Maths for Clinicians 10

• What is a randomised controlled trial (RCT)?

  • Experimental (lab, clinical, field trials)
  • Analytical
  • 2 groups: treatment (intervention) & control group
  • All participants have disease
  • Randomly assigned to groups to reduce bias as differences between participants are divided across the groups
  • Single, double or triple blinding
• What is single blind?

  Patients don't know if they're in control or treatment group
• What is double blind?

  Operator doesn't know who's in control or treatment group
• What is triple blind?

  Statistician doesn't know who's in control or treatment group

  Removes bias as they can't use certain tests to obtain certain results

  (Principal investigator keeps code to eventually identify treatment vs placebo)
• What are the strengths of RCT?

  • Direct comparison of treatments
  • Can make causal conclusions
  • Randomisation reduces bias & confounding variables
  • Statistical test of significance easily interpretable
  • Can avoid Type 1 and 2 errors with sufficient statistical power
• What are the weaknesses of RCT?

  • If large sample needed -> expensive & logistically difficult
  • Informed consent may be impossible
  • Results may not be achieved IRL (outside study)
  • If outcomes occur after long delay, increases costs & loss to follow up
  • Ethical implications- RCT impossible without clinical equipoise
• What effects can be seen in RCT?

  • Placebo effect
  • Operator effect
  • Statistical test effect
• What are Type 1 errors?

  Concluding that there is an effect when there isn't one
• What are Type 2 errors?

  Concluding that there isn't an effect when there is