Biostats Flashcards
sensitivity
TP / TP+FN
A / A+C
specificity
TN / TN+FP
D / B+D
positive predictive value
TP / TP+FP
A / A+B
negative predictive value
TN / TN+FN
D / C+D
what is the relationship between NPV and pretest probability?
inverse
case-control compares:
risk factor frequency
patients selected based on if they have the disease or not
looks backwards to see if they had past exposure to a risk factor
odds ratio
used in case-control
(A/C) / (B/D) = (case;exposed) / (control;not exposed)
(A x D) / (B x C)
cohort study compares:
disease incidence
both groups start with the disease; patients selected based on exposure
looks forward or backward at various factors over time
relative risk
used in cohort
(A / A+B) / (C / C+D)
(risk;exposed) / (risk;not exposed)
cross-sectional study compares:
disease prevalence
prevalence
disease burden in a population at any given time
incidence
number of new cases in a population in a given time
power
1 - beta (FN)
power is related to sample size
false negative
type II error
beta error
false positive
type I error
alpha error
standard deviations
1 SD = 68%
2 SD = 95%
3 SD = 99.7%
normal distribution
mean = median = mode
postively (right) skewed distribution
mean > median > mode
negatively (left) skewed distribution
mean < median < mode
case-fatality rate
fatal cases / total case occurences
risk
rate of an adverse event
affected / total
absolute risk reduction
control risk — treatment risk
(affected/total;control) — (affected/total;treatment)
relative risk reduction
ARR / control rate
[control risk — treatment risk] / control risk
[(affected/total;control) — (affected/total;treatment)] / (affected/total;control)
absolute risk increase
treatment risk — control risk
(affected/total;treatment) — (affected/total;control)
number needed to harm
1 / absolute risk increase
1 / [treatment risk — control risk]
1 / [(affected/total;treatment) — (affected/total;control)]
number needed to treat
1 / absolute risk reduction
1 / [control risk — treatment risk]
1 / [(affected/total;control) — (affected/total;treatment)]
the samller the NNT, the better
positive likelihood ratio
sensitivity / (1 - specificity)
[TP / (TP + FN)] / [1 — (TN / (TN + FP)]
negative liklihood ratio
(1 - sensitivity) / specificity
[1 — (TP / (TP + FN)] / (TN / (TN + FP)
confidence interval
the null is 1.0
if it includes the null (1.0) it is not statistically significant
attributable risk perentage
percentage of disease in an exposed group that can be attributed to the exposure
[(risk;treatment) — (risk;control)] / (risk;treatment)
(relative risk — 1) / relative risk
IV qualitative X DV qualitative
Chi square
IV quantitative X DV qualitative
logisitc regression
IV quantitative X DV quantitative
linear regression
IV qualitative x DV quantitative
t-test, ANOVA