How to deal with non-participation in long-term follow-up

What is it about?

The present editorial comments a manuscript by Lutsey et al [2019], which highlighted that restrictive and, to a lesser extent, obstructive lung disease in midlife were associated with greater risk of incident dementia and mild cognitive impairment in the elderly.

Featured Image

Why is it important?

Lutsey et al [2019] performed a study with a very long follow-up (27 years). As a consequence participation rate was rather low: only 42% of the initial cohort took part in the final neurocognitive examination, and this could have biased their results. However: 1) Most of the people lost to follow-up could not participate simply because they had died. Moreover, participants and non-participants still alive had similar baseline characteristics, and largely differed from deceased subjects. 2) Potential selection bias was dealt with in multivariable analysis by adopting inverse probability weighting, where larger weights arere assigned to subjects less likely to participate. 3) The Authors were able to collect some information also on participating subjects through administrative health databases, which supported the findings recorded on participating subjects.

Perspectives