Background: Bias with regard to participation in epidemiological studies can have a large impact on the\r\ngeneralizability of results. Our aim was to investigate the direction and magnitude of potential bias by comparing\r\nhealth-related factors among participants and non-participants in a MRI-study based on HUNT, a large Norwegian\r\nhealth survey.\r\nMethods: Of 14,033 individuals aged 50ââ?¬â??65, who had participated in all three large public health surveys within the\r\nNorwegian county of Nord-TrÃ?¸ndelag (HUNT 1, 2 and 3), 1,560 who lived within 45 minutes of travel from the city\r\nof Levanger were invited to a MRI study (MRI-HUNT). The sample of participants in MRI-HUNT (n = 1,006) were\r\ncompared with those who were invited but did not participate (n = 554) and with those who were eligible but not\r\ninvited (n = 12,473), using univariate analyses and logistic regression analyses adjusting for age and education level.\r\nResults: Self-reported health did not differ between the three groups, but participants had a higher education level\r\nand were somewhat younger than the two other groups. In the adjusted multivariate analyses, obesity was\r\nconsistently less prevalent among participants. Significant differences in blood pressure and cholesterol were also\r\nfound.\r\nConclusion: This is the first large population-based study comparing participants and non-participants in an MRI\r\nstudy with regard to general health. The groups were not widely different, but participants had a higher level of\r\neducation, and were less likely to be obese and have hypertension, and were slightly younger than nonparticipants.\r\nThe observed differences between participants and non-invited individuals are probably partly\r\nexplained by the inclusion criterion that participants had to live within 45 minutes of transport to where the MRI\r\nexamination took place. One will expect that the participants have somewhat less brain morphological changes\r\nrelated to cardiovascular risk factors than the general population. Such consequences underline the crucial\r\nimportance of evaluation of non-participants in MRI studies.
Loading....