Experimental Studies

Experimental Studies

1. Random assignment to treatment

   • allows for validity of statistical tests
   • avoids bias in the assignment of treatment to individuals
   • creates comprable groups
   • eliminates selection bias which ensures that groups that are exposed and unexposed are representative.
   • Have stronger casual inference than observational

2. Problems w? Clinical trials

   • Ethical Issues
     • Safety of treatment
     • withholding treatment
     • types of patients included in study
     • ethics of placebo
     • ethics of blinding
   • Statistical power
   • lack of generalizability or external validity
   • Patient drop out
   • maintaining study integrity and data quality
   • Patient accrual

3. Intent-to-treat: a way to analyze w/out bias. Analyze as you randomize