Module 2 Key:

Question 1

True or False

Le CRP n’est pas une réponse à la question de recherche clinique

Answer 1

True

PRC is variable we are measuring : does drug lower cardiovascular disease in these patients. What I am measuring are cholesterol levels. This measurement is not the answer to the question but a necessary measurement.

Question 2

What are the types of CRP?

Answer 2

Dichotomique (deux valeurs possibles)

Ex : succès/échec au traitement

Continue

Ex : tension artérielle, en mm Hg

Incidence (temps de survenue des événements)

Ce CRP est particulier car on s’intéresse à la fois à la survenue d’un événement spécifique durant le suivi d’un participant (ex : événement cardiovasculaire, infection au VIH, rémission cancer, décès, etc.) et au moment de la survenue de l’événement (ex : 1 an après le début des traitements).

Le suivi des participants aura donc une durée maximale, après laquelle l’étude se
terminera. Cet arrêt de suivi sous-entend 2 groupes de participants

1. Un groupe dont le CRP est survenu avant la fin de l’étude
   
   1. Either they died or got the virus etc.,
2. Un groupe dont le CRP n’est pas survenu avant la fin de l’étude
   
   1. Administrative censoring!! (censure administrative)

Ordinale

Une variable ordinale exprime le rang; c’est l’ordre qui confère une information, non la valeur. Par exemple, le stade d’un cancer décrit l’étendue de la maladie dans le corps. Par contre, l’ordre ou le rang est important; 4 est un stade pire que 3, etc. Une variable ordinale est un type de variable qualitative mais l’ordre confère une information importante qui statistiquement peut être utilisée.

Dénombrement (Enumeration)

A count variable (sometimes called adiscretevariable ) takes non-negative integer values and in most contexts non-continuous integer values.Here are examples of count variables:

1. Viral load: The quantity of virus present in an individual’s body, generally expressed as the number of copies of the virus per unit volume of biological fluid (such as blood or saliva).
2. Blood Cells:Such as the number of red blood cells, white blood cells, or platelets in a blood sample.
3. Number of lesions or tumors:In medical imaging, the count of lesions, tumors or abnormalities.

Catégorielle

The order (rank) as well as the value do not confer any information for a categorical variable.This variable is qualitative, not quantitative. The classic example of a categorical variable is blood type (0, A, B, and AB).

Question 3

L’incidence est aussi appelée le ___________________

Answer 3

taux d’incidence.

ex: 10 infections pour 100 personnes-années

Question 4

Define and give example:

Différence de proportions

Answer 4

Après une période suivant le recrutement, on mesure quantitativement une variable biologique dichotomique (ex : succès/échec au traitement, selon une définition spécifique). Cette mesure est une proportion.

Si le CRP est le décès (CRP dichotomique), la différence de proportions est la différence entre la proportion de décès observé dans le groupe expérimental et la proportion de décès observée dans le groupe placebo.

Question 5

Describe Cumulative impact vs incidence rate using an example:

Answer 5

Consider a river with water flowing. Water flows at a certain rate; e.g. 100 liters per minute. The flow rate is similar to the incidence rate. If you divert the river into a reservoir for 10 minutes, you will accumulate 1000 litres of water in 10 minutes. This is similar to cumulative impact (cumulative 10-minute impact).

Similarly, in a population where HIV infection is of interest, infections occur at a certain rate over time (e.g., 10 infections per 100 person-years).

Question 6

Explain this:

The incidence rate is often presented with the % sign, but it is not a proportion.

Answer 6

The % sign is shorthand for “per 100 person/year”.

Question 7

Difference Incidence vs Prevalence:

Answer 7

Incidence: What is the chance of getting this disease?

2 people per 100 per year

Prevalence: In a given population how many people have this disease?

20 000 out of 400 000 have this disease

Question 8

The incidence may increase over time while the prevalence may decrease

Give an example

Answer 8

(e.g., if new treatments are introduced into the population then you can still be as likely to get it but then you can cure it)

Question 9

Give an example:

the incidence may decrease over time and the prevalence may increase

Answer 9

(e.g., if new treatments increase survival).

Question 10

True or False

incidence rate can be > 100%.

Answer 10

True

For example, chlamydia can be contracted more than once by the same person, so the number of incident cases of chlamydia infections may exceed the number of person-years.

Question 11

What design is this?
consists of comparing 2 interventions, one of which is the intervention of interest (e.g. new experimental treatment) and the other is the control intervention (placebo or active control). Treatment is randomly allocated to participants (randomization).

Answer 11

Randomized Controlled Clinical Trial with 2 Parallel Groups

Question 12

Given this experiment describe the objective:
standard treatment is: (Ibuprofen + Placebo). The two experimental treatments are: (Morphine + Ibuprofen) and (Morphine + Placebo)

Answer 12

The two main objectives of this study were (1) to compare treatment (Morphine + Ibuprofen) to treatment (Ibuprofen + Placebo) and (2) to compare treatment (Morphine + Placebo) to treatment (Ibuprofen + Placebo)

Question 13

Why must a placebo be included with each intervention here:
standard treatment is: (Ibuprofen + Placebo). The two experimental treatments are: (Morphine + Ibuprofen) and (Morphine + Placebo)

Answer 13

in order to ensure masking of the treatments, a placebo must be included in the treatments with a single treatment. It is not possible to mask the Ibuprofen treatment alone and the Morphine treatment alone with a group with two treatments (Morphine + Ibuprofen).

Question 14

Describe the principle of Design 3:Crossover Clinical Trial:

Answer 14

Design 3 uses each patient as their own control. With this design:

• Each patient is exposed to more than one procedure
• In the case of 2 procedures, each patient is exposed to the 2 interventions, one after the other (2 time periods)
• A withdrawal period is necessary in order to eliminate all effects (harmful or beneficial) of one intervention before starting another intervention
• The order in which a patient receives the interventions is randomized (placebo before experimental or experimental before placebo). This is different from Designs 1 and 2 where the allocation of treatments to participants is randomized.

Question 15

Why Design 3 cannot be used with “end-point” outcome criteria?

Answer 15

Since the patient must return to their initial state after weaning, Design 3 cannot be used when treatments have the potential to completely cure the patient or patient dies after 1 treatment.

Question 16

Bias vs variability vs type of error

Answer 16

Biasis asystematicerror while precision (variability) is arandom error.

If you take only old people in one group they will all have the same features

Question 17

Describe confounding bias:

Answer 17

This occurs when the difference observed is not necessarily caused by the intervention but rather by the non-comparability of the participants in terms of certain factors, such as age.
Random allocation of participants to interventions (randomization) reduces confounding bias but not all other types of bias.

Question 18

Describe internal validity:

Answer 18

Study will be qualified as excellent if the response is precise and unbiased. On the other hand, a study has low internal validity if the response is imprecise and/or there are multiple potentials for bias.

Question 19

Describe attrition and issue

Answer 19

attrition: When people leave study

if there is attrition, it is possible that the groups are no longer comparable and that attrition induces confounding biases.

Randomization alone does not guarantee control of confounding bias. Attrition must also be minimized. An RCT with high internal validity will therefore necessarily have very little attrition. It is therefore not surprising that a great deal of effort is put into minimizing attrition in RCTs.

Question 20

What can help reduce Bias caused by incorrect and/or subjective measurement of CRP:

Answer 20

A single-blind study indicates that the participant is unaware of the intervention they are receiving. Placebo and intervention are masked where they have the same shape, color, dosage, etc. Double-blind, the study staff is also masked.

Question 21

Describe the concept of External Validity:

Answer 21

For example, if an RCT enrolled patients between the ages of 18 and 25 only, how can it be determined whether the response is applicable to patients >25 years of age or <18 years of age? We have no evidence that the procedure is effective and safe in a patient who is 40 years old (for example). None of the patients in the study were > 25 years old. This limits the generalization of the results to a more heterogeneous patient population.

Typically, the more exclusion criteria there are in an RCT, the more external validity will be affected.

Question 22

True or False
Un seul traitement actif peut être évalué dans un essai clinique randomisé avec contrôle placebo et groupes parallèles.

Answer 22

False

You can test multiple treatments and multiple placebos