Samenvatting ARM alle colleges + werkgroepen uitwerkingen (in detail)

ADVANCED RESEARCH METHODS

ONDERWERPEN PAGINA
Week 1 CAUSAL INFERENCE, DAGs DAGs 2
Causal inference 4
Positivity, consistency, exchangeability 6
Exchangeability (dieper kijkje) 7
Colliders 13
Confounders 14

Week 2 & 3 PROPORTIONS & OLS (coefficient) 22
PROBABILITIES Mediation 27
Statistical significance 30
Beyond P <0.05 33
Logistic Regression 44
OLS vs. Logistic Regression 48
OR, RR, RD 51
Fallacy 57

Week 4 INTRO QUALITATIVE Qualitative methods 69
RESEARCH METHODS & DISCOURSE Discourse analysis 70
ANALYSIS Do’s and Don’ts 71

Week 5 ETHNOGRAPHY Ethnography 82
Theory & ethnography 84
Subtle realism & relativism 87
Deductief, infuctief, abductief 88
Nine observational dimensions (Reeves) 89

Week 6 QUALITATIVE & Why quantitative in Healthcare? 96
QUANTITATIVE METHODS Why qualitative in Healthcare? 101




1

, Knowledge video 1: Directed acyclic graphs
DAG theory
You will be able to use DAGs to examine the effect of one variable on the other without the disruptive influence of
other variables.

Directed Acyclic Graphs (DAGs) are graphical representations of the causal structure underlying a research question:




▪ DAGs help to visualize the causal structure underlying a research question
▪ You need a priori theoretical/subject knowledge about the causal structure to draw a DAG (e.g., from previous
studies, literature, common sense, …)
▪ Collect data on all relevant variables
▪ ‘Simple’ rules can be applied to determine for which variables to adjust in regression analysis and how to interpret
the results

DAG terminology

1 PATHS 2 CAUSAL PATHS AND BACKDOOR PATHS
RQ: Influence of X on Y? RQ: Influence of X on Y?




▪ A path is any route between exposure X and ▪ A causal path follows the direction of the arrows
outcome Y ▪ A backdoor path does not → arrows can go into
▪ Paths do not have to follow the direction of the different directions.
arrows. The arrows of the path can also go into
different directions. Which are causal or backdoor paths? →
How many paths between X and Y? Causal:
In this DAG there are 4 paths. X→Y
X→V→Y
1) X→Y
2) X→V→Y Backdoor:
3) XL→Y XL→Y
4) X→WY X→WY




2

,3 OPEN AND CLOSED PATHS, COLLIDERS 4 BLOCKING OPEN PATHS
RQ: Influence of X on Y? RQ: Influence of X on Y?




▪ All paths are open, unless they collide somewhere in ▪ Open (causal or backdoor) paths transmit
a variable on a path association
▪ A path is closed if arrows collide in one variable on ▪ The association between X and Y consists of the
that path combination of all open paths between them
▪ Here: all paths except X → W  Y
How many paths are open and closed?




3 are open and 1 is closed. ▪ An open path is blocked when we adjust for a
The yellow one is the closed path, because two arrows variable (L) along the path
collide on a variable (W). W is hereby the collider. ▪ This means that we remove the disruptive influence
of L from the association between X and Y
▪ How? By including variable L in the regression
analysis
▪ Backdoor paths always need to be closed
▪ Causal paths need to be open/closed depending on
RQ

5 OPENING BLOCKED PATH
RQ: Influence of X on Y?




▪ Including a collider (W) in the analysis means you open the blocked backdoor path. If you close the path you
can remove the collider.
▪ This introduces bias in the association between X and Y




Lecture 1: Causal inference, directed acyclic
graphs

3

, 1 INTRODUCTIONS TO CAUSAL INFERENCEhhhhhhhhhhhhhhhhvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv

LEARNING GOALS
▪ Explain the potential outcomes approach in causal inference
▪ Apply the approach in critical thinking about causal effect estimation
▪ Define ‘causal effect’
▪ Apply the concepts of consistency, positivity, and exchangeability in randomised and observational trials


Proven clinical results: 70% less imperfections
in 4 weeks. True Match Minerals foundations
tested under dermatological control with 41
women. Average reduction of the most visible
imperfections linked to oily skin.

→ they say that this is a powder you should definitely buy. But does it really lead to less imperfections on your skin.
Does A lead to B (causal effect)? Would you buy this powder? They tested it on 41 persons, that is not a lot. It is a
worldwide number 1 powder, but they tested it on only 41 women. It is also not really clear to what they compare
their product.
• Small sample size for this type of research sample is too low. Maybe for another study 41 women would have
been enough. But not here.
• Study performed of financed by commercial company (L’Oréal). This isn’t always a problem
• No control group which means that essential data is missing. And what would have happened when these
women wouldn’t have used this powder. Potential regression to the mean, because if you wait 4 weeks, most
imperfections will heal.

What do we want to know?
In causal inference:
▪ We are not interested in the outcome per se (i.e., 70% less imperfections), but …
▪ We are interested in the role of the treatment in achieving this outcome (i.e., without True Match Minerals
powder, would there have been less skin imperfections?)

Conclusion:
→ We do not have that information
→ No causal claim can be made based on L’Oréal study

Causal effect
Formal definition by Hernàn and Robins (2020):

‘In an individual, a treatment has a causal effect if the outcome under treatment 1 would be different from the
outcome under treatment 2.’
→ if the only difference between the people is if they have had the treatment or not, then you can say that
you have a causal effect. You need to have this information.

To assess this, we need information on:
→ What would have happened?
→ What will happen?

Assume that we know what would have happened in the L’Oréal study:
▪ Woman A treated with True Match Minerals powder: 2 bad spots
▪ Had woman A not been treated with True Match Minerals powder: 5 bad spots

→ Individual treatment effect: -3 spots (or 60% less imperfections)
→ Average treatment effect: average of individual effects in a population



4