Clinical Trials

EPIDEMIOLOGY METHODS II

Sawitri Assanangkornchai, MD, PhD

 Describe the rationale for
conducting a clinical trial
Objectives
Describe the steps in
After completing
designing a clinical trial
this session, the
students should be
able to… Describe the possible
biases in conducting a
clinical trial

Know how to register a

clinical trial

2
 Definition of a clinical trial 3

A clinical trial is
u any investigation in human subjects intended to discover or verify
the clinical, pharmacological, pharmacodynamic effects of the
product(s), the effectiveness of medicine, surgery, radiotherapy,
education program with the object of ascertaining its safety and/or
efficacy.

ICH/GCP group
 4
Phases of drug trials
III. Therapeutic
Confirmatory
I. Human Pharmacology
nlarge-scale, obtain
§ first trial in humans, small n efficacy and safety
of healthy volunteers.
§ preliminary evaluation of
safety, II. Therapeutic Exploratory IV. Therapeutic Use
pharmacokinetic/dynamic ndrug’s efficacy, § post-marketing trials after local
profiles therapeutic dose range and regulatory approval; compares
short-term safety (side efficacy and benefit in “real world”
effects, toxicity) in patients settings

https://www.fda.gov/media/84920/download
Study of the Pharmacokinetics (PK) and 5
Pharmacodynamics (PD) of the drug in humans

– PK: what the body does to the drug (Absorption, Distribution,

Metabolism, Excretion)
– PD: what the drug does to the body

https://www.fda.gov/media/84920/download
 6

Discussion 1
u Why conduct a clinical trial?
u When should we conduct a
clinical trial?
u What are advantages and
disadvantages of clinical trials?
Why conduct a clinical trial? 7

u What kind of research questions can the trial answer?

Hypothesis testing e.g. effectiveness of intervention and a causal
relationship

u Why can the trial establish a causal relationship?

u Temporal sequence
u Confounding and bias are minimized

“RCT gives the strongest evidence for a causal relationship”

 8

Advantages
u The cause precedes the
effects
u Possible confounding
factors do not confuse the Disadvantages
results u Share many of the
u The interventions are disadvantages of cohort
compared efficiently studies due to prospective
collection
u Ethical problems
u May restrict the
generalizability of results
 9

Discussion 2
u Give an example of a clinical
trial that you have come
across
u Identify what are its
u Research question
u Participants
u Intervention
u Control
u Outcome
 Single intervention 10

open label. no control group. in the first phase

u All subjects receive the same intervention for a period of time.

baseline endpoint

Wash out Investigational

(if appropriate) treatment
Parallel group design
u Subjects are randomly assigned to two or more comparable
groups, each of which receives different treatments

placebo

R
drugs

Investigational Wash out bersihkan dari segala effect dari

treatment (if appropriate) obat2an sebelumnya

Placebo/Active R Randomization point

control

11
advantage :
lower sample size
Cross-over design
good for chronic conditions

u Subjects are randomized to groups that receive the same treatments

but in different sequence; the treatment periods may be separated
by a washout period during which no treatment is received
treatmen treatmen

baseline
R
first end poin endpoin
data
measure the outcome outtcome

placebo placebo

Investigational Wash out

treatment (if appropriate)

Placebo/Active R Randomization point

control

12
 Randomized Polypill
Crossover Trial in People
Aged 50 and Over
• Polypill: three blood pressure lowering agents
(amlodipine 2.5 mg, losartan 25 mg and
hydrochlorothiazide 12.5 mg), all at half
standard doses and simvastatin 40 mg
(standard dose)
• Participants: individuals aged 50+ without a
history of cardiovascular disease
• Outcome: blood pressure and low density
lipoprotein (LDL) cholesterol at the end of
each 12 week period

file:///Users/sawitri/Downloads/pone.0041297.pdf
 Factorial design disadvatage : need alot of subject 14

u all doses of one treatment are studied

in combination with all doses of the
other treatment.

Treatment I
Treatment A & Treatment B over naothing

Dose c Dose d
Treatment A & Placebo B treatment a + b over a / b

Treatment II
Dose a
R ac ad
Placebo A & Treatment B

Dose b bc bd
Placebo A & Placebo B
Vitamin D and marine omega 3 fatty
acid supplementation and incident
autoimmune disease: VITAL
randomized controlled trial

§ two-by-two factorial design

§ Vitamin D (2000 IU/day) or

matched placebo, and omega 3
fatty acids (1000 mg/day) or
matched placebo

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8791065/
Cluster
randomizatized
trial
we randomised the cluster (school / clinic)

Community Risk Factors for Ocular

Chlamydia Infection in Niger: Pre-
Treatment Results from a Cluster-
Randomized Trachoma Trial

https://journals.plos.org/plosntds/article?id=10.1371/jou
rnal.pntd.0001586
 Hypothesis 1 1

Superior: a new therapy is

"better" than the established the sample size should be large
standard treatment more than +1

Equivalence: a new treatment

is neither worse or better than
a conventional treatment 0 - -1

Non-inferior: a new treatment

is not inferior to a conventional
treatment -1

http://www.pvanuden.com/2015/04/equivalence-vs-non-inferiority-vs.html
Participants Pyridostigmine
reduces mortality of
patients with severe
u Inclusion criteria SARS-CoV-2
infection: A phase
u Homogeneity 2/3 randomized
u Convenience for follow up controlled trial
u Exclusion criteria
Design: a double-blinded, parallel group randomized,
u Extraordinary subjects : placebo-controlled
contraindication, Participants: Adult (≥ 18-year-old) hospitalized patients with
hypersensitivity, abnormality of positive nasopharyngeal SARS-CoV-2 real-time polymerase-
chain-reaction (RT-PCR) test and imaging study compatible
anatomy or physiology which
with pneumonia, and at least one risk factor for requiring
might influence the result of invasive mechanical ventilation (IMV) or dying, including
the treatment, inability to hypoxemia
participate in the study Intervention: oral pyridostigmine 60 mg/day or a matching
u Sample size estimation placebo for a maximum duration of 14 days, up to hospital
discharge, or until the patient had the primary outcome,
whichever occurred first
Primary outcome: a composite of IMV or death in the 28
days following randomization, and frequency of specific AE

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9644007/ 18
A control group in a 19

clinical trial Types of controls in clinical trials

´ Placebo:
Compares experimental treatment(s) to
u Without a control group, the a dosage form without the experimental
change may be due to treatment.
alternative explanations: ´ No-treatment:
u Predictable improvement Compares two groups, one of which
does not receive the experimental
u Fluctuating disease severity treatment.
´ Dose-comparison:
u Hawthorn effect
Compares two or more active doses of
u Regression to the mean the experimental treatment.
u Practice effect ´ Active:
Compares two or more experimental
treatments.
“No controls, no conclusion” ´ Historical:
Compares observations in a study with
results obtained in an earlier study.
Methods of assigning treatment 20

u Randomization
u Non-randomization
u Stratified
randomization

https://i.stack.imgur.com/wbeHj.jpg
Statistical design of Phase II/III clinical trials for testing
therapeutic interventions in COVID-19 patients

https://bmcmedresmethodol.biomedcentral.com/articles/10.1186/s12874-020-01101-z
Randomization technique 22

u Simple randomization
u tossing a coin, telephone to a randomization center, random
table, random by calculator or computer
u this technique may lead to substantial differences in group size
which will reduce the precision of estimates of the difference.
u Block randomization
u this technique randomizes n individuals into k treatments, in
blocks of size m.
u this will ensure a close balance of the numbers in each group.
Block randomization 23

not blind. becuse we know what

is the next subject

https://www.appliedclinicaltrialsonline.com/view/comparison-techniques-
creating-permuted-blocked-randomization-lists
Different Block sizes

https://www.appliedclinicaltrialsonline.com/view/comparison-techniques-
creating-permuted-blocked-randomization-lists
https://www.randomizer.org/
Mendelian randomization

a method of using
measured variation in
genes of known
function to examine the
causal effect of a
modifiable exposure on
disease in observational
studies

https://journals.sagepub.com/doi/10.5127/pr.038115
 ´ Unblind (open label):

Blinding
Subject, investigator, and evaluator 27
know the treatment.

u A method to keep study ´ Single-blind:

participants, healthcare Either the subject, investigator, or
providers and sometimes evaluator does not know the treatment.
those collecting and
analyzing clinical data ´ Double-blind:
unaware of the assigned Two of the parties (evaluator, subject,
intervention so they will not investigator) do not know the treatment.
be influenced by that
knowledge and therefore ´ Triple blind:
reduce the bias.
Three of the parties (evaluator, subject,
u Blinding is particularly investigator) do not know the treatment.
important when outcome
´ Double-dummy:
measures are subjective.
u Blinding may not always be A method used to enable blinding
appropriate and possible. when two treatments cannot be made
identical in appearance

ACTIVE PLACEBO
Common problems in clinical trial 28

u Compliance, Adherence to treatment: the

extent to which a participant’s behaviors Ø Loss of follow-up: move out from the
comply with medical or health advice study site, low motivation and long
u Behavioural interventions: e.g., diet, exercise, duration of follow up –attrition bias:
meditation –over/under exposure to treatment attrition differs between groups
u Drug interventions: ADR, fear of drug Ø Contamination: members of the
dependence, low motivation ‘control’ group inadvertently receive
the treatment or are exposed to the
u How to increase adherence: designing the
treatment to be easy to follow –small number
intervention
of visits, short duration, one-stop service point, Reduces effect size –Type II error
selecting eligible subjects (concluding –no effect when there
actually is)
u Treatment fidelity: the extent to which the
Ø Co-intervention: treatment/care given
study team complies with the study protocol
in addition to the tested intervention
u To increase fidelity: manual of procedures
Non-differential –decreases power
(MOP), training
Differential –causes bias
 29
Data analysis Intention to treat analysis
u Baseline data Patients who switch treatment group or who fail to
u Intention to treat analysis comply with the prescribed treatment should be
analyzed according to their original group assignment.
u Number needed to treat Advantages:
u Subgroup analysis u Random allocation is preserved.
u Interim analysis u Measure the real-world effect where patient
compliance is typically less than perfect.
u Missing vs. Non-missing
Disadvantages:
u Reliably Change Index (RCI)
u If many patients switch treatment groups,
differences between the groups will be obscured.
u Noncompliance may underestimate the real
benefit of the treatment; additional analyses
may therefore be considered.
u Intention to treat analysis is not appropriate for
examining adverse effects.
Clinical Trial Registry https://clinicaltrials.gov/

• Reduce publication bias

• The ICMJE member journals will
require registration as a condition
of consideration for publication.
• Trials must register at or before the
onset of patient enrollment.

http://www.clinicaltrials.in.th/index.ph
http://www.anzctr.org.au/Default.aspx
p?meun=home&smenu=home
 31
Preparing a clinical trial protocol

http://www.spirit-statement.org/spirit-statement/
Reporting a trial 32

u CONSORT (Consolidated Standards of Reporting Trials)

http://www.consort-statement.org/
Flow diagram of RCT (CONSORT statement)
Assessed for eligibility (n= )

Enrollment
Excluded /not meeting inclusion
criteria / refused to participate/ other
reasons (n= )

Randomized (n= )

Allocated to intervention /did not Allocated to intervention / did not

Allocation

receive allocated intervention (give receive allocated intervention (give

reasons) (n= ) reasons) (n= )
Follow-up

Lost to f/u/ discontinued intervention Lost to f/u/ discontinued intervention

(give reasons) (n= ) (give reasons) (n= )
Analysis

Analysed/ excluded from analysis Analysed/ excluded from analysis

(give reasons) (n= ) (give reasons) (n= )
References 34

u Cummings SR, Grady D, Hulley SB. Designing a Randomized Blinded Trial. In: Hulley SB,
Cummings SR, eds. Designing clinical research, 4th Ed. Philadelphia: Lippincott Williams &
Wilkins, 2013: 137-150.

u Grady D, Cummings SR, Hulley SB. Alternative Clinical Trial Designs and Implementation
Issues. In: Hulley SB, Cummings SR, eds. Designing clinical research, 4th Ed. Philadelphia:
Lippincott Williams & Wilkins, 2013: 151-166.

u Efird J. Blocked randomization with randomly selected block sizes. Int J Environ Res Public
Health. 2011;8(1):15-20. doi:10.3390/ijerph8010015

u The CONSORT Statement:. http://www.consort-statement.org/

u The SPIRIT Statement: https://www.spirit-statement.org/