Chapter 10- Randomized trials

Question 1

Experimental study

Answer 1

A type of study where an investigator manipulates a treatment/factor of interest
(exposure, risk factor, etc.) and measures health outcomes

Question 2

Randomized trials

Answer 2

Start with a defined population. Participants are randomly assigned to a treatment group. Compare risk of disease outcome among people with and without treatment to determine whether receiving treatment is related to the outcome

Question 3

Limitations/considerations of randomized trials (3)

Answer 3

1. Ethics in clinical research
2. Selection of subjects
3. Treatment concealment + control regime

Question 4

Belmont report

Answer 4

Contained 3 key principles:
Respect for Persons
Beneficence
Justice

Question 5

Respect for persons

Answer 5

The idea that individuals should be treated as autonomous

Question 6

Beneficence

Answer 6

Effort should be made to reduce harm and maximize benefits for participants. This is a risk-benefit assessment where expected benefits should far outweigh the potential harms.

Question 7

Justice

Answer 7

Benefits of research should be evenly distributed among all involved species. There should be clear principles for inclusion/exclusion
of populations in research- eligibility criteria in a study is an example of this

Question 8

3 elements of informed consent

Answer 8

1. Information
2. Comprehension
3. Voluntariness

Question 9

Selection of subjects in RCTs

Answer 9

The criteria for determining who will and will not be included in the study should be described in writing before the study begins. There should be no element of subjective decision making on the part of the investigator.

Question 10

Masking (blinding)

Answer 10

Subjects should not know which group they are assigned to, especially when the outcome is a subjective measure. This can be accomplished using a placebo. The observers or data collectors should also be blinded in regard to which group the patient is in

Question 11

Considerations for selection of control regime

Answer 11

1. Ethical: What is standard of care?
2. Practical: What control can help to maintain blinding?

Question 12

How does randomization reduce bias?

Answer 12

Assignment of treatment group due to chance alone. It aims to reduce potential bias in treatment-outcome relationships that can arise consciously or unconsciously from investigators or participants

Question 13

Planned crossover

Answer 13

Subjects are randomized to new treatment or current treatment. They are observed for a certain period of time. After any changes are measured, patients are switched to the other therapy and observed again. Each patient serves as their own control

Question 14

Carryover

Answer 14

A potential concern of a planned crossover design- once treatment is switched, there should not be any residual effects from the initial treatment impacting the outcome of the second treatment. The order in which the treatments are given may cause psychological effects that can impact treatment

Question 15

Unplanned crossover

Answer 15

Subjects randomized to one treatment (like surgery) might eventually decide not to have the surgery. Some controls may have their health decline and then need surgery. Therefore, patients may need to be switched between groups. This may make it challenging to analyze the data

Question 16

2 data analysis options when patients crossover

Answer 16

1. “Intent-to-treat” (treatment assigned at start of study)
2. “As treated” (actual treatment received)

Question 17

How does randomization address information bias?

Answer 17

“Blinding” of participants and investigators helps to minimize outcome because treatment
assignment is unknown

Question 18

Special design or randomization features of randomized trials (4)

Answer 18

1. Stratified randomization
2. Cluster randomization
3. Intentional crossover design
4. Factorial design

Question 19

Stratified randomization

Answer 19

Refers to the situation in which strata are constructed based on values of prognostic variables and a randomization scheme is performed separately within each stratum.

Question 20

Cluster randomization

Answer 20

Pre-existing groups, called clusters, of individuals are randomly allocated to treatment arms. For example, clusters may be clinical practices or schools where the individuals are patients and school children, respectively. CRTs can be used when individual randomization to treatment arms is not possible or the intervention is naturally applied to a whole cluster.

Question 21

Factorial design

Answer 21

If 2 drugs are to be tested, the anticipated outcomes for the 2 drugs are different, and their modes of action are independent, researchers can use the same study population to test both drugs. A 2X2 design can be used, where one condition receives drug A and B, one receives B only, one receives A only, and one receives neither. Participants go through all 4 conditions