Abstract Aim Early palliative care (EPC) in oncology has shown sparse evidence of a positive impact on patient outcomes, quality of care outcomes and costs. Patients and methods Data for this secondary analysis were taken from a trial of 207 outpatients with metastatic pancreatic cancer randomly assigned to receive standard cancer care plus on-demand EPC (standard arm) or standard cancer care plus systematic EPC (interventional arm). After 20 months' follow-up, 149 (80%) had died. Outcome measures were frequency, type and timing of chemotherapy administration, use of resources, place of death and overall survival. Results Some indices of end-of-life (EoL) aggressiveness had a favourable impact from systematic EPC. Interventional arm patients showed higher use of hospice services: a significantly longer median and mean period of hospice care ( P  = 0.025 for both indexes) and a significantly higher median and mean number of hospice admissions (both P  < 0.010). In the experimental arm, chemotherapy was performed in the last 30 days of life in a significantly inferior rate with respect to control arm: 18.7% versus 27.8% (adjusted P  = 0.036). Other non-significant differences were seen in favour of experimental arm. Conclusions Systematic EPC showed a significant impact on some indicators of EoL treatment aggressiveness. These data, reinforced by multiple non-significant differences in most of the other items, suggest that quality of care is improved by this approach. This study is registered on ClinicalTrials.gov ( NCT01996540 ).