L1: Research question
Medical research: the ultimate goal is to improve medical practice
Prevention/risk factor: etiology
o Is a high caloric diet a risk factor for cardiovascular disease?
Diagnosis: focuses on making a diagnosis
o What is the probability of having a hip fracture if the affected leg is shorter
and in exorotation?
Treatment: effects of medical treatment
o Does chloroquine treatment reduce risk of mortality among COVID-19
patients admitted to the ICU?
Prognosis (how does the disease develop)
o What is the probability of dying within 5 years after breast cancer diagnosis?
A good research question should be answerable!
Components of research questions:
PICO
Patient/population
Intervention
Comparator
Outcome (clinical outcome you’re
interested in)
DDO
Domain
Determinant
Outcome
L2: Randomized trials
Regression to the mean:
Outcome = combination of
different phenomena
What is needed to identify treatment effect?
Compare 2 or more groups
Those groups should be comparable with respect to NC, EF and V
o And differ only with respect to treatment
,In that case, an observed difference in the outcome between the groups can be attributed to
the only aspects that the two groups differ on: the treatment (T)
We need comparability with respect to the natural course of the disease, extraneous factors,
and error processes.
To achieve comparability, a randomized control trial typically consists of 3 design elements:
Randomization (including concealment of allocation)
Blinding
Standardization
Randomization
The treatment of interest is randomly allocated to participants in a randomized control trial.
Concealment of treatment allocation
Physician who asked patients to participate does not know what treatment the next patient
will receive, nor does the patient himself. As a result, treatment allocation independent of
patient characteristics (so will be truly comparable)
Blinding
Participants should not know which treatment they receive, because that could influence
their behavior
- This also applies to treating physicians / nurses / relatives, etc.
- Binding aims to keep the groups comparable during follow-up
Concealment of treatment allocation occurs before and during patient enrollment. It
ensures that the person enrolling participants does not know which treatment the next
participant will receive, preventing selection bias. This is achieved through methods like
sealed envelopes or centralized randomization.
Blinding occurs after participants are assigned to a treatment group. It ensures that
participants, healthcare providers, and/or researchers do not know who is receiving which
treatment, reducing performance and detection bias.
In short, concealment prevents bias at the randomization stage, while blinding prevents
bias during the study and analysis.
Voorbeelden van blinding zijn placebo (tasts/looks/smells exactly like the active treatment
but does not contain the active compound)
Of active comparator: een ander geneesmiddel gebruiken met dezelfde werking
Standardisation
All procedures, measurements, and treatments are carried out in a consistent and uniform
way. This ensures that the results are reliable, comparable, and not influenced by external
factors minimizes error processes and improves interpretability of treatment effect
, L3: Sample size calculation
Randomized clinical trial (RCT): aim is to compare two treatments, where patients are
recruited and randomized to treatment A or B
How many patients are needed?
- Too few: not able to detect differences
- Too many: costs too much money, not ethical
Practical factors for deciding sample size:
o Number of eligible patients treated at center;
o Number of patients willing to participate
o Time
o Money
Statistical: how big of an effect can be detected with a given number of patients?
Hypothesis testing
1) Decide on a null hypothesis H0 about the population
H0: there is no difference between groups
2) Take a representative sample of the population
3) Calculate the observed difference in the sample
4) Calculate the p-value, the probability to observe at least this difference if H- is true.
This is done by a statistical test
5) If p-value is small, smaller than a prespecified value ‘α’ we reject H0.
Value ‘α’: significance level
Type 1 error: H0 is rejected, but H0 is true in the population (probability: α)
Type 2 error: H0 is not rejected, but it is not true in the population (probability: β)
Power: the probability of finding a significant effect in your sample when the effect is really
present in the population. Depends on:
Relevant difference (effect size)
Sample size
Standard deviation
Medical research: the ultimate goal is to improve medical practice
Prevention/risk factor: etiology
o Is a high caloric diet a risk factor for cardiovascular disease?
Diagnosis: focuses on making a diagnosis
o What is the probability of having a hip fracture if the affected leg is shorter
and in exorotation?
Treatment: effects of medical treatment
o Does chloroquine treatment reduce risk of mortality among COVID-19
patients admitted to the ICU?
Prognosis (how does the disease develop)
o What is the probability of dying within 5 years after breast cancer diagnosis?
A good research question should be answerable!
Components of research questions:
PICO
Patient/population
Intervention
Comparator
Outcome (clinical outcome you’re
interested in)
DDO
Domain
Determinant
Outcome
L2: Randomized trials
Regression to the mean:
Outcome = combination of
different phenomena
What is needed to identify treatment effect?
Compare 2 or more groups
Those groups should be comparable with respect to NC, EF and V
o And differ only with respect to treatment
,In that case, an observed difference in the outcome between the groups can be attributed to
the only aspects that the two groups differ on: the treatment (T)
We need comparability with respect to the natural course of the disease, extraneous factors,
and error processes.
To achieve comparability, a randomized control trial typically consists of 3 design elements:
Randomization (including concealment of allocation)
Blinding
Standardization
Randomization
The treatment of interest is randomly allocated to participants in a randomized control trial.
Concealment of treatment allocation
Physician who asked patients to participate does not know what treatment the next patient
will receive, nor does the patient himself. As a result, treatment allocation independent of
patient characteristics (so will be truly comparable)
Blinding
Participants should not know which treatment they receive, because that could influence
their behavior
- This also applies to treating physicians / nurses / relatives, etc.
- Binding aims to keep the groups comparable during follow-up
Concealment of treatment allocation occurs before and during patient enrollment. It
ensures that the person enrolling participants does not know which treatment the next
participant will receive, preventing selection bias. This is achieved through methods like
sealed envelopes or centralized randomization.
Blinding occurs after participants are assigned to a treatment group. It ensures that
participants, healthcare providers, and/or researchers do not know who is receiving which
treatment, reducing performance and detection bias.
In short, concealment prevents bias at the randomization stage, while blinding prevents
bias during the study and analysis.
Voorbeelden van blinding zijn placebo (tasts/looks/smells exactly like the active treatment
but does not contain the active compound)
Of active comparator: een ander geneesmiddel gebruiken met dezelfde werking
Standardisation
All procedures, measurements, and treatments are carried out in a consistent and uniform
way. This ensures that the results are reliable, comparable, and not influenced by external
factors minimizes error processes and improves interpretability of treatment effect
, L3: Sample size calculation
Randomized clinical trial (RCT): aim is to compare two treatments, where patients are
recruited and randomized to treatment A or B
How many patients are needed?
- Too few: not able to detect differences
- Too many: costs too much money, not ethical
Practical factors for deciding sample size:
o Number of eligible patients treated at center;
o Number of patients willing to participate
o Time
o Money
Statistical: how big of an effect can be detected with a given number of patients?
Hypothesis testing
1) Decide on a null hypothesis H0 about the population
H0: there is no difference between groups
2) Take a representative sample of the population
3) Calculate the observed difference in the sample
4) Calculate the p-value, the probability to observe at least this difference if H- is true.
This is done by a statistical test
5) If p-value is small, smaller than a prespecified value ‘α’ we reject H0.
Value ‘α’: significance level
Type 1 error: H0 is rejected, but H0 is true in the population (probability: α)
Type 2 error: H0 is not rejected, but it is not true in the population (probability: β)
Power: the probability of finding a significant effect in your sample when the effect is really
present in the population. Depends on:
Relevant difference (effect size)
Sample size
Standard deviation
