Development and validation of the SURVIPALL score: a simple tool to predict 3-month survival and support timely referral to palliative care in oncology patients

Estimating short-term survival in advanced cancer is challenging, yet timely referral to palliative care (PC) is essential for symptom management, decision-making, and care planning. Many existing prognostic tools rely on subjective assessments or are validated only for very short survival windows. This study aimed to identify factors associated with survival shorter than 3 months in patients with advanced cancer referred to PC and to develop a simple prognostic tool. A retrospective cohort study was conducted including adult oncological patients referred from the Oncology Department to an outpatient palliative care unit in 2022. Patients were classified according to whether survival was shorter or longer than three months after referral. Univariate analyses were performed in the overall cohort, and multivariable logistic regression was conducted in patients with complete data. Model discrimination was assessed using receiver operating characteristic curve analysis. Temporal validation was performed in an independent cohort. Two hundred and forty-six patients were included, of whom 132 (53.7%) died within three months. Albumin measurements were available for 181 patients (73.6%), included in the univariate and multivariable analyses. Multivariable logistic regression identified anorexia, albumin < 3.5 g/dL, presence of metastases and neutrophil-to-lymphocyte ratio > 3 as independent predictors of death; ECOG performance status ≥ 3 was additionally retained in the final model based on clinical relevance. These five variables were incorporated into the SURVIPALL scale, assigning one point to each variable except neutrophil-to-lymphocyte ratio > 3, which was assigned two points. The model achieved an area under the curve (AUC) of 0.807. A score ≥ 3 yielded a sensitivity of 94.4% and a negative predictive value of 87.5% for predicting survival shorter than three months. Temporal validation in the 2023 cohort ( n = 187) demonstrated similar performance (AUC 0.800; sensitivity 93.9%; negative predictive value 85%). The SURVIPALL scale is a simple and rapidly applicable prognostic tool based on routinely available variables. A score ≥ 3 accurately identifies patients with an estimated survival shorter than 3 months, supporting its potential use to optimise referral timing, guide resource allocation, and enhance communication with patients and families. Further studies should explore its applicability across diverse care settings.