The objective of study was to identify characteristics associated with being home at 6 months in 80 patients on prolonged mechanical ventilation (PMV) (> or =7 days). At 6 months, 47.5% were home, 13.8% institutionalized, and 38.8% deceased and classified "not home." Using signal detection methodology (SDM), four mutually exclusive groups at high and low probability of being home were identified. The best outcome (94.4% home) was achieved by patients with an admission Charlson Comorbidity Score < or =3 and an Acute Physiology Score (APS) < or =21 and the worst outcome (23.4% home) by patients with an admission Charlson Comorbidity Score >3 and Health Assessment Questionnaire score >2.7. SDM provided an effective means of identifying subgroups likely to be discharged home using available information.