Open ld-archer opened 1 year ago
C-Reactive Protein (inflammation indicator) distribution from ELSA is positively skewed, so need to be log transformed before use (if using).
From Shankar et al. 2011:
'Given the known relationship between marital status and many of the outcomes considered in this analysis, we repeated all analysis with a social isolation index excluding the marital status dimension. The pattern of results remained identical for health behaviors and cholesterol. In the case of blood pressure, CRP and fibrinogen, results were no longer significant when marital status was excluded from the index, suggesting that these effects are substantially driven by marital status.'
Lots of studied into loneliness/social isolation mention some biomarkers of inflammation, as inflammation is seen as a mechanism leading from these behaviours to poor health. This information is measured by ELSA, but doesn't seem to be available in the harmonised outputs. Need to investigate this and see if they are easily available, and if we can include them in the work.