The Funen Neck and Chest Pain study: analysing non-response bias by using national vital statistic data

To describe the Funen Neck and Chest Pain (FNCP) study and carry out a comprehensive non-response analysis of the quality of the survey.

The FNCP questionnaire was sent out to 7000 randomly selected individuals aged 20-71 years living in Funen County, Denmark. A full description of the FNCP survey, analysis of selection bias (representativeness of the background population), selective bias (non-responder bias), and item non-response bias was performed by using Danish vital statistics.

The adjusted response rate was 60%. Men, retired individuals, and individuals with lower income tended to be late responders. Women, people aged 50+, married individuals, and individuals with two or three children, with a higher educational level, living in a single house or with a high income were more likely to participate. Conversely, men, younger individuals, singles or divorced persons, individuals living in residential caring homes, and with lower educational level were less likely to participate. Adjustments based on design and logical omissions gave an overall missing data rate of 1.0%. In general, the frequency of missing data increased with age and was higher for women. The frequency of inconsistent answers was 0.34%.

A comprehensive non-response analysis showed some sociodemographic discrepancies between the FNCP study and the background population. The pattern of missing data was strongly associated with the design of the questionnaire and with participants' willingness to answer only questions they considered relevant. These factors must be taken into consideration when results from the FNCP study are presented.