Do tools aimed at avoiding hospital admission operate at different mortality thresholds? A systematic review

To determine whether front-door discharge decision tools operate at different mortality thresholds.

Three databases  searched, for studies testing, deriving or validating front-door risk prediction tools or discharge decision aids, with  defined discharge 'cut-off', reporting mortality or readmission rates. Studies supporting tools' inclusion in national guidelines were also included.

Twenty-four studies were included, frequently for acute chest pain. Mortality rates among those discharged based on tools 0-1.7%. Eight studies reported readmission rates, 0-8% among those discharged early or deemed low-risk.

Although mortality rates were lower for those deemed low-risk by decision aids than those admitted or control groups, readmission rates tended to be higher among low-risk or discharged patients, than among control group or admitted patients.