The length of hospitalization after acute myocardial infarction determined by risk calculation.

The feasibility of the use of a Cox model for risk assessment of individualized hospital discharge after myocardial infarction was evaluated. First, a previously developed prognostic index computed at the fifth day after admission was tested on a new population of 1140 patients. It was confirmed that after 5 days half of the patients (52%) could be discharged with low risk. Second, a new competing risk variant of the Cox model that updates prognosis according to the occurrence of complications was developed that describes the risk of death, cardiac arrest, and cardiogenic shock within 44 days after hospital admission. With a risk of one of these events being below 2% during a 14 day period after proposed discharge, 453 patients of 966 survivors (47%) could be discharged after only 5 days. A longer stay (of up to 30 days) was proposed for 338 patients (35%) to achieve the same level of risk. The savings in hospitalization days would be 15%. These results were confirmed in a new sample of 197 patients from the same institution who were discharged according to the proposed system. Of the 169 day 5 survivors, 67% were discharged on days 6 through 15 and this resulted in only two unexpected deaths and a 20% savings in hospitalization days. We conclude that individually determined discharge time is feasible without increased risk of death of severe complications after early discharge.