Biostats Flashcards
observational studies
cross-sectional - frequency of disease and risk factors, shows association (not causality)
case-control - OR (pts with COPD had higher odds of a history of smoking than those without COPD)
odds ratio = [disease/no disease in exposure group]/ [disease/no exposure with no exposure]
cohort - RR
relative risk = %disease with risk factor/% disease without risk factor
clinical trial
does the drug SWIM
1) healthy volunteers - safe?, assess drug properties
2) small no of pts with disease - work?
3) RCT - improvement?
4) postmarketing surveillance of pts after treatment is approved - stay on the market?
PPV and NPV
depend on disease prevalence
PPV = true pos/all pos NPV = true neg/all neg
sensitivity
specificity
sensitivity:
100% sensitivity rules out disease, low false-negative rate
= 1-false negative rate
cut-off values:
LOWERING cutoff point - increased sensitivity, decreased specificity, increased NPV, decreased PPV (more positive tests)
likelihood ratio (?)
positive LR = TP rate/FP rate
pos LR > 10
neg LR < 0.1
risk
attributable risk - difference in risk between exposed and unexposed groups
= (risk in exposed - risk in unexposed)/risk in exposed
absolute and relative risk reduction
NNT = 1/ARR
NNH = 1/AR (of disease with expsoure)
prevalence and incidence
prevalence/(1-prevalence) = incidence rate x avg duration of disease
precision
aka reliability
increased precision –> lower SD, increased statistical power (1-b)
kids in cars
kids < 12 should not ride in a set with a front-facing airbag
medicare
65+
<65 with certain disabilities and ESRD
suicidal pt
if serious - suggest pt stay in hospital voluntarily
- pt can be hospitalized involuntarily if he/she refuses
pt is angry about the delay
apologize
stay away from efforts to explain the delay
kids
infant (0-12 mo) = Parents Start Observing
- primitive reflexes disappear
- posture
- picks and pincer grasp, points
- social smile, stranger anxiety, separation anxiety
- orients to voice/name/gestures
- object permanence
- mama, dada by 10 mo
toddler (1-3) - Child Rearing Working
- cruises, climbs stairs, cutlery, kicks ball
- recreation (parallel play), rapproachment (to and from mom), realization of gender (by 36 mo)
- words - 200 words by age 2, 2-word sentences
preschool (3-5) - Dont Forget theyre still Learning
- drives tricycle, drawings, dexterity
- freedom, friends (cooperative play)
- language - 1000 words by age 3
- legends - story telling
ages 5-7 - begin to understand that death is permanent
consent for minors
minor < 18
parental consent is not required for sex (contraception, pregnancy), drugs (substance abuse), emergency/trauma
abortion:
many states require parental notification or consent for minors for an abortion
biases
confounding bias - cross-over studies, etc. to reduce bias
latent period
- example - antioxidants reduce risk with 5 years…
standard error
= SD/square root of sample size
estimate of variability of sample means around the TRUE population mean
distributions
positive skew = data concentrated on left
mean > median > mode
incorrect results
type 1 error - stating there is an effect/difference when none exists
a is probability of making a type 1 eorr
type 2
b is the probability of making a type 2 error
power = 1-b, power is the probability of rejecting the null hypothesis when it is false
- increase power by increasing sample size, increasing expected effect size, increasing precision
confidence interval
95% CI has a 0.95 probability of containing the true population mean
95% CI = mean + 1.96SE
99% CI = mean + 2.58SE
Chi-squared
for categorical outcomes
- can be for 2 or more groups
also used to compare expected frequency of occurrence with observed frequency of occurence
competence v.s. capacity
capacity determine by physician, competency determined by judge
reportable disease
STIs, hepatitis, food poisoning reported to public officials –> public officials will notify people at risk
Tb, Ebola have mandatory treatment
causes of death
less than 1 year - congenital, premie, maternal pregnancy complication
1-14yr - accident, cancer, congenital
15-34 - accident, suicide, homicide
35-44 - accident, cancer, heart disease
45-65 - cancer, heart disease, accident
65+ - heart disease, cancer, chronic respiratory disease
PDSA cycle
process improvement model to test change
plan, do, study, act
errors
active - wrong IV pump dose entered, immediate error
latent - error waiting to happen, different types of IV pumps used in one hospital
error analysis
- RCA - fishbone
- failure mode and effects analysis - forward-looking approach
Berkson’s bias
selection bias when hospitalized patients are the control group
HIPAA
consent to share medical info- including with family and friends
when pt is incapacitated - basic info can be shared, in pts best interest
