Clinical Research Management

1
Clinical Research Management
Apinun Aramrattana, MD, PhD Department of Family
Medicine, Faculty of Medicine, Chiang Mai
University
Objectives Students can understand on how to get
clinical research done satisfactorily?
2
Clinical Research Management
How to get clinical research done
satisfactorily? Do It 4 Goods 1.Good
Reasons 2.Good Thoughts 3.Good Plans 4.Good
Practices
3
1.Good Reasons
Why do we need/have to it?
Abnormality
Diagnosis
Frequency
Treatment
Risk
Prognosis
Prevention
Causation
4
Abnormality
Diagnosis
Frequency
Risk
1.Good Reasons
Prognosis
Treatment
Prevention
Cause
Outcomes of Interest
Death
Disease
Which one(s)? How much?
Disability
Discomfort
Dissatisfaction
Destitution
5
Abnormality
Diagnosis
Frequency
1.Good Reasons
Prevention
Prognosis
Treatment
Expected Outcomes
Death
Disease
Disability
Discomfort
Dissatisfaction
Destitution
2.Good Thoughts
Details of problem
Describe
Size of Problem
Observe
Compare
Strategies / Designs
Treatment
Experiment
Test
Prevention
6
2.Good Thoughts
Outcome
Bias Chance Truth

X
X
X
X

Systematic error
Bias

Chance
Statistical (random) error
X
X
X
X
7
2.Good Thoughts
Statistical error
mean
1 SD
-1 SD
2 SD
2 SD
68.26
95.44
8
2.Good Thoughts
Statistical error
9
True value
2.Good Thoughts
Systematic error
Study value
mean
2 SD
2 SD
Statistical error
10
True value
2.Good Thoughts
Systematic error
Study value 1
Study value 2
mean
2 SD
2 SD
2 - Valid but less reliable
1 - Invalid but more reliable
Statistical error
11
2.Good Thoughts
Is this X abnormal?
Abnormality
Normal range ?
Related to a disease?
symptom J
sign K
Symptoms signs?
Symptom X
.
Diagnostic criteria?
Sensitivity specificity?
Who recorded?
Retrospective study?
How good ?
12
Diagnosis
2.Good Thoughts
How good this test X is?
Sensitivity
Accuracy
Specificity
Positive predictive value
Negative predictive value
13
Diagnosis
2.Good Thoughts
How good this test X is?
Gold standard
Definite Dx.
Non-Dx.
Sensitivity
Specificity
???? ???????
a
d
( a )
- ( b )
- ( d )
( c )
a b
c d
a
b
d
c
14
Descriptive Study Case reports Case series
2.Good Thoughts
Frequency

• details of the problem Y
• results of lab. in the records
• use clinical details in the records
• new surveys/records
• simple and cheap

Systematic studies Prevalence / Incidence
studiesCorrelational studies
15
2.Good Thoughts
Frequency
How frequent we find Y in X populations ?
Retrospective study
Dx .. ????? ... ...
Dx .. ????? ... Y...
Dx .. ????? ... ...
Dx .. ????? ... ...
Y
Y
Prevalence

x
16
2.Good Thoughts
Frequency
How frequent we find Y in X populations ?
Prospective study
Y
x
x
x
x
Y
x
x
x
Y
Y
Incidence

x
Time
17
2.Good Thoughts
Frequency
How frequent we find Y in X populations ?
Validity?
How to dx. Y? how is it recorded?
Retrospective study
X who? Where from ?
Y
Prevalence

x
Prospective study
Adequate follow-up?
Y
Incidence

x
Time
18
2.Good Thoughts
Frequency
Size of Risk / Prognosis
Validity?
Retrospective study
Prospective study
Y
x
x
x
x
Y
x
x
x
Y
Death before arrival
Incidence
Prevalence
19
Risk / Prognostic factors
Cohort study
Is Y a bad prognostic factor ?
Y
Relative Risk
Y
Y is a bad prognostic factor
R R gt 1
20
Risk / Prognostic factors
Rare condition
Is Y a bad prognostic factor ?
Y
? ! ?
Y
?
Relative Risk
Too long Cannot follow-up
21
Risk / Prognostic factors
Rare condition
Is Y a bad prognostic factor ?
Retrospective study
Odds
Y
Case control study
Odds Ratio
Y
O R gt 1
Y is a bad prognostic factor
22
Risk / Prognostic factors
Rare condition
Is Y a bad prognostic factor ?
?
How ask?
Retrospective study
Y
Recall bias?
Odds Ratio
?
Y
Choices of controls ?
23
2.Good Thoughts
Treatment
Is A a good treatment?
Regression to mean ?
unpredictable?
attention?
A
P2 lt P1
Problem at baseline ( P1 )
Problem at follow up ( P2 )
Uncontrolled trial
24
Controlled Trial
Randomized
Treatment
Regression to mean ?
unpredictable?
attention?
A
Pt lt Pc
T
P
Randomization
C
25
Controlled Trial
Randomized
Treatment
Compliance
Inclusion
Contamination
A
Cointervention
Pt lt Pc
T
P
Randomization
C
Blinding
Exclusion
26
Treatment
Cross-over Randomized Controlled Trial
Wash out period
Inclusion
P
P
P
P
P
T
C
Pt lt Pc
Randomization
Blinding
P
P
P
P
P
C
T
Exclusion
27
Prevention
Randomized Controlled Trial
Phase of trial I - Healthy volunteers -
safety? II - Volunteer patients - small
number trial III - Selected patients -
efficacy trial
- effectiveness trial
28
Cause
Associated factor is not a cause.
Temporality
Large relative risk
Dose-response effects
Reversibility
Consistency
Biological plausibility
Specificity
Analogy
29
Bias
1. Selection bias
2. Measurement bias
3. Confounding bias
30
Selection Bias
- Sampling for descriptive study
- Sampling for observation of comparison
- Randomly allocation for intervention study
31
Measurement Bias
- Hard data soft data
- Select highest sensitivity and specificity
- Reduce inter-observer variation
32
Confounding Bias
False association
Main question factor y
Y
Effect
Association
Z
Another factor factor Z
Association Confounding
33
Internal Validity
Validity

Chance
External Validity
Conclusion
34
Power of the study 1 ß 50
Diff. reallygt 0?
0
Difference
H0
H1
Type I (?) error 5 Significant
35
N gt n
Power ?
0
Difference
H0
H1
Type I error lt5 Sig. difference
Type II error 20 No sig. difference
36
3.Good Plans
WHO Recommended format for a Research Protocol

• Project summary
• General information
• Rationale
• Study goals and objectives
• Safety Considerations
• Follow-Up
• Data Management and Statistical Analysis
• Quality Assurance
• Expected Outcomes of the Study
• Dissemination of Results and Publication Policy
• Duration of the Project

37
3.Good Plans
WHO Recommended format for a Research Protocol

• Problems Anticipated
• Project Management
• Ethics
• Informed Consent Forms
• Budget
• Other support for the Project
• Collaboration with other scientists or research
  institutions Curriculum Vitae of investigators
• Other research activities of the investigators
• Financing and Insurance

38
4.Good Practices

• Good Clinical Practice, GCP
• Good Laboratory Practice, GLP
• Good Participatory Practice, GPP

Good Papers !