DIT Medicine in Society Flashcards
Case control study? What is the outcome of it?
compares group of people with disease to a group without it. look for prior exposure.
RETROSPECTIVE
OBSERVATIONAL
Outcome is odds ratio.
Cohort studies? What is outcome? What is outcome?
A cohort is a group of people with similar characteristics.
You follow people with exposures or characteristics and see how ti goes.
OBSERVAITONAL
Prospective or Retrospective
Relative Risk is outcome. How many times greater risk of bad outcome is for those exposed
Cross sectional study? What is outcome?
Data of people at that point of time.
Strictly observational
Prevalence. no causality can happen.
Twin study vs adoption study?
Twin compares monozytotic to dizygotic.
Adoption is to compare heritability vs environmental
Clinical trial phases?
- Is drug safe?
- does drug work?
- Does drug work as well or better? (often in medical journal)
- Can it stay? (post marketing surveillance)
What is meta analysis?
Combining studies to strengthen statistical significance. Pool the studies to gather to get inclusion criteria.
Make the two by 2 table for diagnostic statistics.
Diseasse?
\+ -
Test + TP FP
- FN TN
"Disease always comes to on top"Sensitivity?
Don’t miss anything. Sensitive. If you have it, you test positive.
TP/(TP+FN)
Specificity?
If you test negative, you don’t have it. Didn’t give a false positive. OPPOSITE OF sensitifvity
TN/(TN+FP)
Vertical calculation in the table (as is sensitivity)
Positive predictive value?
Proportion that a tested positive is true positive. So divide out the false positives+true positive
TP/(TP+FP)
negative predictive value?
Means you don’t have false positives.
TN/(TN+FN)
Horizontal tests
How does a higher prevalence impact positive predictive value?
What about negative predictive value?
higher prevalence increases positive predictive value, lowers the negative
(makes sense, think about if nobody has a disease, do you believe it?)
What is incidence X average disease duration?
It is about the prevalence
Odds ratio?
the name. RATIO OF ODDS.
odds of people with disease was exposed over risk VS odds that the group without the disease was exposed.
(A/C)/(B/D) or (A/B)/(C/D)
Relative risk?
C/(C+D)
If low prevalence, the two are about equal
Attributable risk?
AR is incidence of disease in exposed - incidence of unexposed
Easy. 30% in exposed - 5% unexposed means attributable risk is 25%
Absolute risk reduction?
Easy. It is seeing if intervention decreases incidence by how much
Vaccinated 5% risk and unvaccinated is 25%? then ARR is 20%
Number needed to treat?
1/Absolute risk reduction
Number needed to harm?
1/Attributable risk
What is Berkson bias?
a type of selection bias of only looking at inpatients
What is a crossover study?
Patients are their own controls. For a period of time each patient gets either placebo or treatment
Type I error vs type II?
Type I is alpha, and alpha is the chance f there being alpha error (so false positive)
Type II is the other other, so you accept the null hypothesis when there was a difference you didn’t see. False negative
What is the alpha value in statistics?
Alpha is the level set by investigator in which P value is judged. Usually 0.05
Reject null if below alpha
What is power?
a
Standard Error of Mean?
Standard deviation/square root n
Confidence Interval?
Mean +/- Z SEM
Z:
1.6 (90%)
2 (95%)
2.6 (99%)
what is correlation coefficient interpreted?
R of 1 is perfect correlation
R of zero is no relationship
R<0 is negative correlation
three types of disease prevention?
Primary: prevent
Secondary: Screen
Tertiary: Treat
PST and PST
Kubler Ross stages of grief?
DABGA
Denial Anger Bargaining Grieving Acceptance
