Methodologic artifacts in adult sun-protection trends.

Abstract

BACKGROUND: Skin cancer is the most common of all cancers and a public health concern of increasing magnitude in the U.S. PURPOSE: A mid-year review of Healthy People 2010 found that the percentage of adults engaging in sun-safe practices increased from 59% in 2000 to 71% in 2005. This paper examines whether this increase in sun-safe practices in adults is an artifact caused by the change from 2000 to 2005 in the operational definition of "sun-safe practices" in the National Health Interview Survey (NHIS). METHODS: This study analyzed the 2000 and 2005 NHIS data sets in 2009 to examine changes in sun-safe practice prevalence estimates and to estimate the relative odds of practicing sun-safe behaviors associated with gender, age, race/ethnicity, education, and geographic region. RESULTS: Contrary to the Healthy People 2010 mid-year review, it was found that engaging in sun-safe practices by adults decreased from 59% (2000) to 55% (2005). A question order effect and use of images for use of sun-safe hats likely caused an artificial change in "cover-up" behavior estimates from 31% (2000) to 18% (2005) and wide-brim hat usage from 27% (2000) to 12% (2005). When eliminating data from these two questions added in 2005, the relative odds of practicing sun-safe behaviors was significantly lower for men in both 2000 and 2005 but were not different for other demographic characteristics. CONCLUSIONS: The current analyses suggest that the increase in sun-safe behaviors in adults and a notable decrease in "hat use" may be due to methodologic artifacts. When operational definitions change, caution should be used in interpreting estimates over time.