Module 4 Key

Question 1

When is a PAS written?

Answer 1

The PAS is usually written and defined before the study is carried out and therefore before the results have been seen.

Question 2

What is the PAS?

Answer 2

Le plan d’analyse statistique (PAS) d’une étude est un document qui définit les principales analyses statistiques. Le PAS est obligatoire dans une demande de subvention et dans le protocole de l’étude.

Question 3

What is the multiplicityproblem?

Answer 3

changing categories/tests in a study based on results so that you can get better results

Question 4

What is the CONSORT diagram.

Answer 4

CONSORT diagram allows you to quickly identify how many participants/patients were approached to participate in the study, how many participated in the study, how many were lost to follow-up, and how many were included in the primary CRP analysis.

Question 5

Intention de traiter ITT vs Intention du protocole ITP

Answer 5

In short, the ITT does not allowany exclusionof patients in the main analyses: a patient recruited and randomized is de facto included in the main analysis. One of the advantages of ITT is that it preserves the integrity of randomization and therefore protection against confounding bias

The ITP analysis aims to include only the “perfect” patient: that is, the one who uses the treatment as prescribed, comes to each of the visits prescribed by the protocol, and adheres to all other procedures of the ‘study. Ultimately, those are most likely to respond to treatment.

Question 6

What is the issue with an ITT analysis? IS it perfect?

Answer 6

ITT dilutes the magnitude of response by underestimating benefit and overestimating safety.

For example, participants who are not adherent to the treatment, why keep them in the main analysis and not exclude them?Since these patients have not used the treatment, they cannot benefit from it if it proves effective.

Question 7

What is the issue with ITP?

Answer 7

ITP exposes the study to confusion bias because the more exclusions there are, the further we move away from the initial randomization and the comparability of the groups

Question 8

Why is it more difficult to carry out an ITT analysis?

Answer 8

An ITT analysis does not allow any exclusion of patients who were randomized for any reason.This includes missing values.For example, a randomized patient who is absent during their clinical visit to measure their CRP.This generates a missing CRP value for this patient.In practice, this means that it is very difficult to carry out an ITT analysis in the pure sense of the term.

Question 9

Which type of analysis has a higher internal validity between Intention de traiter et Intention de protocole

Answer 9

ITT analysis is much less susceptible to bias than ITP analysis.The internal validity of the TT analysis is therefore greater than that of the ITP analysis.

Question 10

Généralement, la première table d’un article

Answer 10

présente les statistiques descriptives des variables mesurées à la visite de recrutement (statistiques socio-démographiques et cliniques)

Question 11

What is IQRis theinterquartile range

Answer 11

(IQR = Q3 - Q1);that is, half of the observations are between Q1 and Q3

Question 12

What does the line inside a box plot represent?

Answer 12

median

Question 13

The IQR and standard deviation are measures of _________ while the median and mean are measures of _______________

Answer 13

dispersion
central tendency

Question 14

The _______ is more robust than the mean because it is less sensitive to extreme values and skewed distributions.

Answer 14

median

Question 15

What is d’écart-type?

Answer 15

standard deviation

Question 16

How to get varience from d’écart-type

Answer 16

The square of the standard deviation is called the variance

Question 17

What is stadard error

Answer 17

Similar to standard deviation but for the sample means say in a repeated experiment. Here x is the mean of the essay and mean is the mean of means.

Question 18

In a contingecy table, how would you calculate le produit des proportions marginales associées au succès du traitement dans le groupe de patients avec le traitement expérimental? if Num total = 100 and sucessfult = 40

Answer 18

P(A) is probability of of being sucessful
P(B) is probability of being in experimental group
so this question is askin what is P(A and B)

P(A) = 40/100 = 0.4
P(B) = 0.5 (assuming a 50/50)
p(A&B) = 0.4*0.5 = 0.2

Question 19

Given a dichotomus variable with p =0.2 calculate varience and std

Answer 19

std = sqrt(0.2 (1-0.2))
std = sqrt(0.2*0.8)
std = 0.4
var = sqrt(std)
var = sqrt(0.4)
var = 0.632

Question 20

Ture or False;

plus il y a de pertes au suivi, plus la validité interne de l’étude est affaiblie

Answer 20

True

les pertes au suivi sont un très bon indicateur de la qualité d’une étude