generated from statOmics/Rmd-website
-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathcovidVegetableConsumption.Rmd
88 lines (54 loc) · 3.12 KB
/
covidVegetableConsumption.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
---
title: "Covid-19 and vegetable consumption"
author: "Lieven Clement"
date: "statOmics, Ghent University (https://statomics.github.io)"
---
https://www.medrxiv.org/content/10.1101/2020.07.17.20155846v1
---
{width=70%}
---
{width=70%}
"The authors state: The negative ecological association between COVID-19 mortality and the consumption of cabbage and cucumber supports the a priori hypothesis previously reported.
In this hypothesis, we proposed that vegetables such as Brassica - with an antioxidant activity reducing insulin resistance - may also be associated with low COVID-19 mortality in countries.
"
"
Though our results do not allow to infer causality, they do reinforce our a priory hypothesis that the ingestion of anti-oxidant foods acting on insulin intolerance may have reduced the severity of COVID-19.
"
---
---
- Many hypotheses are assessed ?!
- Causality ?!
- Experimental design: Observational study
- Based on the data we cannot provide recommendations at the subject level
---
{width=70%}
- Importance of Data Exploration!
- Data does not exhibit the trend
- Data shows evidence for two clusters: above and below 200 deaths/million
- Model for cucumber does not model the data correctly: Overestimation of death rate for many countries e.g. Portugal, Hungary, ...
---
```{r echo=FALSE,out.width="10%",out.extra='style="float:right; padding:10px"'}
knitr::include_graphics("./figures/fear.png")
```
Papers are merged, published, and 39 times cited
Bousquet et al. (2021). Cabbage and fermented vegetables: From death rate heterogeneity in countries to candidates for mitigation strategies of severe COVID-19. Allergy 76:735–750.
---
In the example we covered issues with three important branches in statistics
1. Statistical inference
- Is there an association between COVID-19 mortality and food consumption?
- Issue in study:
- data dredging, p-hacking, ...: when you assess many hypotheses you will allways find strong patterns by random change $\rightarrow$ correct for multiple testing!
- Assumptions of the models do not hold
- Confounding
2. Experimental design
- Confounding: countries do not only differ in consumption of a vegetable but also in may other variables (demographical, COVID measures, healty care, ...) that are associated with COVID mortality.
- Difficult to draw causal conclusions from observational studies.
- In experimental studies: randomisation! $\rightarrow$ so that the groups only differ in the treatment.
- Experimental studies are therefore the golden standard
3. Data exploration and visualisation
- Crucial to get insight in the data!
- Assess model assumptions
---