Level of Medical Intervention in Geriatric Settings: Decision Factors and Correlation With Mortality.

OBJECTIVES:Level of medical intervention (LMI) has to be adapted to each patient in geriatric care. LMI scales intend to help nonintensive care (NIC) decisions, giving priority to patient choice and collegial discussion. In the present study, we aimed to assess the parameters associated with the NIC decision and whether these parameters differ from those associated with in-hospital mortality. DESIGN:Prospective observational study. SETTING AND PARTICIPANTS:All consecutive patients from a French 62-bed acute geriatric unit over 1 year. METHODS:Factors from the geriatric assessment associated with the decision of NIC were compared with those associated with in-hospital and 1-year mortality, in univariate and multivariate analyses. RESULTS:In total, 1654 consecutive patients (median age 87 years) were included. Collegial reflection led to NIC decision for 532 patients (32%). In-hospital and 1-year mortality were 22% and 54% in the NIC group vs 2% and 27% in the rest of the cohort (P < .001 for both). In multivariable analysis, high Charlson Comorbidity Index [odds ratio (OR) 1.15, 95% confidence interval (CI) 1.06-1.23, per point], severe neurocognitive disorders (OR 2.78, 95% CI 1.67-4.55), dependence (OR 1.92, 95% CI 1.45-2.59), and nursing home residence (OR 2.38, 95% CI 1.85-3.13) were highly associated with NIC decision but not with in-hospital mortality. Conversely, acute diseases had little impact on LMI despite their high short-term prognostic burden. CONCLUSIONS AND IMPLICATIONS:Neurocognitive disorders and dependence were strongly associated with NIC decision, even though they were not significantly associated with in-hospital mortality. The decision-making process of LMI therefore seems to go beyond the notion of short-term survival.