BIostat Flashcards
what is an example of continuous data
logical order with values that increase or decrease by the same amount - HR 120 is x2 of HR 60
interval data- no meaningful 0 ( 0 equals something) 0 degree C
ratio data- meaningful 0( 0 means 0) HR is 0
what is an example of discrete data or categorical data
nominal- male female, yes no
ordinal- order = 1,2,3,4,5 pain scales, NYGH i-IV
what is the SD of data
it shows how spread out the value is from the mean, the lwoere the more closer to the mean
what are some charecteristics of normal or gaussian distribution
mean median and mode are the same and are at the center
68% of the values fall within 1 SD and 95 % of the values fall within 2 SD
when is better to use a median and how to even better the measurement
when the data is skewed adn to better add more participants
skewed meaning
data is in the direction of the outlier ( right meaning skewed to the right higher values)
what are independant variables
changed manipulated by the researcher for example drugs, dose, placebos, patients , age etc
what are dependant variables
effected by independant variables
- HF progression, HBA1c, BP cholesterol values
what is the alpha level
alpha level is the max permissable error margin commonly sett at 5%
what is a confidence interval
1- alpha
most of the time 95% confident data is true
if value touches 0 data is not statistically significant
RRR, OR, HR
result is statistically significant if data doesnt cross 1
what is a type I error
error where null is rejected but actually it should be accepted
false positive
** alpha <5 means 95 % correct and 5 % chance for TI error
what is a type II error
where null is rejected shouldve accepted
false negative
Power is used to avoid a type II error
P= 1-b
Power 90% means 90% of avoiding type 2 error
treatment relative risk meaning
if > 1 the treatment is risk with the endpoints
<1 treatment is safer wtih the endpoints than control group
“as likely” as the control group
what is the Relative risk reduction
% of patient less likely to have a disease after treated
“less likely” than control
What is the absolute risk reduction
includes reduction in risk and the incidence rate of the outcome
X amount of pt will benefit from the treatment
can be used to find NNT and NNH
Number needed to treat
1/ ARR round up
example - for every 9 pt who recieve the treatment per year 1 pt can prevent the disease
Number needed to harm
1/ARR rounded down
example- for every 90 pt treated with the treatment 1 patient can get harmed
what is an odds ratio
an event will occur verses the probability that it will not occur
used mostly in case control studies
what and when is hazard ratio used
when there is a disease progression or death
similar to RR
when is a T test used
a student t test is when there is continuous data and 2 variable a treatment and a control group
when is ANOVA or F test used
when there is continuous data with 3 or more samples
Chi square is used when
there is discrete/ categorical data ( such as pain scale) between treatment and control
sensitivity determines
the true positive, higher the sensitivity the better reults
specificity determines
the true negative
the more specific it is the more able it is to differentiate
