Journal Club (méthodologie) Introduction à la lecture d’articles médicaux (Partie 2)
Introduction à la lecture d’articles médicaux (2)
Dr. Ioannis KOKKINAKIS
Chef de Clinique
Service de Médecine Interne - Hôpital du Valais – Sion
20.06.2017
| Page 2 |
Qu’est-ce qu’on a vu la dernière fois?
•
1. Grille de Lecture pour ERC
•
2. Autres Grilles des Lecture
•
3. Recommandations Int.
•
Lien pour télécharger les présentations:
https://www.dropbox.com/sh/m7da6csj6idm7xi/AAAKCvNlMtX2xsGS9xjmBqcHa?dl=0
| Page 3 |
Qu’est-ce qu’on va voir …
• 1. Rappel de la grille de lecture
• 2. Types d’études
• 3. Etudes observationnels
• 4. Etudes expérimentales
• 5. Métanalyses
| Page 4 |
Rev Mal Resp 2002, 19, 505-514
| Page 5 |
Rev Mal Resp 2002, 19, 505-514
| Page 6 |
Notre Grille de Lecture (SSMI)
| Page 7 |
Autres Grilles de Lecture
| Page 8 |
Reference: http://libguides.csuchico.edu/c.php?g=414398&p=2822479
| Page 9 |
Types d’études
Experimental study :
•
the investigators directly manipulate or
assign participants to different interventions or
environments.
•
Experimental studies that involve humans are called
clinical trials.
•
Controlled trials / Uncontrolled trials
•
Randomised / Non randomised
Observational study :
•
the investigator cannot control the assignment of
treatment to subjects because the participants or
conditions are not being directly assigned by the
researcher.
•
Examines predetermined treatments, interventions,
policies, and their effects
Trapp R.G. (2004). Chapter 2. Study Designs in Medical Research.
| Page 10 |
Reference: http://libguides.csuchico.edu/c.php?g=414398&p=2822479
| Page 11 |
Référence: http://cisncancer.org/research/
| Page 12 |
Types d’études
| Page 13 |
Prevalence
| Page 14 |
| Page 15 |
Incidence
| Page 16 |
| Page 17 |
| Page 18 |
| Page 19 |
Ecologic studies
•
Outcome + exposure measured at a group (or population) level
•
Using data already available, such as routine surveys.
•
Relationship between the outcome and exposure measures at the group level is assessed.
Proportion of the adult population
Proportion of children with Asthma.
who smoke
in 56 countries
-
OMS
International
Allergy in Childhood
Study of Asthma and
.
Mitchell EA &Stewart AW, Eur J Epidemiol 2001, 17: 667-673.
| Page 20 |
Image of American Institute of Cancer Research - http://cisncancer.org/research/
| Page 21 |
Choisir le type d’étude…
•
•
Fumeurs
Non fumeurs
Cohor
t
prospective
•
•
BPCO
Pas de BPCO
•
Fumeurs
•
Sarcoïdose
•
Non fumeurs
Case control
•
Group contrôle
•
Prévalence et association
Cross sectional
Asthme – Tabac dans une population
| Page 22 |
How to chose the study design.
| Page 23 |
•
•
If an association between an exposure and outcome
has been observed in a study, this can be either due to
a genuine underlying effect in the reference population,
or as a result of some kind of “error” and the apparent
effect is not genuine.
Therefore, before concluding that any difference in the
outcome between the exposure groups is due to the
effect of the exposure, it is important to discount the
possibility of other explanations.
| Page 24 |
Risks
to consider
….
•
•
•
•
Chance or random error
Confounding
Bias
Reverse causation
| Page 25 |
•
Chance or random error
•
Confounding
•
Bias
•
Reverse causation
Random error occurs due to chance alone, determining the difference in the
results between the study sample and those which would have been observed if
the whole target population was studied - when the study sample is taken from the
target population, the results of the study are unlikely to be identical to that of the
target population since the selection of participants would be prone to random error
(or random variation).
| Page 26 |
Confounding is where the distribution of risk
factors for the outcome (other than the
exposure of interest), between the cases and
controls is unequal.
A confounder is a factor that can partly or
wholly explain the apparent association
between the outcome and the exposure of
interest; or obscure a real association between
the outcome and the exposure of interest.
To be a confounder, a factor must be:
-
associated with the exposure of interest
-
independently associated with the outcome
of interest
-
not be on the casual pathway between the
exposure and the outcome
•
•
•
•
Chance or random error
Confounding
Bias
Reverse causation
| Page 27 |
Confounding
Smoking
Alcohol
Cardiovascular disease
-associated with the exposure of interest
-independently associated with the
outcome of interest
-not be on the casual pathway between the
exposure and the outcome
| Page 28 |
•
•
•
•
Chance or random error
Confounding
Bias
Reverse causation
| Page 29 |
Selection bias
•
Selection bias can occur if the exposed and unexposed groups are not
comparable in terms of factors related to the outcome, apart from the exposure
of interest.
•
However, this tends to be more of a problem in studies where participants are
selected on the basis of their exposure status, for example, when an external
comparison group is used.
•
Completeness of follow-up
•
Case ascertainment differs between the exposure groups
| Page 30 |
•
•
Observer bias
Occurs when the degree of accuracy with which the study investigator records information
on the outcome depends on the exposure status of study participants, and vice versa.
It is more likely to occur when there is a greater element of subjective judgment required for
classifying the outcome or exposure.
| Page 31 |
Reporter bias
Occurs when the degree of accuracy with which study participants report information on
the outcome depends on their exposure status, and vice versa.
Image of American Institute of Cancer Research - http://cisncancer.org/research/
| Page 32 |
Recall bias
Image of American Institute of Cancer Research - http://cisncancer.org/research/
| Page 33 |
Reverse causality is where the exposure might be
the result rather than the cause of the outcome.
This is more of an issue in a CCS, where
measurement of the exposure is done after the
outcome has been diagnosed – so it can be
difficult to establish the time sequence of
exposure and outcome.
Cross Sectional Study
Mental Status
Dementia
•
•
•
•
Chance or random error
Confounding
Bias
Reverse causation
Opoid Use
| Page 34 |
| Page 35 |
•
•
•
•
Factorial design trials
Non-inferiority and equivalence trials
Cluster randomised trials
Crossover trials
| Page 36 |
•
•
•
•
•
•
•
•
Alternative Clinical Trial Designs
Factorial design trials
Non-inferiority and equivalence trials
Cluster randomised trials
Crossover trials
Factorial design trials
In a factorial design trial two (or more) experimental treatments are assessed within a
single trial. Suppose the experimental treatments under investigation are treatments A and
B. Essentially each randomised patient is randomised twice as follows:
Treatment A vs the control of treatment A
and
Treatment B vs the control of treatment B
1. A + B
2. A + control for B
3. control for A + B
4. control for A + control for B
| Page 37 |
•
•
•
•
•
•
•
•
•
Alternative Clinical Trial Designs
Factorial design trials
Non-inferiority and equivalence trials
Cluster randomised trials
Crossover trials
Non-inferiority and equivalence trials
Most trials are conducted to assess the superiority of some treatment over another or in
comparison to a placebo. However, there are circumstances where the reason for
conducting a trial is to establish that two treatments are the same in respect of patient
response. The aim of an equivalence trial is for the observed treatment difference to be
close to zero with a tight confidence interval.
In fact true equivalence trials are relatively rare. Most aim to establish the non-inferiority of
one treatment over another i.e. to establish that one treatment is at least as good as the
alternative in respect of patient response and are more correctly referred to as non-
inferiority trials.
Establishing equivalence can have important bearings on the subsequent use of the
treatment.
For example, the treatment under investigation may have fewer side-effects, may be easier
to administer, may be cheaper or may be useful as an alternative treatment if response is
poor on the established therapy.
| Page 38 |
•
•
•
•
Alternative Clinical
Factorial design trials
Non-inferiority and equivalence trials
Cluster randomised trials
Crossover trials
Trial Designs
| Page 39 |
•
•
•
•
Alternative Clinical
Factorial design trials
Non-inferiority and equivalence trials
Cluster randomised trials
Crossover trials
Trial Designs
| Page 40 |
| Page 41 |
Metanalysis
| Page 42 |
| Page 43 |
Forest plot
| Page 44 |
Further reading…
•
http://www.nejm.org/page/clinical-trials-series
| Page 45 |
© LE PEILLET Damien, Olaf et parentés
Design original par BootstrapMade
Design original par BootstrapMade