Cohort

- pick exposed and non exposed groups and compare their incidence of disease

retrospective cohort

-examines two groups with different exposures and compares their outcomes retrospectively usually by chart review -the exposure outcome would have already occurred at the time of the study and the data is collected after-the-fact

ANOVA ??? MANOVA ???

ANOVA - used to compare the means of 3 or more independent groups with a single dependent variable (parametric data) MANOVA - more than one dependent variable and when 3 or more independent groups

Wilcoxan Rank Sum test ???

- compares the MEDIANS of two independent populations with non-parametric data

Mann-Whitney U test ???

- Used for comparing the MEANS of two independent groups that are nonparametric

Kruskal Wallis Test ???

- used for THREE or more independent groups with data that is assumed to be non-parametric

Student’s t test ??? Paired t test ???

Student’s t test - compares the means of 2 sets of continuous data that have a normal distribution Paired t test - groups are paired (ie before and after an intervention)

Chi square & Fisher’s Exact test ???

- comparison of proportions of two groups of categorical variables -if > 5 measurements → Chi square -if <5 measurements → Fisher Exact

Kolmogorov-Smirnov & Shapiro-Wilk ???

- tests to determine whether data have a normal distribution

case-control study ???

-is an observational analytic study following subjects on the basis of whether they do or do not have a disease to determine the possible effect of an exposure

case series ???

is a descriptive study which describes a group of patients with the same disease process

Prevalence

-equals the number of existing cases of a disease for a total population at a given point in time -rate of disease at any one point in time that includes new and old cases

Incidence

-rate of new cases of a certain disease at any one point in time -measurement of the instantaneous rate of developing a disease in a population and equals the number of new cases of disease in a given time period over the total person – time of observation

Relative Risk

Relative Risk (same as cumulative incidence ratio): (risk=proportion) Unitless

= Proportion of disease in exposed/Proportion of disease in non exposed

= (a/a+b)/(c/c+d)

In words, if RR=2 then the exposed are 2 times more likely to develop disease than the nonexposed.

Lower limit is 0

Ratio of 1 means no association

Upper limit is infinity which depends on the incidence of the NONEXPOSED group

Odds ratio

Odds ratio: (ratio=odds) Unitless

= Odds of disease among exposed/odds of disease among non exposed

= ratio of dz:no dz in exposed group/ratio of dz:no dz in non exposed group

= (a/b)/(c/d) = a*d/b*c

Mann Whitney U test / Wilcoxon rank-sum test/ Wilcoxon signed-rank test

-(nonparametric)

-Comparison of median ranks for an ordinal type or continuous data

-ie. Difference in exam scores in two groups of students who were given computerized compared to paper exam format

Students t test

- (parametric)

-Comparison of mean for a continuous variable

-Change in cholesterol level between two groups randomized to two different medications

Analysis of variance (ANOVA)

-(parametric)

-Comparison of mean between more than two independent groups

-Difference in the white cell count in multiple groups of patients after diagnosed with a specific illness

Pearson coefficient of correlation

-(parametric)

-Measurement of degree of association or strength of linear relationship between two variables that are continuous

-Relationship between immunization rate and infant mortality rate

Spearman rank correlation

-(nonparametric)

-Measurement of degree of association between ranks of two variables (either both ordinal or one is ordinal and other is continuous)

-Relationship between height and weight in low birth weight babies over a time period