“…This is mainly due to extensive data demands, in the form of large, high quality, longitudinal panel data spanning decades, but also the advanced methodological nature of the CAD debate, which often means restricting analysis to a well know example. As a high quality data infrastructure on ageing has matured, and a consensus on how to investigate CAD has emerged, exemplified by Wilson, Shuey and Elder’s three central tenets, the occurrence and strength of this process has been tested in contexts other than the US: across the whole of Europe [ 25 , 35 ], in individual European countries such as Germany [ 19 ], Sweden [ 18 ] and Switzerland [ 5 ], as well as comparing several European countries with contrasting positions in terms of their welfare systems [ 32 , 33 ]. These studies illustrate that the US context is exceptionally inductive to CAD processes, combining high initial levels of health problems, with a large educational health gap, both widening over time within cohorts, as well as widening up for younger cohorts [ 33 ].…”