Biostatistics Flashcards
Cross-sectional study
What is happening at a particular point in time? Good for determining disease prevalence and risk factors.
Case-control study
What happened in the past? Compares a group with disease to a matched group without disease and searches for associations. Calculate Odds Ratio (OR).
Cohort study
Can be prospective (Who will develop disease?) or retrospective (Who developed disease?). Compares a group with a certain exposure/risk to a group without that exposure and asks who developed or will develop disease? Calculate Relative Risk (RR).
Phase I of drug trial
Small number of healthy volunteers. Is it safe?
Assesses safety, toxicity, pharmacokinetics.
Phase II of drug trial
Small number of patients with the disease of interest. Does it work?
Phase III of drug trial
Large number of patients randomly assigned either to treatment or to best available treatment (standard of care). Is it as good as or better?
Phase IV of drug trial
Post-marketing surveillance of patients after treatment FDA approved. Can it stay? Detects rare or long-term adverse effects.
Evaluation of diagnostic tests for sensitivity and specificity
———Dz+ Dz-
Test+ TP (a) | FP (b)
Test- FN (c) | TN (d)
Sensitivity
= a/a+c
Specificity
= d/b+d
Positive predictive value (PPV)
= a/a+b = true positive/all positives
Negative predictive value (NPV)
= d/c+d = true negative/all negatives
Contingency table for quantifying risk
——–Dz+ Dz- (Disease/outcome)
Ex + a b
Ex - c d
(Exposure/Risk factor/Intervention)
Odds Ratio (OR)
= (a/c)/(b/d)= ad/bc
=Odds that the cases were exposed to risk versus the controls.
Used for case-control studies.
Relative Risk (RR)
= [a/(a+b)] / [c/(c+d)]
= Risk of developing dz in the exposed group/ risk of developing dz in the unexposed group.
Used for cohort studies.
IF prevalence of disease is low, then OR approximately equals RR.
RRR= 1-RR
Attributable risk (Absolute Risk Increase)
= [a/(a+b)] - [c/(c+d)]
= The difference in risk between the exposed group and unexposed group (or the difference in risk attributable to the exposure).
Used to calculate the NNH.
Relative risk reduction (RRR)
= 1 - RR
= 1- [a/(a+b)] / [c/(c+d)]
= Proportion of risk reduction attributable to the intervention.
Absolute Risk Reduction (ARR)
= [c/(c+d)] - [a/(a+b)]
= The difference in risk between the unexposed group and the exposed group (or the difference in risk attributable to the intervention).
Used to calculate the NNT.
Number Needed to Treat (NNT)
NNT = 1/ Absolute Risk Redution
= the number of patients that need to undergo an intervention for 1 patient to benefit.
Number Needed to Harm (NNH)
NNH = 1/ Attributable Risk
= the number of patients that need to be exposed to a risk factor (or intervention) for 1 patient to be harmed.
Incidence
= # new cases in a time period/# of ppl at risk
Prevalence
= # existing cases at one point in time/ # of ppl at risk
True negative
=Specificity * # pts without disease
= [d/b+d]*(b+d)
True positive
=Sensitivity * # pts with disease
=[a/a+c] *(a+c)
Positive likelihood ratio
=Sensitivity/ (1 - Specificity)
Negative likelihood ratio
=(1 - Sensitivity)/Specificity
95% Confidence Interval
= mean +/- 1.96 *SD/sqrt(n)
Type I error
Occurs when research rejects the null hypothesis (finds a difference), but the null hypothesis is true (there is no real difference). Aka a FALSE POSITIVE ERROR.
Alpha
The probability of making a Type I error
Type II error
Occurs with research fails to reject a null hypothesis (find a difference), but the null hypothesis is false (there is a real difference). Aka a FALSE NEGATIVE ERROR.
Beta
The probability of making a Type II error. Power of a study = 1 - beta
Statistical power
= (1 - beta). It represents a study’s ability to detect a difference (reject the null hypothesis) when on truly exists. Power is typically set at 80% and depends on the sample size and difference between outcomes.
