Obesity and Heart Disease

Share:

Description

To better understand whether the relationship between heart disease and obesity could be due to the social stigma associated with obesity, researchers examined cardiovascular disease rates of obese and non-obese women in American Samoa, where obesity was considered socially desirable. 3,112 women were categorized according to whether they were obese or not and whether they died from cardiovascular disease (CVD).

Usage

1

Format

A data frame with 2 observations on the following 3 variables.

Obesity

a factor with levels "NotObese" and "obese"

Deaths

the number of women who died from CVD

NonDeaths

the number that died from other causes

Source

Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cenage Learning.

References

Crews, D.E. (1988). Cardiovascular Mortality in American Samoa, Human Biology 60: 417–433.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
str(case1801)
attach(case1801)
   
## EXPLORATION
myTable             <- cbind(Deaths,NonDeaths)  # Form a 2 by 2 table of counts
row.names(myTable)  <- Obesity  # Assign the levels of Obesity as row names  
myTable   # Show the table

## INFERENCE (4 methods for getting p-values and confidence intervals)
prop.test(myTable, alternative="greater", correct=FALSE) # Compare 2 proportions
prop.test(myTable, alternative="greater", correct=TRUE) # ...with cont. corect.  
prop.test(myTable,correct=TRUE) # 2-sided alternative (default) to get CI
chisq.test(myTable) # Pearson's Chi-Squared Test
fisher.test(myTable, alternative="greater") # Fisher's exact test  
fisher.test(myTable) # 2-sided alternative to get CI for odds ratio
myGlm1  <- glm(myTable ~ Obesity, family=binomial)  # Logistic reg (CH 21)
summary(myGlm1)  # Get p-value-- 0.734
beta    <- myGlm1$coef
exp(beta[2])  #Odds of death are estimated to be 17% higher for  obese women 
exp(confint(myGlm1,2)) # 95% confidence interval  


## GRAPHICAL DISPLAY FOR PRESENTATION
myTable
#        Deaths NonDeaths
#Obese        16      2045
#NotObese      7      1044
prop.test(16,(16+2045)) #For one proportion, est: 0.0078 95% CI: 0.0046 to 0.013
prop.test(7,(7+1044)) #For one proportion, est: 0067 95% CI: 0.0029 to 0.014
pHat    <- c(0.007763222, 0.006660324)*1000 # Get estimated deaths per 1,000 women
lower95 <- c(0.00459943, 0.002921568)*1000
upper95 <- c(0.01287243, 0.014318321)*1000

if(require(Hmisc)) { # Use Hmisc library 
  myObj   <- Cbind(pHat,lower95,upper95) 
  Dotplot(Obesity ~ myObj,   # Draw a dot plot of estimates and CIs
    xlab="Estimated CVD Deaths Per 1,000 Women (and 95% Confidence Intervals)",
    ylab="Weight Category", ylim=c(.5,2.5), cex=2)
}


detach(case1801)

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.