BACKGROUND: The incidence of withholding and withdrawing life support from the critically ill has increased in recent years. The aim of this study was to assess the degree of consistency between the weight assigned by intensivists to different determinants and their relation to end-of-life decisions, and to evaluate the current concepts in withholding or withdrawing intensive care in Nordic countries. METHODS: Forty-one intensivists from Nordic countries completed a questionnaire sent by e-mail: consistency between contributing factors and the decisions regarding 10 actual cases was evaluated by logistic regression analysis and by the classification (leave-one-out) method. Concepts in management after the withdrawal decision were also analyzed. RESULTS: The median (range) number of withdrawals per physician was four (range 0-10) out of 10 cases. No single factor was an independent covariant of all decisions made. The classification method revealed that approximately 70% only of decisions could be predicted correctly. Different actions taken after a decision to withdraw intensive care varied from 9.8% (discontinuing ventilator therapy) to 97.6% (informing relatives). CONCLUSIONS: No generally accepted grounds for end-of-life decisions could be detected among Nordic intensivists. In addition, the current concept of management after decision to withdraw therapy varies markedly. This study has implications in further assessment of the individual decision-making process and the uniformity of actions after withdrawal decisions.
BACKGROUND: The incidence of withholding and withdrawing life support from the critically ill has increased in recent years. The aim of this study was to assess the degree of consistency between the weight assigned by intensivists to different determinants and their relation to end-of-life decisions, and to evaluate the current concepts in withholding or withdrawing intensive care in Nordic countries. METHODS: Forty-one intensivists from Nordic countries completed a questionnaire sent by e-mail: consistency between contributing factors and the decisions regarding 10 actual cases was evaluated by logistic regression analysis and by the classification (leave-one-out) method. Concepts in management after the withdrawal decision were also analyzed. RESULTS: The median (range) number of withdrawals per physician was four (range 0-10) out of 10 cases. No single factor was an independent covariant of all decisions made. The classification method revealed that approximately 70% only of decisions could be predicted correctly. Different actions taken after a decision to withdraw intensive care varied from 9.8% (discontinuing ventilator therapy) to 97.6% (informing relatives). CONCLUSIONS: No generally accepted grounds for end-of-life decisions could be detected among Nordic intensivists. In addition, the current concept of management after decision to withdraw therapy varies markedly. This study has implications in further assessment of the individual decision-making process and the uniformity of actions after withdrawal decisions.