klmk

• Relative Risk (RR)

  A measure of how likely an event is to happen in one group relative to another

  
• Odds Ratio (OR)

  The ratio of the two odds; happening vs not happening

  
• Number Needed to Treat (NNT)


  Measure of impact of medicine or a new treatment; how many to treat to impact one

  
• Relative Risk (RR) is the ratio of the probability of an event occurring with an exposure versus the probability of the event occurring without the exposure

  
• Relative Risk = 1 means the 'event' is equally likely in both groups

  
• Relative Risk >1 means the 'event' is more likely

  
• Relative Risk <1 means the event is less likely

  
• Calculating Relative Risk (RR)


  Risk (Test) / Risk (Control)

  
• Relative Risk is sometimes called Risk Ratio. Synonymous terms

  
• If a new treatment increases the likelihood of a cure, then RR for new treatment will be >1, which is good news - greater 'risk' of cure
• Odds


  Ratio of the number of cases where the relevant event occurred to the number where the event did not occur

  
• Calculating Odds Ratio (OR)

  Odds (Test) / Odds (Control)

  
• The null hypothesis for OR is 1. The current result indicates reduced odds of expulsion with the new design
• Absolute Risk Difference (ARD)

  Risk (Test) - Risk (Control)

  
• Calculating Number Needed to Treat (NNT)

  1 / Absolute Risk Difference

  
• Ideal NNT would be 1
• If the events are rare, the risks and the odds take very similar values

  
• For rare events

  The OR will approximate to RR
• For more common events

  The OR will diverge more strongly from RR from (null = 1)
• It's a common practice to report 95% CI for RR and OR
• It's less common to see authors quoting 95% CI for the NNT
• The Absolute Risk Difference (ARD) is calculated by subtracting one risk from the other
• The null hypothesis value for the ARD is zero. For RR and OR, null is 1
• The NNT is the reciprocal of the ARD (1/ARD)
• The main value of NNT is in health economics