Admission-day predictors of terminal outcome in advanced cancer inpatients: Insights from a large Indian palliative care cohort.

A Allwin Jesuraj (All India Institute of Medical Sciences (AIIMS), New Delhi, India) S Seema Mishra R Rakesh Garg N Nishkarsh Gupta S Sachidanand Jee Bharati (All India Institute of Medical Sciences (AIIMS), New Delhi, India) B Brajesh Kumar Ratre (All India Institute of Medical Sciences (AIIMS), New Delhi, India) A Anuja Pandit S Saurabh Vig

Abstract

339 Background: Inpatient mortality in palliative care units ranges 10-30% globally, yet prognostic data from low- and middle-income countries (LMICs) remain limited. Whether admission-day clinical and laboratory variables can identify high-risk patients in resource-limited settings is unclear. Methods: Retrospective analysis of a prospectively maintained database from two inpatient palliative care wards at a tertiary cancer centre in North India. All consecutive index admissions (January-December 2025) were included. Primary outcome: terminal event (inpatient death or discharge for end-of-life home-based care). Predictors: ECOG performance status, admission reason, number of complaints, serum albumin, haemoglobin, estimated glomerular filtration rate, total leukocyte count, age, sex, weekend admission, cancer type, distance. Multivariable logistic regression with stepwise selection identified independent predictors. Sensitivity analysis examined mortality alone. Results: Among 1,458 patients (mean age 52.3±14.1 years, 51% male), terminal events occurred in 275 (18.9%) and deaths in 182 (12.5%). Multivariable analysis retained 6 independent predictors (Table). ECOG PS 4 showed strongest association (OR 42.3). Symptom-based predictors included dyspnoea (OR 33.8), delirium (OR 23.3), and symptom complexity with ≥3 complaints (OR 17.1). Laboratory markers included severe renal impairment (OR 3.3) and albumin (OR 0.55 per g/dL, protective). Weekend admission was associated with increased risk (OR 2.5). Sensitivity analysis using mortality alone yielded consistent findings. Conclusions: Basic admission day variables can reliably forecast terminal outcomes among palliative care inpatients. ECOG PS, initial symptoms, albumin, and kidney function, all accessible at the bedside without advanced tests, help guide triage, care planning, and resource distribution in LMICs. The results encourage early prognosis discussions and care preparation. Independent admission-day predictors of terminal event in palliative care inpatients (n=1458). Predictor OR (95% CI) P-value ECOG PS 4 (vs 1-2) 42.3 (17.8 - 100.6) <0.001 Dyspnoea (vs pain) 33.8 (8.7 - 132.1) <0.001 Delirium (vs pain) 23.3 (5.9 - 92.6) <0.001 ≥3 complaints (vs 0-1) 17.1 (3.5 - 83.2) <0.001 Severe CKD (vs normal) 3.3 (1.8 - 6.1) <0.001 Weekend admission 2.5 (1.4 - 4.4) 0.002 Albumin (per g/dL) 0.55 (0.39 - 0.76) <0.001

Article Details

Volume / Issue Vol. 44, Issue 19_suppl
Published July 01, 2026
Pages 339-339
ISSN 0732-183X
Publisher Lippincott Williams & Wilkins

Journal Info

Journal of Clinical Oncology

Lippincott Williams & Wilkins

ISSN: 0732-183X Health Sciences

Authors (8)

A

Allwin Jesuraj

All India Institute of Medical Sciences (AIIMS), New Delhi, India

S

Seema Mishra

R

Rakesh Garg

N

Nishkarsh Gupta

S

Sachidanand Jee Bharati

All India Institute of Medical Sciences (AIIMS), New Delhi, India

B

Brajesh Kumar Ratre

All India Institute of Medical Sciences (AIIMS), New Delhi, India

A

Anuja Pandit

S

Saurabh Vig