In-store pharmacies
  Prescribing
  NSF for older people
  Community pharmacy
  Homoeopathy
  Meningitis and the HAJ
  Education
  The Army
  Folic acid
  Nurse prescribing
  Labelling
  The Journal
  The Society

The Journal

Precision of relative risk

From Ms D. Layton, MRPharmS

I have just read three items reporting results from presentations given at the annual meeting of the American College of Cardiology (PJ, March 24, p382). Although these items inform the readership of encouraging results for innovative therapy, I consider that presenting just the relative risk estimate without a confidence interval is misleading.

All estimates of effect and impact should be accompanied by a 95 per cent CI so that readers can gauge the precision of the relative risk estimate. For example, an overall relative risk of 0.81 will be non-significant if the 95 per cent CI crosses unity.

Deborah Layton
Drug Safety Research Unit, Southampton

Trial results reported in The Journal are significant unless otherwise stated. However, we accept that relative risk estimates should include a confidence interval.—EDITOR.

Previous Topic (Labelling)
Next Topic (The Society)
Send your letter to The Editor