Medical Statistics Case Study

Predicting Carotid IMT

An applied walkthrough of multiple regression analysis examining cardiovascular risk factors.

N = 100
Medical Dataset
SPSS

Learning Goals

1

Follow a real medical dataset example.

2

Analyze regression reporting in research.

3

Interpret coefficients meaningfully.

1) Research Context

Researchers are examining how cardiovascular risk factors affect Carotid Intima–Media Thickness (IMT), a critical indicator of atherosclerosis.

Dependent Variable (Y)
IMT (Intima-Media Thickness)
Cholesterol
Age
Blood Pressure
BMI

2) Running the Regression

Software: SPSS v28
Method: Enter (Standard Regression)
Stats: R², ANOVA, Coefficients, VIF, Durbin–Watson
Checks: Residual Plots (Normality, Homoscedasticity)

3) Key Results Analysis

Model Fit

46% Explained Variance (Adj. R²)

The model statistically significantly predicts IMT,
F(4, 95) = 21.3, p < .001

Diagnostics

No Multicollinearity (VIF < 2.0)
Assumptions Met (Residuals OK)

Coefficients Breakdown

Dependent Variable: IMT

Age

Most consistent effect

p < .001 B = 0.015

BMI

Largest effect size

p < .05 B = 0.020

Cholesterol

Significant predictor

p < .01 B = 0.002

Blood Pressure

Not significant

p = .08 B = 0.007
Interpretation: Age and Cholesterol are significant. BMI has the largest effect size (B=0.020), meaning it strongly predicts thickness, though with slightly less certainty than Age. Blood pressure was not a significant predictor in this specific dataset.

4) Writing up the results

APA Style

"A multiple regression was conducted to examine predictors of carotid intima–media thickness. The overall model was significant, F(4, 95) = 21.3, p < .001, and explained 46% of the variance in IMT (Adjusted R² = .46). Age (B = .015, p < .001), cholesterol (B = .002, p < .01), and BMI (B = .020, p < .05) were significant predictors, whereas blood pressure was not (p = .08)."

5) Quick Check-in

Download Full PDF Handout