Sensitivity - probability of testing positive if you have the disease

Probability that a sick patient will have a positive test

Test’s ability to detect disease = TP/(TP+FN)

• With high sensitivity a negative test result means patient is very unlikely to have disease

Specificity - probability of testing negative if you don’t have the disease

Probability that a healthy patient will have a negative test

Test’s ability to state no disease is present = TN/(TN+FP)

• With high specificity, a positive test result means a patient is very likely to have the disease

Positive Predictive Value

TP/(TP+FP)

Likelihood that with a positive result, the patient actually has the disease

Negative Predictive Value

TN/(TN+FN)

Likelihood that with a negative result; the patent does not have the disease

Predictive value – Depends on disease prevalence while Sensitivity and specificity does not

Relative risk – (Incidence of exposed/incidence of unexposed)

• (a/a+b)/(c/c+d)
• < 1 = negative association. > 1 positive.  = 1 is no association
• Relative risk reduction (RRR) = 1 - Relative Risk (RR)

Odds ratio

• (a/c)/(b/d) = (AxD)/(BXC)

Confidence interval

• Mean of your estimate +/- the variation in that estimate
• Range of values you expect your estimate to fall between if you redo your test within a certain level of confidence
• Decrease in sample size will increase the confidence interval

Null hypothesis

• No difference exists between two groups

P Value

• <0.05 rejects the null hypothesis = > 95% likelihood that the difference between the populations is true and did not occur by chance

Type I error

• False positive Rejects null hypothesis when the null hypothesis is true
• Falsely assumed there is no difference when difference exists

Type II error

• False negative Accepts the null hypothesis incorrectly. Caused by small sample size.
• Falsely assumed there is a difference when no difference exists

Power of the test - Probability of making correct conclusion = 1 – probability of type II error = Beta

• Likelihood that the conclusion of the test is true
• Larger sample size increases power

Measures of central tendency = mean, median, mode

Variance - spread of data around a mean

Parameter - population

Prevalence - total cases in a population at a specific point of time

• Proportion of disease affecting a particular population = (True Positives + False negatives)/Total number of studied individuals
• Answers - how many individuals have this disease right now?

Incidence - number of NEW cases in a specific population over a specific period of time

• Incidence - how many individuals per year newly acquire this disease?

Categorical data - Represents characteristics, gender, language