Lecture 1 research question
Medical research
Prevention/risk factor (etiology)
Is a high caloric diet a risk factor for cardiovascular disease?
Diagnosis
What is the probability of having a hip fracture if the affected leg is shorter and in
exorotation?
Treatment
Does chloroquine treatment reduce the risk of mortality among COVID-19 patients
admitted to the ICU?
Prognosis
What is the probability of dying within 5 years after breast cancer diagnosis?
Research questions
- Arise in practice
- Starting point when designing a study/reading a paper
- Answerable
- Standard elements
Components
PICO
Patient (population)
Intervention
Comparator
Outcome
DDO
Domain
Determinant
Outcome
Example 1
Treatment, because it is about comparing two treatment options.
,Example 2
Diagnosis, because it is about suspecting a certain disease.
Example 3
Etiology, because it is about a risk factor.
Lecture 2 randomised controlled trial
Study to measure the effect of a new diuretic drug for patients with hypertension
- Take a group of people with hypertension
- Give one group the new diuretic drug and the other group not
- Measure if the blood pressure decreased
Classification: treatment
P – patients with hypertension
I – new diuretic drug
C – no diuretic treatment
O – lower blood pressure
Reasons why the blood pressure has normalized
- Started the pharmacological therapy
- Stopped smoking
- Started going to the gym
- More adherent to his beta blocker treatment
- Afraid of the doctor the first time (white coat effect)
, - Blood pressure fluctuates over time
- The second time a more precise measurement method was used
Regression to the mean
You select patients with high blood pressure in a peak, after this moment, the blood
pressure can only go down.
Outcome model
Treatment (T)
Natural course (NC)
Extraneous factors (EF)
Error processes (V), natural variation
Our interest lies in the treatment.
Outcome with treatment = T + NC + EF + V
Outcome without treatment = NC + EF + V
We make a comparison between 2 groups, they should be comparable to NC, EF and V. The
only difference is the treatment.
Design elements
Randomisation
Blinding
Standardisation
Randomisation
It needs to be randomly selected who gets the treatment and who doesn’t or treatment A
and B. This is necessary to make the two groups comparable.
How?
Set of envelopes in desk drawer? Risk of looking in the envelope.
Random number generator at computer? Another answer asked, repeat this until the
wanted answer.
Concealment of treatment allocation
Physician or investigator and patient both don’t know what the patient will receive.
As a result, treatment allocation is independent of patient characteristics. Therefore, the
treatment groups are totally comparable.
Blinding
Participants don’t know which treatment they receive, because that could influence their
behaviour. This is also applying to physicians, nurses and relatives.
This is needed to keep the group comparable during follow-up.
Cannot be done for example when one group receives a physical program and the other
doesn’t, because then it is obvious in which group you are.
, 1. Placebo
Tastes, looks, smells the same, but does not contain the active compound.
Can be done with drugs.
This is sometimes difficult, for example with surgery or physiotherapy.
2. Active comparator
Different diuretic treatment for example.
Blinded outcome assessment
The one who measures the outcome should also not know the treatment status.
Example: blood pressure, quality of life
Exception: all-cause death
Standardisation
o Intervention (what dosage, how to administrate, what frequency)
o Concomitant care (clear instructions what physicians can and cannot do)
o Outcome assessment (how should the outcome be measured)
Minimize the error processes.
Improve the interpretability of treatment effect.
Lecture 3 sample size calculations
How many patients do you use in a randomised control trial?
Too few not able to detect differences.
Too many lot of time and money, unethical.
Factors deciding sample size
Practical
o Number of eligible patients treated at center
o Number of patients willing to participate
o Time
o Money
Statistical
o How big of an effect can be detected with a given number of patients?
Medical research
Prevention/risk factor (etiology)
Is a high caloric diet a risk factor for cardiovascular disease?
Diagnosis
What is the probability of having a hip fracture if the affected leg is shorter and in
exorotation?
Treatment
Does chloroquine treatment reduce the risk of mortality among COVID-19 patients
admitted to the ICU?
Prognosis
What is the probability of dying within 5 years after breast cancer diagnosis?
Research questions
- Arise in practice
- Starting point when designing a study/reading a paper
- Answerable
- Standard elements
Components
PICO
Patient (population)
Intervention
Comparator
Outcome
DDO
Domain
Determinant
Outcome
Example 1
Treatment, because it is about comparing two treatment options.
,Example 2
Diagnosis, because it is about suspecting a certain disease.
Example 3
Etiology, because it is about a risk factor.
Lecture 2 randomised controlled trial
Study to measure the effect of a new diuretic drug for patients with hypertension
- Take a group of people with hypertension
- Give one group the new diuretic drug and the other group not
- Measure if the blood pressure decreased
Classification: treatment
P – patients with hypertension
I – new diuretic drug
C – no diuretic treatment
O – lower blood pressure
Reasons why the blood pressure has normalized
- Started the pharmacological therapy
- Stopped smoking
- Started going to the gym
- More adherent to his beta blocker treatment
- Afraid of the doctor the first time (white coat effect)
, - Blood pressure fluctuates over time
- The second time a more precise measurement method was used
Regression to the mean
You select patients with high blood pressure in a peak, after this moment, the blood
pressure can only go down.
Outcome model
Treatment (T)
Natural course (NC)
Extraneous factors (EF)
Error processes (V), natural variation
Our interest lies in the treatment.
Outcome with treatment = T + NC + EF + V
Outcome without treatment = NC + EF + V
We make a comparison between 2 groups, they should be comparable to NC, EF and V. The
only difference is the treatment.
Design elements
Randomisation
Blinding
Standardisation
Randomisation
It needs to be randomly selected who gets the treatment and who doesn’t or treatment A
and B. This is necessary to make the two groups comparable.
How?
Set of envelopes in desk drawer? Risk of looking in the envelope.
Random number generator at computer? Another answer asked, repeat this until the
wanted answer.
Concealment of treatment allocation
Physician or investigator and patient both don’t know what the patient will receive.
As a result, treatment allocation is independent of patient characteristics. Therefore, the
treatment groups are totally comparable.
Blinding
Participants don’t know which treatment they receive, because that could influence their
behaviour. This is also applying to physicians, nurses and relatives.
This is needed to keep the group comparable during follow-up.
Cannot be done for example when one group receives a physical program and the other
doesn’t, because then it is obvious in which group you are.
, 1. Placebo
Tastes, looks, smells the same, but does not contain the active compound.
Can be done with drugs.
This is sometimes difficult, for example with surgery or physiotherapy.
2. Active comparator
Different diuretic treatment for example.
Blinded outcome assessment
The one who measures the outcome should also not know the treatment status.
Example: blood pressure, quality of life
Exception: all-cause death
Standardisation
o Intervention (what dosage, how to administrate, what frequency)
o Concomitant care (clear instructions what physicians can and cannot do)
o Outcome assessment (how should the outcome be measured)
Minimize the error processes.
Improve the interpretability of treatment effect.
Lecture 3 sample size calculations
How many patients do you use in a randomised control trial?
Too few not able to detect differences.
Too many lot of time and money, unethical.
Factors deciding sample size
Practical
o Number of eligible patients treated at center
o Number of patients willing to participate
o Time
o Money
Statistical
o How big of an effect can be detected with a given number of patients?
