Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

EVIDENCE-BASED MEDICINE > 4,5 – Clinical Trials > Flashcards

4,5 – Clinical Trials Flashcards

1
Q

How do we select our sample?

A
  • General population
  • Eligible population (study population)
  • Sample
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

External validity

A
  • How applicable the results are to the GENERAL population of interest
  • Impacted greatly by subject selection
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Control group types

A
  • Historical
  • Concurrent
  • Cross-over trials
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Historical control groups

A
  • Before and after
  • Don’t know what else has changed (ex. weather, management factors)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Concurrent control groups

A
  • Formed at the same time
  • “parallel arm trials”
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Cross-over trials

A
  • Utilize the same animals as treatment and control groups
  • Order of treatments must be randomized
  • Need to figure out a wash out period (time between the 2 treatments)
  • Common
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

3 key elements of clinical trial design

A
  1. Outcome measures
  2. Bias
  3. Chance
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Outcome measures

A
  • VARIABLES measured to see if the treatment works or not
  • What is most clinically relevant?
  • Want to be able to measure reliably
  • Want it to be objective
    o Sometimes need subjective (blinding is important and case outcomes)
  • Specific (ex. do post-mortem vs. just died)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

How many outcomes?

A
  • Can be a problem to look at TOO many
  • Author should PRIORITIZE them
  • Want 1 or 2 primary
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Example: possible outcomes for BRD trial

A
  • Mortality: objective
  • Morbidity: case definition
  • Serological conversion
  • Average daily gain
  • Feed efficiency (need to measure at the pen level NOT the individual level)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Experimental unit

A
  • Smallest INDEPENDENT unit to which treatment is allocated
  • Ex. leg of animal, udder quarter, individual animal, pen of animals
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Experimental unit=pen of animals

A
  • Need to analyze results on pen basis NOT the individual animal
  • Ex. 2 pens, n=2 (not that much)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

‘herd immunity’

A
  • Vaccinated or treated animals may protect or reduce challenge to ALL individuals in herd or group
  • If vaccinates and non-vaccinates are COMMINGLED in pen
  • Ex. BVD vaccine trials, anthelmintic trials (half not dewormed=benefits)
  • MINIMIZES DIFFERENCES in outcomes
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

3 reasons for a outcome variable difference between treatments (ex. castration with meloxicam or not)

A
  • Meloxicam is effective at reducing pain
  • Bias is present
    o Something other than the treatment
  • Outcomes variable difference could occur simply by chance
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Bias

A
  • Did some factor other then the treatment cause difference in outcome?
  • A ‘systematic’ difference between the treatment and the control group which could affect the outcome measure
  • If bias=can’t rely on it
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

4 times when bias might occur

A
  1. Selection/randomization bias
  2. Performance bias (cointervention)
  3. Exclusion bias
  4. Detection
17
Q

Selection/randomization bias

A
  • Systematic difference in the treatment and control groups attributable to lack of randomization
18
Q

Performance bias (cointervention)

A
  • Systematic difference in care provided apart from the intervention being evaluate
19
Q

Exclusion bias

A
  • Systematic differences in withdrawals from the trial
20
Q

Detection bias

A
  • Systematic differences in outcome assessment or follow-up
21
Q

Sources of selection bias

A
  • Non-randomized clinical trials
  • Vaccinate the first half of the truck load
  • eliminated by RANDOMIZATION
22
Q

Randomization

A
  • necessary to avoid bias
    o eliminates trial entry biases
  • protects against systematic differences in control and treatment group
  • baseline comparison of know covariate helps to evaluate effectiveness of randomization
23
Q

How to randomize

A
  • Simple randomization
  • Systematic randomization
24
Q

Simple randomization

A
  • Toss a coin
  • Random number tables or number generator
25
Q

Systematic randomization

A
  • Not true randomization
  • Ex. randomizing first animal and then alternating assignment b/w treatment and control groups
26
Q

Confounding variables

A
  • Factors that may be related to outcome measure that are NOT equally distributed between treatment and control groups
  • *proper randomization should balance most of them
27
Q

How to overcome selection bias

A
  • RANDOMIZATION
  • Remove possible bias at many levels
  • Stratified randomization or blocking
28
Q

Stratified randomization or blocking

A
  • Divide into different strata according to key confounding variable (ex. breed, age, sex)
  • Then use simple randomization within each subgroup
  • *especially when sample sizes are SMALL
29
Q

Matching

A
  • Pair of treatment and control subjects chosen to be as similar as possible
  • Ex. age, sex, breed, time of admission
30
Q

Performance and exclusion bias

A
  • Treatment and control may receive UNEQUAL observation or ancillary treatments
  • Some may be lost due to follow-up
31
Q

Problems with ancillary treatment

A
  • Contamination
  • Co-intervention
32
Q

Contamination

A
  • Control patients accidentally receive the experimental treatment
33
Q

Co-intervention

A
  • Performance of additional diagnostic or therapeutic acts on experimental but NOT control patients
34
Q

Detection bias

A
  • Diagnostic decisions can be affected by knowledge or treatment status
  • More important in subjective measures
  • Process to measure may be different=potentially bias the trial
35
Q

Blinding

A
  • Subject and observer unaware of treatment group allocation
36
Q

Single blinding

A
  • Only subjects are unaware of which treatment group they belong to
37
Q

Double blinding

A
  • Subjects and observers are unaware of treatment groups
38
Q

Triple blinding

A
  • Statistical analysis is blinded

EVIDENCE-BASED MEDICINE (8 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions