HLSC 2P27 Lecture Notes - Lecture 5: Clinical Endpoint, Surrogate Endpoint, Biomarker
24 Jan 2017
School
Department
Course
Professor
Document Summary
Highest risk for confounding is in a case control study, we have no way of doing anything about the confounders that occurred in the past. In a prospective cohort study, confounding is possible but can be controlled. Bias: a systematic error that leads conclusions away from the truth, can occur at any point in the trial (beginning, middle, end) Internal: study had measure what it said it would (even if the hypothesis is wrong, the study can still be internally valid) External: taking those study results and extrapolating them to the general population as a whole. Incidence 10% in those who did not, 6% who did consume antioxidants. Trick to external validity is to specify your population (age, gender, region, etc. ) It is interventional, we are doing something (usually asking people to take something, or use a device that is being tested) People who agree to take part have a process of inform and consent.
