High Yield- Stats Stats

Question Answer
What are the 3 types of data? nominal, ordinal, continuous
Which data type is parametric? continuous
Which data type are non-parametric? nominal and ordinal
What tests are used for 2 independent samples for Continuous data? 1. Student's t-test
2. Mann-Whitney U
What tests are used for 2 independent samples for Ordinal data? 1. Mann-Whitney U
2. Wilcoxon Rank Sum
What tests are used for 2 independent samples for Nominal data? 1. Chi-square
2. Fisher's Exact
What tests are used for related/paired samples for Continuous data? Paired t-test
What tests are used for related/paired samples for Ordinal data? 1. Sign test
2. Wilcoxon Signed Rank
What tests are used for related/paired samples for Nominal data? McNemar Test
What tests are used for ? 3 independent samples for Continuous data? 1-way ANOVA
What tests are used for ? 3 related samples for Continuous data? 2-way ANOVA
What tests are used for ? 3 independent samples for Ordinal data? Kruskal-Wallis one way ANOVA
What tests are used for ? 3 related samples for Ordinal data? Freidman 2 way ANOVA
What tests are used for ? 3 independent samples for Nominal data? Chi-square for k independent t samples
What tests are used for ? 3 related samples for Nominal data? Cochran Q
Measures of correlation for Continuous data? Pearson's Correlation
Measures of correlation for Ordinal data? 1. Spearman
2. Kendal rank
3. Kendal Coe
Measures of correlation for Nominal data? Contingency coefficient
Mode is usually associated with what type of data? nominal
Median is usually associated with what type of data? 1. ordinal
2. continuous not normally distributed (outliers)
Mean is usually associated with what type of data? continuous
When the data is normally distributed (Bell shaped curve), what happens to the mean, median, and mode? mean=median=mode
standard deviation formula? SD= vVariance
Standard Error of the Mean formula? SEM= SD/vn
What is 95 % confidence interval? 95% confident that the true population mean falls between this interval
Positively (right) skewed data, what happens to mean median mode? Mean > median > mode
Negatively (left) skewed data, what happens to mean median mode? Mode > median > mean
Nominal data is also known as? Categorical
Difference between interval data and ratio data? Interval has no absolute zero whereas ratio data does
When do you use Fisher Exact vs Chi square in nominal data? Fisher Exact is used for sample size < 40
When do you use Mann-Whitney vs Student's t-test for continuous data? Mann-Whitney is used for non-parametric data or outliers
Power formula? Power = 1-beta
What is beta? The probability of making a type 2 error
What is a type 2 error? Accepting the null when there is a difference
What 2 factors impact Power the most? alpha and sample size
What are the percentages of 1, 2, and 3 standard deviations away from the mean? 68, 95, 99.7 %
What does the p-value tell you? it determines the likelihood that a difference is seen is due to change or random error
Does a p-value tell you clinical significance? No
Is it statistically significant if a 95% CI for a mean contains "0.00?" No; you must accept the null
Is it statistically significant if a 95% CI for Odds ratio or Relative Risk contain "1.00?" No; you must accept the null
What is relative risk? the probability of an event occurring (for example, developing a disease, being injured) in an exposed group to the probability of the event occurring in a non-exposed group.
What type of study is relative risk done? cohort or clinical trials (prospective study)
What type of study is odds ratio done? case control study (retrospective)
Relative risk formula? (incidence rate in exposed pts) ? (incidence rate in non-exposed pts)
Relative risk reduction? 1 minus relative risk
Number needed to treat (NNT) formula? 1/ARR
Absolute Risk Reduction (ARR) formula? AR (control) – AR (treatment)
AR (absolute risk)? AR (absolute risk) = the number of events (good or bad) in treated or control groups, divided by the number of people in that group.
What is Odds Ratio? it is the Relative Risk in "retrospective study"
Odds Ratio formula? OR = AD/BC
correlation coefficient? r= -1 to +1; strength of "relationship" between 2 variables. It DOES NOT imply causation or predictability
At what Power is acceptable? 0.8
What is descriptive study design? document experience; case reports, clinical series
What are experimental study design? clinical trials, educational intervention, health-care trial
What are observational study design? case control, cohort, cross-sectional, correlational
What is the Naranjo ADR Probability Scale? the likelihood of whether an ADR (adverse drug reaction) is actually due to the drug rather than the result of other factors
Case-control study? retrospective, Odds Ratio
Cohort study? prospective, Relative Risk
Cross-Over study? prospective, wash-out period, then switch interventions; pts are serving as their own controls.
What is Double Dummy? investigator and patient doesn't know the treatment groups
CONSORT statement? Consolidated Standards of Reporting Trials. Help readers better understand study design. Aid in assessing the validity of the results
Selection Bias? Subjects not randomly assigned
Sampling bias? Sample not representive of population. e.g. Pediatric sample extrapolated to adults
Hawthorne effect? Subjects change their normal behavior because they know they are being studied.
Internal Validity? the independent variable was in fact responsible for the change seen in the dependent variable
External Validity? Is concerned with the generalizability of the results from sample to population
Cost-minimization? meds are therapeutically equal but financial costs are different
Cost-effectiveness? Used when interventions are not considered equal. Good value for amount of money you paid. You may not pick the least expensive. measured in nonmonetary values, e.g. years of life.
Cost-benefits? quantitative approach for deciding whether to go ahead with a decision. measured in dollars
Cost-utility? Quality of adjusted life-year(QALY)
Direct Medical Costs? costs incurred for medical products and services; costs of drugs, medical supplies, and equipment, laboratory and diagnostic tests, hospitalizations, and physician visits
Direct Non-medical costs? costs for nonmedical services that are results of illness or disease, e.g. travel time, living situation, bus fare, child care
Indirect Costs? costs of reduced productivity; lost productivity measured in lost income from wages
Intangible Costs? nonfinancial outcomes of disease and medical care, e.g. pain, suffering, inconvenience, and grief
Sensitivity? proportion of pt with disease who test positive (all true positives); used to rule "OUT" disease. SNOUT
sensitivity formula? SN= TP/(TP+FN)
Specificity? proportion of pt without disease who test negative (all true negatives). used to rule in disease. SPIN; used in HIV
specificity formula? SP= TN/(TN+FP)
Precision? reliable; consistent and reproducible test results BUT does not mean it is accurate to true value
Accuracy? trueness of test. on the mark.
Accuracy formula? Accuracy = (TP + TN)/ (total number of samples)
Receiver Operator Characteristic (ROC)? a decrease in sensitivity results in an increase in specificity.
Positive Predictive Value (PPV)? probability that a patient with positive test result actually has the disease
Negative Predictive Value (NPV)? probability that a patient with negative test result DOES NOT have the disease
Positive Predictive Value equation? PPV= TP/(TP+FP)
Negative Predictive Value equation? NPV= TN/(TN+FN)
Prevalence? Total # of cases at a given time/total population at that time; it's a snapshot; High prevalence to RULE IN a diagnosis; Low prevalence to RULE out
Incidence? new cases in a time period/total population at risk during that time
Positive Likelihood Ratio equation? LR+ =(TP%/FP%)
Negative Likelihood Ratio equation? LR- =(FN%/TN%)
Systematic Review? answers a defined research question by collecting and summarising all evidence. Studies are qualitative
Meta-Analysis? combine the statistical results from multiple studies in an effort to increase power to resolve uncertainty when reports disagree. Studies are quantitative
Which well known group does systematic reviews? Cochrane Collaboration
Which publishing criteria is used to provide minimum set of items for systematic reviews and meta-analysis? PRISMA= Preferred Reporting Items for Systematic Reviews and Meta-Analysis