Purpose: To assess physician decisions about ICU admission for life-sustaining treatments (LSTs).
Methods: Observational simulation study of physician decisions for patients aged ≥80 years. Each patient was allocated at random to four physicians who made decisions based on actual bed availability and existence of an additional bed before and after obtaining information on patient preferences. The simulations involved non-invasive ventilation (NIV), invasive mechanical ventilation (IMV), and renal replacement therapy after a period of IMV (RRT after IMV).
Results: The physician participation rate was 100/217 (46 %); males without religious beliefs predominated, and median ICU experience was 9 years. Among participants, 85.7, 78, and 62 % felt that NIV, IMV, or RRT (after IMV) was warranted, respectively. By logistic regression analysis, factors associated with admission were age <85 years, self-sufficiency, and bed availability for NIV and IMV. Factors associated with IMV were previous ICU stay (OR 0.29, 95 % CI 0.13-0.65, p = 0.01) and cancer (OR 0.23, 95 % CI 0.10-0.52, p = 0.003), and factors associated with RRT (after IMV) were living spouse (OR 2.03, 95 % CI 1.04-3.97, p = 0.038) and respiratory disease (OR 0.42, 95 % CI 0.23-0.76, p = 0.004). Agreement among physicians was low for all LSTs. Knowledge of patient preferences changed physician decisions for 39.9, 56, and 57 % of patients who disagreed with the initial physician decisions for NIV, IMV, and RRT (after IMV) respectively. An additional bed increased admissions for NIV and IMV by 38.6 and 13.6 %, respectively.
Conclusions: Physician decisions for elderly patients had low agreement and varied greatly with bed availability and knowledge of patient preferences.