Background:Most people in contemporary western society die of the chronic diseases of old age. Whilst palliative care is appropriate for elderly patients with chronic, non-malignant disease, few of the these patients access such care compared with cancer patients. Objective referral criteria based on accurate estimation of survival may facilitate more timely referral of non-cancer patients most appropriate for specialist palliative care.Objective:To identify tools and predictor variables that might aid clinicians estimate survival and assess palliative status in non-cancer patients aged 65 years and older.Methods:Systematic review and quality assessment using criteria modified from the literature.Results:11 studies that evaluated prognoses in hospitalised and community-based older adults with non-malignant disease were identified. Key generic predictors of survival were increased dependency of activities of daily living, presence of comorbodities, poor nutritional status and weight loss, and abnormal vital signs and laboratory values. Disease-specific predictors of survival were identified for dementia, chronic obstructive pulmonary disorder and congestive heart failure. No study evaluated the relationship between survival and palliative status.Conclusion:Prognostic models that attempt to estimate survival of ?6 months in non-cancer patients have generally poor discrimination, reflecting the unpredicatble nature of most non-malignant disease. However, a number of generic and disease-specific predictor variables were identified that may help clinicans identify older, non-cancer patients with poor prognoses and palliative care needs. Simple, well-validated prognostic models that provide clinicians with objective measures of palliative status in non-cancer patients are needed. Additionally, research that evaluates the effect of general and specialist palliative are on psychosocial outcomes in non-cancer patients and their carers is needed.