Population-based surveys on the frequency of common skin diseases in adults - is there a risk of response bias?

Population-based surveys on the frequency of common skin diseases are impor tant in determining the health needs of a community. As they rely on volunt ary presentation, there is a risk of response bias which may compromise the quality of the data obtained. The aim was to determine in what way and to what degree response bias may occur in a population-based survey on the fre quency of common skin diseases in adults. A follow-up study was conducted o n 1043 out of 2500 adults who did not attend for examination as part of a r andomized population-based survey on the frequency of common skin diseases amongst adults in Maryborough, Central Victoria, Australia. Nonrespondents were more likely to be at the extremes of age, retired, unemployed and less likely to report that they had a history of skin disease than the responde nts. Subsequent examination of a sample of the nonrespondents revealed they were more likely to have skin cancers and Campbell de Morgan angiomas than the respondents. These differences cannot be fully explained on the basis of an age-related response bias alone. Response bias is a risk in populatio n-based surveys of common skin diseases which rely on voluntary presentatio n. Some attempt should be made to sample the nonrespondents in these survey s to determine the nature and extent of any bias and to adjust for it, if n ecessary.