The Herald story is based on a one-year change with the number deaths of people aged under 25 falling in 2017 but rising last year. One-year changes are meaningless. Here is the longer trend showing that deaths of undr-25s remain at around only 5% of the total in sharp contrast to those of the 35-44-year-olds:
Health trends, as any fool knows, are best understood in longer blocks of time because of the dangers in over-reacting to one-year fluctuations. For example, for this same group, there were 100 deaths in 2002 but only 47 by 2005, then 94 in 2007, then 65 in 2010, 32 in 2013, 46 in 2014 and 30 in 2015.