The competing risk between in-hospital mortality and recovery: A pitfall in COVID-19 survival analysis research