P3 Epidemiology Final

dichotomous outcomes

descriptive measures

frequency of event in specific time period

allows identification of high risk groups and possible clues of causality

numerator- freq
denominator- population
time period
multiplier - / how many

relationship between numerator and denominator when they are unconnected quantities

numerator and denominator are always related

often shown as percent

rate- X / delta t

ratio- X / Y

proportion- X / Y+X

any departure from well being

any departure from actually living...aka death

incidence and prevalence

number of new cases

number of cases in population

proportion of population at risk that develops disease over specific peroid

denominator must be observed entire duration followup

CI = number new cases dx during time
______________________
Pop at risk during same time

same shit only with diff length of folow up times

Numberof new cases in specific time
IR = __________________________
total person time of observation in pop at risk...huh

number of existing cases in entire pop

number of ppl w/ dx at single point in time
_________________________
total number pop at that time

number ppl w/ disease at any point during time peroid
______________________________
number of persons in pop

a bath tub....
incidence water going in
prevelence water inside
and death or recovery is how fast water leave and prevelance goes down

total number deaths from all causes per 1000 persons during specific time
______________________
total number persons at that time

total number of deaths from all causes in specific age group
___________________________
total number in that age group

deaths from specific cause
________________________
total pop that period

propensity of disease to cause death


number deaths from dx
_____________________
# ppl w/ dx

proportion of deaths from specific cause

deaths from dx
_________________
total deaths

statistical relationship tween 2 or more variables

exposure- potential causal characteristics (behavior or tx)

outcome- consequence of behavior or tx

risk ratio

odds ratio

risk of developing event in exposed individual to that of unexposed individual

cumulative incidence exposed
________________________
cumulative incidence unexposed

-risk is equal in exposed and unexposed groups

-if greater than 1 than exposure is associated

NOTE TO REMEMBER- for calculations with risk ratio recognize that total is for each group not whole study!!!!

ratio of number events and nonevents in cases to number of events and nonevents in control....

double positives * double negative
__________________________
mixed * mixed

greater- odds for event in exposed group higher

less- event less liekly in exposed group

analytic- tests predetermined hypotheses about association tween exposure and outcome variables

framework for determining statistical significance

really how many times have we learned this now....

means no difference (for u morons out there)

theres an association

I = P = positive

II = N = negative

power matters when null is rejected truthfully

probability that effect as large as one seen in study could be from chance alone

used to determine conclusion of study

inverse relationship

sample up, P down

0.05

use appropriate stat anal method

compare calc value to table value at specified P value

IF calc value larger...data significant

relative risk reduction

absolute risk reduction

number needed to treat

number neede to harm

extent to which exposure reduces a risk in comparison to unexposed

proportion of risk in untreated group...

RRR= (Ru - Re) / Ru

or 1- RR

simple...absolute value difference event rate of exposed and unexposed groups

ARR = Re - Ru

from what i understand these Re or Ru are cumulative incidence things

D- methods and procedures for summarizing and describing data

I- methods used to make statements about the populations based on sample

specific value of variable

ex female for gender

exhaustive

mutually exclusive

Nominal

Ordinal

Interval

Ratio

Discrete vs continous
Dependant vs independant

Negative

Positive

None

Curvilinear - ex age vs intelligence

takes into account the two components involved in measurement

Observed score = true ability + random error

random- from factors that randomly effect measurement
-not consistent effects
-add variability w/o changing average
-called noise

Systematic
-affect measurement across sample
-positive or negative consistently
-called bias

R- repeatability or consistency

V- best availible approximation of the truth

External

Construct

Internal

Conclusion

can we generalize this to other persons, places, times

get a random large sample to get this

are you using the correct instruments to measure what we are attempting to understand

are we executing this correctly design wise

do we see cause and effect

is this the true cause or its it something else thats causing it

to the results and findings match the conclusion given

are we reading the data correctly and taking it into account

distribution

central tendency

dispersion

spread of values around the central tendency

range and standard deviation

will show how certain scores relate to the mean of sample

decision making information about hypotheisiszed values of an unknown parameter

systematic error avoided since it can be eliminated

random error just kinda happens

F they do not...

interval estimation associating a measure statistical variation with a point estimate

has two numerical values defining range of values covering parameter being estimated

range of values around point estimate tween upper and lower limits

inversely

smaller random error more precise point estimate

do it

regression modelling!

linear regression

regression coefficient

logistical

odds ratio

cox regression

hazard ratio....
(thats not what i heard in drug lit....)

when other IV effect our results

Confounding
Mediation
Moderation

changes relationship related to both

different interpretations when ignored or included

accounts for all or part of relationship

IV causes interveneing variable in turn causes DV

differs conceptually from mediation- confounder is not intermediate in causal sequence

alters strength or direction of relationship

relationship is diff at diff lvls of moderator

therapy appropriate

medication used as recommended

the point below which desired effect is unlikely to be achieved

extent behavoir coincides medical advice

patients failure consume medication according to directions

agreement tween pt and HCP regarding ultilization behavior

defines medical compliance conformity to recommendation dosing, timing, freq

length of time taking medication in time...

initial noncompliance

partial complinace

compliance

hypercompliance

direct measures- directly observed therapy

indirect measures- self report, provider esimates

contemporary emphasis on rights of ppl take risks to save themselves

dont care