015) Cite this as: BMJ 2015;351:h4631

This article has a correction. Please see:

  1. Philip Sedgwick, reader in medical statistics and medical education,
  2. Katherine Joekes, senior lecturer in clinical communication
  1. 1Institute for Medical and Biomedical Education, St George’s, University of London, London, UK
  1. Correspondence to: P Sedgwick p.sedgwick{at}

The impact of isoniazid prophylaxis on mortality and tuberculosis in children with HIV was investigated using a double blind placebo controlled trial. The intervention was isoniazid given with co-trimoxazole either daily or three times a week. Control treatment was placebo isoniazid given with co-trimoxazole. The setting was two tertiary healthcare centres in South Africa. Participants were children with HIV aged 8 weeks and older. In total, 277 children were recruited and randomised to the intervention (n=139) or control treatment (n=138).1

The primary outcomes included the length of time after randomisation until death from any cause and the length of time after randomisation until the occurrence of tuberculosis. The initial results of the trial were reported after participants had been followed for a maximum of 500 days (median 5.7 months) During follow-up, mortality was significantly lower in the isoniazid group than in the placebo group (8% (n=11) v 16% (n=21); hazard ratio 0.46, 95% confidence interval 0.22 to 0.95). Furthermore, the risk of tuberculosis was also significantly reduced in the isoniazid group (4% (n=5) v 10% (n=13); 0.28, 0.10 to 0.78). The researchers concluded that for children with HIV, isoniazid prophylaxis has an early survival benefit and reduces the risk of tuberculosis.

Which of the following statements, if any, are true?

  • a) When calculating the hazard ratio of death, it was assumed that the death rate was constant during follow-up for each treatment group

  • b) The intervention group had …

View Full Text

Log in

Log in through your institution


* For online subscription