Critical Appraisal Bias - Randomization/Concealment

Question 1

Benefits of randomizations

Answer 1

Treatment and control group is more balanced in regards to known/unknown prognostic factors

Question 2

Negatives of Observational Studies

Answer 2

Tend to over estimate treatment effects because of unbalanced groups based on prognostic factors

Question 3

Observational Studies and Bias

Answer 3

Can try to match patients based on their prognostic factors
- Problem: Some prognostic factors are unknown at the time

Question 4

Were patient’s randomized
- Effect on risk of bias

Answer 4

Yes, Low risk of bias

Question 5

Randomization concealment

Answer 5

Is not the same as blinding
- Refers to if the process of randomization (Allocation Concealment)

It is important to ensure those involved in gathering participants are separate from the randomization process

Question 6

Importance of Allocation Concealment

Answer 6

Trials that did not use it had 40% larger estimates of effect
- Treatment was overestimated
- Can allocate certain patients with specific prognostic into certain groups

Question 7

Simple Randomization

Answer 7

Equal chance of allocation into both groups
- Does not work with small sample sizes

Question 8

Blocked Randomization

Answer 8

Randomize in blocks to get relatively equal number of patients

Pro: Similar Sized groups to help imbalance
Con: Can be predictable

Question 9

Stratified Randomization

Answer 9

Stratify patients based on known prognostic factors
- Then randomly allocate those within the same strata into treatment groups

Helps balance factors like treatment site, age, sex

Question 10

Blinding vs Allocation Concealment

Answer 10

Blinding is having the patient, researcher, and clinicians unaware of what intervention has been assigned to the patient
- Cant always have blinding

Allocation concealment is preventing the recruiter from predicting the assignment of intervention
- Can always have allocation concealment

Question 11

Strong Concealment Mechanisms

Answer 11

Web Based
Telephone System

Question 12

Method of Sequence

Answer 12

Randomization sequences

Question 13

Prognostic Balance

Answer 13

Have to ensure that patients of the study are from similar prognostic groups or else it can cause bias.
- Patients from hospital compared to patients in community

Question 14

Sample size effect on prognostic factors

Answer 14

The larger the sample size the more similar the groups will be

Question 15

How to acheive prognostic balance

Answer 15

Main purpose of randomization is to achieve prognostically similar groups
- Does not always work

Increasing sample size will make groups more similar

Question 16

How to test if 2 groups are different

Answer 16

P-Value
- 1 in 20 chance to fail

Magnitude of the Differences is the preferred method