Product HomeAbout the Product Web Site Resources
Statistics for Business and Economics: Excel/Minitab Enhanced
Heinz Kohler
    buy online



Statistics in the News: Chapter 12 Estimation

Assessing the Risk of Cardiovascular Disease

Pharmaceutical companies are forever searching for new and better drugs that may dramatically impact their profits. For that reason, they seek to learn all they can about the causes of diseases and how these causes might respond to various drugs. The Framingham Heart Study recently provided interesting data on the relationship between blood pressure and the incidence of cardiovascular disease (including such "events" as congestive heart failure, myocardial infarction, and strokes).

The 10-year study involved 6,859 subjects-all of whom were initially free of hypertension (high blood pressure) and cardiovascular disease. The subjects were divided into three blood pressure categories, as defined in Table A.

Table A. Blood Pressure Categories

Pressure

Optimal

Normal

High-Normal

Systolic

<120 mm Hg

120-129 mm Hg

130-139 mm Hg

Diastolic

<80 mm Hg

80-84 mm Hg

85-89 mm Hg

The study showed a clear positive relationship between blood pressure and the incidence of cardiovascular disease: The higher the blood pressure, the greater was the risk of such disease. Some of the relevant data for men and women appear in Table B.

Table B. The 10-Year Incidence of Cardiovascular Disease

Blood Pressure Category

Number of Subjects in Study

Number of Cardiovascular Events

Incidence in Percent

95% Confidence Interval

Men

Optimal

1,005

55

5.8

4.2 - 7.4

Normal

1,059

96

7.6

6.0 - 9.1

High-Normal

903

108

10.1

8.1 - 12.1

Women

Optimal

1,875

26

1.9

1.1 - 2.7

Normal

1,126

40

2.8

1.9 - 3.8

High-Normal

891

72

4.4

3.2 - 5.5

Note the rising incidence numbers (along with their associated confidence intervals in the last column of Table B), as blood pressure rises. Compared to optimal blood pressure, the authors noted, higher blood pressure categories for either sex were invariably associated with greater risk of cardiovascular disease.

Table C provides an alternative view. It shows how much more hazardous higher blood pressure is after adjustments are made for the subjects' age, body mass index, total cholesterol levels, presence or absence of diabetes mellitus, and smoking status. The authors fitted a complex regression model for the purpose. Statisticians will be interested to learn that the p values associated with the Table C data (indicating the likelihood of being able to attribute the results to mere sampling error) were 0.01 for men and < 0.001 for women.

Table C. HAZARD RATIOS

Blood Pressure Category

Hazard Ratio

95% Confidence Interval

Men

Optimal

1.0

 

Normal

1.3

1.0 - 1.9

High-Normal

1.6

1.1 - 2.2

Women

Optimal

1.0

 

Normal

1.5

0.9 - 2.5

High-Normal

2.5

1.6 - 4.1

Sources: Adapted from Ramachandran S. Vasan et al, "Impact of High-Normal Blood Pressure on the Risk of Cardiovascular Disease," The New England Journal of Medicine, November 1, 2001, pp. 1291-1297, and Julio A. Panza, "High-Normal Blood Pressure-More 'High' Than 'Normal'" ibid., pp. 1337-1340.


Copyright © 2003 South-Western. All Rights Reserved. Disclaimer
Webmaster