Card Set Information

2013-10-16 13:37:51

Show Answers:

  1. 1. What is the utility of RCTs? 3

    2. What are the advantages of RCTs?

    3. Limitations?  4

    4. What are important concepts in RCT design? (4)
    1. used to evaluate new drugs/treatments, evaluate new medical technology, and assess new screening programs.

    2. No selection bias, certainty about exposure (compliance should be monitored), and gold standard of study designs

    3. Expensive, lengthy, lack of generalizability, and ethical issues.

    4. (1) Concurrent controls (2) Placebos (3) randomization (4) double-blinded technique
  2. 1. Why is randomization so important? (3)

    2. What are some issues with randomization? (2)

    3. What do you do if age and sex strongly affect prognosis?

    4. What is single-blind design?

    5. Placebo - what can it help control? Do people often guess? When do you have to tell people about the existence of placebo? What else is placebo group important for?
    1. To prevent patient/PI bias, to establish balance at BL in terms of observable and unobservable confounders, and to create basis for statistical testing that may rely on randomization

    2. If subjects dropout based on certain characteristic to a greater extent in the intervention or control groups, benefits of randomization are lost. ALSO, if one/both arms are not compliant with group assignment - could skew results

    3. Stratified randomization - first by sex, then by age.

    4. Subject unaware of group assignment.

    5. Crossover, yes, pharmaceutical trials. Studying side effects.
  3. 1. Why do we have to use ITT?

    2. What do we look at in terms of outcome?

    3. How are results/analyses similar to cohort studies? 3

    4. Draw power table: decision vs. reality

    5. What is a?

    6. What is B?

    7. What is power?
    1. We know true intervention effect is diluted by crossovers, but if data were not analyzed this way, it would violate randomization process introducing confounding

    2. Cumulative incidence - avg probability that an individual in the RCT will develop disease over follow-up period (like COHORT!)

    3. Uses cumulative incidence and survival analysis-product limit analysis. (1-probability of not developing disease in any of the time intervals) Risk ratio

    5. a = probability of making type I error - probability of concluding treatments are different when they're not

    6. Probability of making a type II error - probability of concluding treatments are the same, when they're different.

    7. Power = 1-B --> probability of correctly concluding that the treatments differ
  4. What must be specified in order to estimate the sample size needed in a randomized trial? (5)
    • 1. Difference in response rates detected
    • 2. Estimate of the response rate in one of the groups
    • 3. Level of statistical significance (a= 0.05)
    • 4. The value of the power desired (1-B)
    • 5. Whether test should be one or two sided.
  5. 1. What happens in a planned crossover?

    2. When do you use a factorial design?
    1. Patient serves as his/her own control, time for washout is important.

    • 2. When testing two drugs/nutrients and the outcomes for the two are different and modes of action are independent. Basically, using the same population to test two drugs.

  6. What is the formula for efficacy?

    What is the formula for number needed to treat (NNIT)

    Rate = mortality rate. The rates are %.
  7. What part is external validity? What part is internal validity in a randomized trial?
  8. What are ethical considerations?
    1. Is randomization ethical? Yes if we don't know if Rx will work

    2. Is it ethical to randomize? Yes or else we may use ineffective treatments on patients.

    3. Can patients really provide informed consent?

    4. Should trial be stopped earlier than planned?

    5. Should placebo be used?

    6. Benefits MUST outweigh human experimentation.
  9. What are advantages of community trials? (1)

    Disadvantages? (3)
    1. REpresent only way to estimate directly impact of change in behavior or modifiable exposure on incidence of disase


    1. Inferior to clinical trials with respect to ability to control entrance into study, delivery of intervention, and monitoring of outcomes

    2. Fewer study units are capable of being randomized, affecting comparability

    3. Affected by population dynamics, secular trends, and nonintervention influences.