Medical Application of Statistics

Special Medical Application for Statistics- 

Use- Apply statistics to medical practice 

Special Terms include: 

Double Blind Study:Subjects are randomly assigned to different treatment groups. Neither the investigator nor the subject knows which is receiving an experimental treatment or a control experience. This often times is a drug verses the placebo. 

Reliability: How well a test can be reproduced with Results similar to original test. 

Meta-analysis: Pooling results from several previous studies to achieve greater statistical power. Data is acquired from previous studies. 

Case Control Study: Observational study, sample chosen based on the presence/ absence of disease. An example of this can often times be seen in a study with smoking and lung cancer.

Cohort Study: Observational study, based on the presence/ absence of risk factors and subjects followed over a period of time. 

Clinical Trial: Experimental study, compares therapeutic benefit of two or more treatments. 

Prevalence: A proportion of Individuals in a study population with a disease at a specific point in time. 

Incidence: Number of new cases diagnosed over a specific time interval. 

Sensitivity: Probability that a person with the disease will be correctly identified by a test for the disease. 

Odds Ratio: Expresses chance of having or developing a disease in a group exposed to risk factors. 

Chi Square

Chi Square Test- 

Use: Examine Differences in distributions of nominal data. This was developed by Pearson in the early 1900’s. 

The Chi Square test contains two types- 

·     Goodness of fit- This test determines whether an observed distribution fits a theoretical distribution. This is commonly used in Genetics. 

·     Contingency Analysis- This test determines whether an observed distribution from one sample is equal to an observed distribution from a second sample. An example of this is if one wanted to look at AIDS in North America compared to AIDS in Africa. 

Linear Regression/ Correlation

Linear Regression/ Correlation 

Use: Tests the relationship among variables. It can have a positive relationship, a negative relationship, or no relationship. 

Correlation:  

*Tests for significant relationships between two variables.

* Measures the strength of association between to variables

* results in the correlation coefficient 

Mechanics of Correlation Analysis: 

*Statistical Hypothesis 

* Calculate the correlation coefficient 

* Conduct Statistical Tests 

* Coefficient of determination 

Multiple Comparisons Test

Multiple Comparisons tests- 

Use: If the results of an ANOVA yield a significant difference. ANOVA results only indicate that a difference exists among the mean, not where the difference is. 

Multiple Comparison Tests are divided into three broad categories. 

*Generic multiple comparisons tests- This test evaluates all possible pairs/ combinations of means. It includes Tukey’s HSD test and Student- Newman Kuels. 

*Control group tests- Evaluates the differences between experimental group verses the control group. This includes the Dunnett’s test. 

*Multiple contrasts tests- Can be used like the traditional test methods to evaluate differences among pairs of means but it is better used to evaluate homogeneous groups of means against other groups or individual means. This includes the Scheffe test.