TYPE OF ERRORS
- Your Hypothesis: difference exists between A and B
- Null Hypothesis: (Contradicts your Hypothesis) there is no difference between A and B
Type I Error: Incorrectly rejecting the null hypothesis
- (The study showed that there is difference but in fact there is not difference, you think your study was successful but in fact it wasn’t!)
- Type I Error = False Positive
- p-value: chance of making type I error
- (p<0.05 = <5% chance of commenting this error)
Type II Error: Incorrectly accepting the null hypothesis
- (The study showed that there is no difference but in fact there is difference, you think your study was not successful but in fact it was!)
- Type II Error = False Negative
- Increasing the power (bigger sample size) decreases this error
Type III Error: Conclusions not supported by data
95% confidence interval: If it includes de value 1, it is not statistically significant.
- The farther away form 1 the stronger the correlation (i.e., 9-10 or 0.1-0.2 has stronger correlation than 2-3 or 0.8-0.9)
Prevalence: # of patients having the disease in the population.
- (It’s higher in long lasting diseases)
Incidence: # of newly diagnosed cases in a population in a given time period of time
SCREENING AND DIAGNOSTIC TESTS
- Sensitivity: (analyzes the tests results (+) or (-) in the Patients with the disease/condition)
- TP / (TP + FN) True positive rate:
- The probability that a patient with the disease will have a positive test result.
- SnOut: a sensitive test with a (-) result its good at ruling-out the disease
- (You can trust Negative results)
- High Sensitivity = Low False Negatives
- Specificity: (analyzes the tests results (+) or (-) in the Patients without the disease/condition)
- TN/(TN+FP) True negative rate:
- The probability that a patient without the disease will have a negative test result
- SpIn: a highly specific test with a (+) result its good at ruling-In the disease
- (You can Trust Positive results)
- High specificity = Low False Positives
Predicted Values are dependent on the prevalence of the disease:
- Positive Predict Value: The probability that a person with a positive test result actually has the disease.
- (Prevalence is directly proportional to PPV)
- Negative predictive value: The probability that a patient with a negative test result really is free of the disease.
- (Prevalence is inversely proportional to NPV)
- Case Control: Retrospective
- Takes patients with the disease and look in the past to see what factors contributed to develop the disease.
- Uses Odds Ratio for the calculations: (TPxTN)/(FPxFN)
- Cohort study: Prospective
- Takes a group of pts exposed to a risk factor and a group of pts not exposed and follows them up for a couple of years to see how the disease develops, or if a drug has effect or not.
- Uses Relative Risk for the calculations:
- Incidence in exposed/incidence in unexposed
Clinical Trial: Randomized, Double blind, Multicenter, Placebo, control.
- Review and statistical
- Combining of data from different studies
- (Increases the power of any single study)
- Also use (also uses Odds Ratio)
- T test: Compares 2 groups (ex: means of weight b/t 2 groups)
- ANOVA: Is a t-test for more than 2 groups.
- ''Non-parametric statistics: for qualitative data analysis. Race, sex, medical problems and diseases, medications)
- Chi-square: compare 2 groups with categorical variables (obese patients with diabetes Vs. Obese patients without diabetes
- Kaplan-Meyer: (small groups) estimate the survival rate
- The average of the Test
- Central tendency in a Normal Distribution
- The confidence interval of the mean gives the answer
Variance: The spread of data around the mean
- Median: The middle value of a set of data.
- Central tendency in a NON normal Distribution
- Mode: The most frequent occurring value
Mode=5, Mean=6.8, Median= 7