Transcript Clinical Trials: a Phase of Product Development

Study design
P.Olliaro
Nov04
Study designs:
observational vs. experimental studies

What happened?


What’s happening?


Case-control study
Cross-sectional study
What will happen?


Cohort study
Clinical trial
What happened? Case-control study
Exposed
Cases
Non-exposed
Exposed
Controls
Non-exposed
Time
Onset of study
Direction of enquiry
What is happening? Cross-sectional study
With oucome
Subjects
Selected for
Study
Without outcome
Time
Onset of study
No direction of enquiry
What will happen? Cohort study
Exposed
OR
Subjects
With outcome
Without outcome
Cohort
Selected
For Study
With outcome
Unexposed
OR
Controls
Without outcome
Time
Onset of study
Direction of enquiry
Randomised Controlled Clinical Trial
With outcome
Experimental
Subjects
Without outcome
Subjects
Meeting
Entry
Criteria
With outcome
Controls
(Treated OR
Untreated)
Onset of study
Without outcome


 Intervention
Direction of enquiry
Time
Trial profile for Controlled Clinical Trials
(e.g. Malaria) - Patients attrition
Total patient population
(# screened)
# Non eligible
(reasons:…)
Total # patients in trial
(# eligible, randomised)
# treated
Test intervention
Withdrawals:
# treatment failures
# lost to follow up
# adverse event
# others
# with outcome
on day X
# controls
(placebo, std Tx)
Withdrawals:
# treatment failures
# lost to follow up
# adverse event
# others
# with outcome
on day X
Issues in design & interpretation of
clinical trials

Randomisation


Overemphasis on significance testing




Treatments still developed/recommended without properly
randomised trials
“magical” p=0.05 barrier. P-values only a guideline to the
strength of evidence contradicting the null hypothesis of no
treatment difference – NOT proof of treatment efficacy
Use interval estimation methods, e.g. confidence intervals
Often trial generate too many data (e.g. interim & subgroup
analyses, multiple endpoints) & significance tests
Size of trial


Often trials do not have enough patients to all reliable
comparison
At planning stage, power calculation should be used realistically
(but often produce samples >> number of patients available!)
Checklist for Assessing Clinical Trials
General Characteristics



Reasons the study is needed
Purpose/Objectives: Major & Subsidiary
Type: Experimental, Observational


Phase: I, II, III, IV, other
Design: Controlled, Uncontrolled
Checklist for Assessing Clinical Trials
Population





Type (Healthy volunteers; Patients)
How chosen/recruited?
Entry/eligibility criteria: Inclusion, Exclusion
Comparability of treatment groups:
demography, prognostic criteria, stage of
disease, associated disease, etc
Similarity of participants to usual patient
population
Checklist for Assessing Clinical Trials
Treatments compared




Dose rationale & details
Dosage form & route of administration
Ancillary therapy
Biopharmaceutics: source, lot No (Test &
Standard medications/Placebo)
Checklist for Assessing Clinical Trials
Experimental Design



Controls (active/inactive; concurrent/historical)
Assignment of treatment: randomised?
Timing
Checklist for Assessing Clinical Trials
Procedures
Terms & measures
 Data quality
Common procedural biases:
 Procedure bias
 Recall bias
 Insensitive measure bias
 Detection bias
 Compliance bias

Checklist for Assessing Clinical Trials
Study outcomes & interpretation



Reliability of assessment
Appropriate sample size
Statistical methods




Use for what? Questions re: differences?
Associations? Predictions?
“fishing expedition”
Multiple significance tests
Migration bias
Checklist for Assessing Clinical Trials
Data Collection




Measurements used to assess goal attainment
(Appropriate type? Sensitivity? Timing?)
Observers (Who? Variable?)
Methods of collection (Standard?
Reproducible?)
Adverse events: Subjective (volunteered,
elicited?); Objective (laboratory, ECG, etc)
Checklist for Assessing Clinical Trials
Bias control
Bias = measurement or systematic errors (≠ random
errors)
 Subject selection






Prevalence or incidence (Neyman) bias: e.g. early fatalities,
“silent” cases
Admission rate bias (Berkson’s fallacy): distortions in RR
Non-response bias or volunteers effect
Procedure selection bias
Concealment of allocation
Blinding:



Subjects
Observers
Others
Checklist for Assessing Clinical Trials
Results






Primary outcome measures
Secondary outcome measures
Drop outs (reasons, effects on results)
Compliance: Participants (with treatment);
Investigators (with protocol)
Subgroup analysis
Predictors of response
Checklist for Assessing Clinical Trials
Data analysis



Comparability of treatment groups
Missing data
Statistical tests: if differences observed, are
they clinically meaningful? If no difference,
insufficient power?
Study
Participant
Jour
Antécedents médicaux
Examen clinique
Température
Frottis / Goutte Epaisse
Papier filtre pour PCR
Papier filtre pour PK
Hématologie
Biochimie
0
X
X
X
X
X
X
X
X
1
2
3
7
14
21
28
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Hospital
files
Report
Entry 1
Analysis
CRF
Entry 2
Publication