What happens when you can’t conduct an RCT?

This preview shows page 1 - 11 out of 73 pages.

P6031—Research Methods and Applications | Quantitative Foundations 1 ReMA | Quantitative Foundations Lecture 19 | November 1 Observational studies: The Cohort Study
P6031—Research Methods and Applications | Quantitative Foundations 2 Last Times Linear correlation between two continuous variables The correlation coefficient Sample correlation: r Population parameter = ρ R 2 The proportion of the total variability in Y that is explained by it’s correlation with X
P6031—Research Methods and Applications | Quantitative Foundations 3 Last Times Linear regression ! " # = % & + % ( ) # b o = the predicted value of Y for individuals with X = 0 b 1 = the expected absolute increase (or decrease) in the expected value of Y for every one-unit increase in X If X is dichotomous, b 1 is the difference in the predicted mean value of Y between groups of X
P6031—Research Methods and Applications | Quantitative Foundations 4 Today Causal inference when you can’t do an experimental study Observational studies Introduce the Cohort Study Measures of association and a range of plausible estimates from a cohort study RR, RD, IRR, OR and 95% CIs
P6031—Research Methods and Applications | Quantitative Foundations 5 Key features of RCTs Random assignment of exposure Intent to Treat Analysis Well-defined, specific exposure Blinding Longitudinal follow-up of participants All of these are designed to limit alternate explanations for an exposure-outcome association other than a causal link.
P6031—Research Methods and Applications | Quantitative Foundations 6 Problem: RCTs aren’t always ethical or feasible Ethics You can’t randomize people to receive something you know is bad for them You can’t randomize people to NOT receive something you know is good for them Feasibility Some exposures can’t really be manipulated Genetic mutations Being born with two X chromosomes Being white
What happens when you can’t conduct an RCT?
P6031—Research Methods and Applications | Quantitative Foundations 7
P6031—Research Methods and Applications | Quantitative Foundations 8 Moving from Experimental to Observational studies Instead of assigning individuals to receive exposure, we can instead simply compare between individuals who are exposed vs unexposed to a putative cause of an outcome of interest.
P6031—Research Methods and Applications | Quantitative Foundations 9 The cohort study An observational study in which a group (or cohort) of individuals free of the outcome of interest are followed longitudinally over time to assess the proportion (or rate) who develop the outcome of interest over some period of follow-up time. Cohort studies compare the difference in the incidence of the outcome of interest between exposure groups.
P6031—Research Methods and Applications | Quantitative Foundations 10 EXPOSURE or INTERVENTION UNEXPOSED or NO INTERVENTION SAMPLE FROM POPULATION

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture