Prediction of survival in a hospital-based continuing care unit

Prediction of survival can be relevant in palliative care in those units with selective admission policies and limited resources, for planning patient management and in discharge planning for those patients expected to go home. In this study, factors most predictive of prognosis were identified. Those factors shown to have no effect on survival included the performance of investigations or procedures, anti-cancer therapy, morphine dose on admission and original admitting ward. Patients admitted primarily for pain control had a significant survival advantage over those patients admitted for palliation of some other symptom. Actual survival correlated well with predicted outcome. Factors most predictive of relative risk of death in a multivariate analysis were dyspnoea, decubitus ulcers, predicted outcome, interventions and a diagnosis of lung cancer. When symptoms alone were analysed, dyspnoea and immobility carried the highest relative risk of death.