Introduction to Biostatistics 04 – Rates and Proportions
Introduction to Biostatistics 04 – Rates and Proportions
This lecture transitions from interval data to nominal data, where outcomes are categorical rather than measured on a scale (e.g., dead vs. alive, treatment vs. placebo). It introduces the Z-test for comparing two proportions and the Chi-square test for analyzing contingency tables. The session also covers essential epidemiology concepts such as Relative Risk and Odds Ratio, which are used to quantify the association between risk factors and clinical outcomes.
Learning Objectives:
By the end of this lecture, students will be able to:
- Define nominal data and explain why arithmetic means are replaced by proportions and rates in this context.
- Calculate the standard error of a proportion and perform a Z-test to compare two independent proportions.
- Apply Yates' correction for continuity to ensure accurate p-values when approximating discrete data with a continuous normal distribution.
- Construct and analyze contingency tables using the Chi-square test to compare observed frequencies against expected frequencies.
- Differentiate between Prospective and Case-Control studies, understanding which designs allow for the calculation of Relative Risk versus Odds Ratio.
Topics covered in this lesson
- Understanding the mean and standard deviation of Bernoulli trials explains how to analyze binary outcomes like success or failure.
- The Z-test for proportions provides a step-by-step method for testing the statistical significance of differences between two rates, such as mortality.
- Calculating expected values shows what categorical data would look like if there were no association between treatment groups and clinical outcomes.
- The Chi-square statistic measures how much real-world observations deviate from what is expected under the null hypothesis.
- Advanced Chi-square applications use row-by-column tables to compare three or more groups while applying corrections for multiple comparisons.
- Relative risk compares the probability of an event occurring in a treated group versus a control group during prospective studies.
- The odds ratio provides a way to estimate risk in retrospective case-control studies where the total population incidence is not known.
- Prioritizing patient outcomes, such as mortality and quality of life, over intermediate physiological markers distinguishes clinical variables from process variables.
External links
