and behavior change (lesson2) - example: the vitamin E case
Paper from 1993
Conclusion: Association of higher vitamin E intake, lower
coronary heart disease in men.
- Not causal, but observational study.
- Enough to start a prevention campaign?
After 1993: peak vit. E intake!
- Higher correlation: vit. E intake & health, stronger neg. correlation
with mortality.
=> associated: education, income, good diet and exercise, neg.,
correlation with smoking.
Meta-analysis 2005: high vit. E dosage even associated with a higher all-cause mortality risk, so in high
quantities it might even be detrimental!
CONCLUSION:
- Health recommendations are more likely to be adopted by individuals who invest more in their
health in other ways;
• Here: not necessarily vitamin E that is responsible for lower risks, but rather third variables
that co-occurred due to common determinants (cf. the educated and higher income people
etc. were the ones in that period taking vitamin E supplements). – multifactorial analysis.
- Health behaviors are correlated (due to common determinants);
- Observational findings can be misleading (confounds).
• Confounds!