And so is the potential cost of not doing anything, so what's your point?
You seem to conjecture that refering someone to a hotline raises the chance of that person commiting suicide, while AFAIK all the evidence points towards suicide hotlines being effective at reducing stress and suicidality.