Biostats Flashcards
mode
most frequent values
median
center value arranged from lowest to highest
Range
highest - lowest
Gaussian or normal distribution curvw
the mean, median, and mode are the same values
Skew
refers to the direction of the tail.
left skew (negative): more higher values
Right Skew (Positive): More lower values
Null Hypothesis
No statistically significant difference in the groups. what the person tries to reject
Alternative hypothesis
There IS a statistically significant difference
what the person tries to prove
Confidence Interval
CI = 1 - alpha
if it does not cross 0 or 1 it is significant
Type 1 error
False positive = Alpha
Type 2 error
False Negative = Beta
Power
Power = 1 - Beta
Is determined by outcome values, rate and significance
Risk
of unfavorable events in group / Total # in group
Relative Risk (RR)
Risk in Txt group / Risk in Control Group
RR = 1 no difference in risk
RR > 1 greater risk of the outcomes
RR < 1 lower (reduced) risk
Relative Risk Reduction (RRR)
1 - RR
Measures how much less likely the events risk is in the treatment group relative to the control group
Absolute Risk Reduction (ARR)
(% risk in control) - (% risk in treatment group)
Number needed to Treat (NNT)
1 / (ARR = % risk in control - % risk in txt)
Always round up no mater what
Odds Ratio (OR)
Estimates the risk associated with the treatment or intervention in a case control study
Hazard Ration (HR)
Hazard rate in txt group / HR in cont group
Discrete Data (Categories)
Nominal (Categorical) Data - ordeer doesnt matter
EX: Sex, ethnicity, marital statis, religous preference
Discrete Data (Rank in ORDER)
Ordinal Data - order matters
Ranked categories Ex: NYHA FC Class ranking, Pain scale 8 is worse than 7
Continous Data (most measurements)
Ratio Data - meaningful values with a true meaningful 0. 0 = none
EX: Blood pressure ( 0 = death), time, wt, ht
Continuous Data
Interval Data - meaningful values but without a meaningful 0
Ex: Temp
Student t-test
Assess significance in studies with continuous data
Chi-Square Test
Assess significance between treatment groups in a discrete data (nominal or ordinance)