Observational, Retrospective

Compares a group with a disease to a group without

Looks for prior risk

Measured by Odds Ratio

ie. Patients with COPD had higher odds of history of smoking than those without COPD

Observational, prospective

Compare a group exposed to risk to group not exposed

Asks, does exposure increase disease likelihood

Measure by Relative Risk

ie. Smokers had higher risk of developing COPD than did nonsmokers

Observational

Assesses frequency of diseas and associated risk factors

does not establish causality

Measurment for Case-control

odds that a risk factor is associated with a disease

Groups are: diseased and non-diseased

a/b

----

c/d

Study of humans

highest quality when randomized, controlled, and double-blinded

Phase I: is it safe

Phase II: is it effective

Phase III: is it more effective than current txt

Phase IV:  are there long term effects

aka reliability

Consistency, reproducability

affected by Random error

aka validity, trueness

affected by Systematic error

difference in risk between exposed and unexposed

proportion of disease occurrences attribulable to risk

(a/(a+b)) - (c/(c+d))

Reduction in risk associated with treatment as compared to placebo

Same equation as Attributable risk with different variables

knowledge of presence of disorder alters recall by subjects

subjects not representative of general population

results not generalizable

information gathered at inappropriate time

affected by fatal disease, deceased cannot respond

2 closely associated variables distort or confuse each other

control by matching variables, ie. if studying poor population, control should also be poor.

Early detection confused with increased survival

Seen in improved screening

Natural history of disease is unchanged

What is the least affected statistical parameter of statistical distribution by outliers

alpha

you saw a difference that did not exist

convicting an innocent man

False positive

p<.05= <5% chance of alpha error

p= probability of type one error

Beta

Stating there is no difference when there is

Setting guilty man free

fail to reject null when it is false

(1-beta)

Probability of rejecting null hypothesis when it is false

(ie. probability of getting it right, accepting alternative hypothesis when it is right)

increases with sample size

1: 68%

1.65: 90%

2: 95%

3: 99.5%

Range of values in which a specified probability of the means would be expected

If the CI for a mean difference includes zero, then there is no significant difference, Ho is not rejected

checks the difference between the means of two groups

ie. mean cholesterol between smokers and non-smokers

r

between -1 and 1

closer to 1 means stronger correlation

1: Prevent occurance (vaccination)

2: early detection

3: reduce disability from illness

Rooting reflex

Orients to voice

Holds head up

Moro reflex dissapears

Social Smile

Sits alone

Crawls

Stranger anxiety

Walks

Babinski reflex dissapears

Few words, separation anxiety

Climbs stairs

Stacks 6 blocks

Object permanence

200 words 2 word sentances