Many people in the United States and Europe use hair dyes. It is estimated that more than one-

third of women over age 18 and about 10 percent of men over age 40 use some type of hair dye.

Over 5,000 different chemicals are used in hair dye products. Early hair dye formulations

contained chemicals, including aromatic amines, that were found to cause cancer in animals. In

the mid- to late 1970s, however, manufacturers changed the components in dye products to

eliminate some of these. It is not known whether some of the chemicals still used in hair dyes

can cause cancer. Given the widespread use of hair dye products, even a small increase in risk

may have a considerable public health impact, so scientists have tried to determine if there is an

association with human cancer.

Over the years, some population studies have found an increased risk of bladder cancer in

hairdressers and barbers. In addition, studies have linked the personal use of hair dyes with

increased risks of certain cancers of the blood and bone marrow, such as non-Hodgkins

lymphoma (NHL) and leukemia, but other studies have not shown such links. Studies of breast

and bladder cancer have also produced conflicting results. Relatively few studies have been

published about the association of hair dye use with the risk of other cancers. Based on its review

of the evidence, IARC concluded that personal use of hair dyes is "not classifiable as to its

carcinogenicity to humans". [Source: National Cancer Institute]

Case study/Questions: You are an epidemiologist studying use of hair coloring as a risk factor for

two specific cancer outcomes: NHL and bladder cancer. You decide to approach the Centers for

Disease Control & Prevention (CDC) for support to conduct your study.

3. (3 pts) What are the major biases you'd expect to encounter in carrying out your study? How

would you set up your study to minimize the impact of these potential biases? What impact

could these biases have on the validity and results of your study?

4. (2 pt) You completethe study and report that indeed, there appears to be an association of

NHL with prior history of long-term use of hair coloring. The association is statistically

significant at the 95% level. What additional information might increase your belief that this is a

true effect, rather than an artifact?