Stats Flashcards
What is a type a error?
Rejecting the null hypothesis when it is in fact true
How do we calculate the power of a study?
1- B
B is chance of incorrectly accepting the null hypothesis when it is not true
Best way to decrease the risk of a B-error?
Increase sample size
Best study design to test the efficacy of an intervention?
Double blinded randomized controlled trial
Phase I-IV of drug testing:
I; tests safety of a drug in a small number of participants
II; studies effectiveness and side effects of drug in <100 participants
III: compares new drug to existing treatments in 100s of patients
IV: post market long term effects of drug
Sensitivity:
Percentage of people that have a disease and are identified as having it
TP/TP + FN
Specificity:
TN/TN +FP
Number needed to treat:
NNT
Number of pts that need to be treated to prevent 1 adverse outcome/event
Need absolute risk reduction %
1-ARR
ARR = Control event rate (%) - Experimental event rate (%)
Prevalence?
How many people have the disease right now
Divide number of pts with disease by total number of pts studied
TP + FN/All pts studied
Incidence?
New cases of a dx arising in a population over a given period
PPV?
TP/TP + FP
NPV:
TN/TN + FN
Hawthorne effect:
Observer effect
Subjects change their behavior when they know they’re being watched
Correlation coefficients;
Used to measure how strong a relationship is between two variables
Most popular one used is Pearson’s R correlation coefficient; -1 to 1
1 means there’s is a strong correlation between the two variable
Prevalence and sensitivity/specificity;’
Prevalence has no effect on sensitivity/specificity
Prevalence will affect the PPV and NPV
