Question

Study Designs - Part 1

INSTRUCTIONS: In working through these questions, take time in designing your study to consider how best to reduce confounding, various biases, and improve generalizability.

Randomized Trials

Imagine that you just read the following report:

Antioxidants Don’t Lessen Strokes for at-Risk Women

(From a nytimes.com article by Nicholas Bakalar, published 8/21/2007)

“Supplements of the antioxidants beta carotene, vitamin C and vitamin E may be good for you, but a new study reports that they have no effect, either alone or in combination, in preventing heart attack, stroke or death among women at risk for cardiovascular disease.”

Researchers randomly assigned more than 8,000 women to take regular doses of vitamin E, vitamin C, beta carotene, or placebos, and followed them for more than nine years. All the women, whose average age was 60, either had cardiovascular disease or were at high risk for it. During the nine years, 1,450 women had a heart attack, a stroke, or cardiac surgery, and there were 365 deaths from heart disease.

Women in the vitamin E group had a slight decrease in disease compared with the placebo group, but it was not statistically significant. Neither beta carotene, nor vitamin C had any statistically significant effect compared with placebo.

Combinations of the antioxidants had no effect either, except for a slight reduction in stroke among those taking both vitamins C and E together. The study appears in the Aug. 13 issue of The Archives of Internal Medicine.

“While the individual supplements may not decrease risk,” said Nancy R. Cook, the lead author and an associate professor of epidemiology at Harvard, “it does seem that diets high in fruits and vegetables that contain these antioxidants are helpful. It may be that we haven’t identified the particular nutrients or combinations of nutrients that might be beneficial.” “

1. Be an epidemiologic critic: What additional information would you want to know about this study before you believed the conclusions? Are there ways in which the study may not have been valid?

2. What recommendation(s) would you make to groups of women considering dietary supplements?

Answer 1

1. The additional information which I would want to know about the study before I believed the conclusions were:

• Whether the sources of Vitamin E, Vitamin C, beta carotene were from fruits and vegetables or from artificial supplements.
• What were the doses of Vitamin E and Vitamin C given to the women?
• Were the effects of antioxidants studied by varying and alternating the concentrations of each antioxidants to determine the results?

Yes, there are ways in which the studies may not have been valid. First, there should have been a separate study for the women who had the cardiovascular diseases and those who had a high risk for developing the same. While the deaths occured, it was important to go through the medical history of the patient and then draw conclusions. Because, while vitamin E and C are recommended for prevention of cardiovascular diseases, beta carotene was not advisory. Further, there may be trace elements which also might play an important role in prevention of the diseases. They may have an regulatory effect over these antioxidants. Those were not identified. So, these are questions which needs to be addressed and in this ways studies may not have been valid.