Critical numbers 2- Summary Statistics I Flashcards
3 types of catergorical data:
- binary- 1 or the other
- ordinal- natural order e.g. level of pain
- nominal- no natural order e.g. blood group
2 types of numerical data
- discrete - can only take certain values e.g. shoe size
- continuos - can take any value e.g. weight
prevelance vs incidence
prevelance: number of existing cases
incidence: number of new cases
what is risk?
probability of an outcome
what is odds?
probailty of outcome occuring vs outcome not occuring (divide outcome occuring by outcome not occuring)
what is rate?
probaility of an outcome per time (add total time even if lost to follow up) per person per year
relative risk/odds/rate
1 divided by the other
- 1 means no difference
- higher than 1, higher risk/odds/rate
- lower than 1, lower risk/odds/rate
relative vs absolute
realaitive is percentage incraese but absolute is actual figures
What does PICO / PECO stand for?
population
intervention/exposure
comparator
outcome
establsihing causality in non-randomised studies?
strength of association- stronger association more liekly to be causal
consistency- shown over differnet studies? differnet location?
specificity- specific exposure outcome relationship
temporality- exposure must precede outcome
biological gradient- dose-response- higher exposure = higher outcome?
plausibility- biological mechanisms explain this?
coherence- compatibilty with current theories
experiment- outcome altered with experimentation
analogy- similar cause-effect relationships esablished
What are ecological, cross sectional, case-control, cohort studies?
Ecological studies – compare areas with different levels of exposure
Cross-sectional studies – compare individuals with different levels of exposure
Case-control studies – compare those with or without the outcome/disease
Cohort studies – follow up individuals with different levels of exposure over time
what is a confounding variable?
another factor linked to both risk and outcome
