What is a type II error?
A type II error (β) is the probability not rejecting the null hypothesis when it is false --> false negative error
What is power?
The probability of rejecting the null hypothesis when it is false.
Power= 1 - β
You can increase power (by decreasing β) by increasing the sample size, expected effect size, or precision of measurement.
What is a type I error?
A type I error (α) is the probability of incorrectly rejecting a null hypothesis --> false positive error α is often set to < 0.05
Formula for calculating confidence interval?
Mean ± Z(SEM)
For 95% CI, Z = 1.96
For 99% CI, Z = 2.58
SEM = SD/√n
How to appropriately use statistical tests?
Compare means of 2 groups --> t-test
Compare means of 2+ groups --> ANOVA
Compare percentages or proportions of categorical variables in 2+ groups --> Chi-square
Describe accurary vs. precision
Accuracy = Validity = trueness of a test measure (absence of systematic error or bias)
Precision = reliability = consistency and reproducibility of a test (absence of random variation)
Sensitivity = true positive rate = a/(a+c)
SNOUT --> a highly sensitive test with a negative result rules out disease (low false negative rate)
A high sensitivity test is good for screening in populations with low prevalence
Specificity = true negative rate = d/(d+b)
SPIN --> a highly specific test with a positive result rules in disease (low false positive rate)
Used for confirmation after a positive screening test or in the case of severe disease
Describe positive predictive value
Positive predictive value = probability that a person actually has the disease given a positive test result = a/(a+b)
As prevalence ↑, PPV ↑
Describe negative predictive value
Negative predictive value = probability that a person is disease-free given a negative test result = d/(c+d)
As prevalence ↑, NPV ↓
Describe the distribution of a normal curve?
Describe how sensitivity and specificity change based on a test's cut-off point?
Sensitivity increases as the cut-off decreases/moves to the left.
Specificity increases as the cut-off increases/moves to the right
How do you calculate likelihood ratios?
Positive likelihood ratio = Sensitivity/(1-Specificity) = likelihood of having the disease given a positive result
Negative likelihood ratio = (1-Sensitivity)/Specificity = likelihood of having the disease given a negative result
How do you calculate attributable risk?
AR = Incidence in exposed - Incidence in unexposed
Attributable risk percent (ARP) = (RR-1)/RR
What is the rare disease assumption?
OR approximates RR when prevalence is low (<1%)
How do you calculate absolute risk reduction and number needed to treat?
ARR = Rate in untreated - Rate in treated
NNT = 1/ARR
How do you calculate relative risk reduction?
RRR = 1 - RR
How do you calculate number needed to treat and number needed to harm?
NNT = 1/(Rate in untreated - Rate in treated) = 1/Absolute risk reduction
NNH = 1/(Adverse event rate in treatment group - Adverse event rate in control group)
How are incidence and prevalence related?
Average duration of disease ≈ Prevalence/Incidence
As duration ↑, prevalence ↑