Class Notes (1,100,000)
US (480,000)
Rutgers (10,000)
22:49 (100)
Lecture 4

11:709:493 Lecture Notes - Lecture 4: Reporting Bias, Random Sequence, Selection Bias


Department
Nutritional Sciences
Course Code
11:709:493
Professor
Tomeska
Lecture
4

This preview shows page 1. to view the full 5 pages of the document.
Cochrane Risk of Bias Tool (for use with clinical trials)
Domains and Questions
- Selection bias
o could the selection of participants bias the outcome?
- Performance bias
o Could knowledge of group assignment bias the results?
- Detection bias
o Could knowledge of group assignment affect the measurement and
analysis?
- Attrition bias
o Could loss of data bias the results?
- Reporting bias
o Did authors report only certain outcomes to support their main
point?
- Other bias
o In nutrition related research was diet intake measurement
adequate for the purpose of the article?
Selection Bias
- Random sequence generation
o Ideal: no way to anticipate future assignments of subjects to groups
Assignment may be anticipated for several reasons:
Knowledge of a deterministic assignment rule, such as
by alteration, date of birth or date of admission
You're Reading a Preview

Unlock to view full version

Only page 1 are available for preview. Some parts have been intentionally blurred.

Knowledge of the sequence of assignments, whether
randomize or not (if a sequence of random assignments
is posted on the wall)
Ability to predict assignments successfully, basedo n
previous assignments (which may sometimes be
possible when randomization methods are used that
attempt to ensure an exact ratio of allocations to
different interventions)
- Allocation concealment
o Just because you randomized DOES NOT mean you’re finished
o Efforts made to generate unpredictable and unbiased sequence are
likely to be ineffective if they’re not protected by adequate
concealment of the allocation sequence from those involved in
enrollment and assignment or participants
Knowledge of the next assignment (for example a table of
random numbers openly posted on the bulletin board can
cause selective enrollment on the basis of prognostic factors;
participants may be deliberately directed to the “appropriate”
intervention by delaying a participants entry into the trial until
the next appropriate allocation appears
Deciphering of allocation schedules may occur even if
concealment was attempted (unsealed allocation envelopes
may be opened, translucent envelopes may be held against a
bright light)
You're Reading a Preview

Unlock to view full version