- Design
- systematic review and random-effects meta-analysis of 39 randomised controlled trials (2002-2024), Cochrane RoB 2 and Egger's test
- Population
- 8,399 adults with cancer; early palliative care defined as starting within 12 weeks of diagnosis or first-line treatment
- Primary outcome
- quality of life and overall survival
- Effect
- quality of life SMD 0.30 (95% CI 0.25-0.35, I² 15.4%); overall survival HR 0.84 (0.78-0.90, I² 25.4%), exploratory
A meta-analysis pooled 39 randomised trials from 2002 to 2024, 8,399 patients, comparing palliative care started within 12 weeks of diagnosis or of starting first-line treatment, alongside usual cancer care, with usual care alone. Most trials were at low risk of bias and most came from North America and Europe.
Quality of life improved with early palliative care: standardised mean difference 0.30 (95% CI 0.25-0.35), with low heterogeneity (I² 15.4%). The overall survival estimate, which the authors label exploratory, was a hazard ratio of 0.84 (0.78-0.90, I² 25.4%). Results were consistent across cancer types and regions.
The survival finding has been contested since the first lung cancer trial reported it, on the reasonable ground that palliative care is not a treatment for cancer. Thirty-nine trials later it has not gone away, and the mechanism most often proposed — fewer futile late-line treatments, fewer terminal admissions, better symptom control and earlier recognition of deterioration — is entirely plausible without the intervention being antineoplastic. The practical obstacle in Indian practice is not evidence but the word: patients and families hear palliative as terminal. Referring at diagnosis, routinely and for everyone with advanced disease, is what removes that signal — the referral means nothing about prognosis if everybody gets one.
- Refer at diagnosis of advanced disease, not at the point of treatment failure
- Make it automatic and universal so the referral carries no prognostic message
- Say 'supportive care' if the word palliative closes the conversation, but make the referral either way
- Twelve weeks from diagnosis or first-line start is the window these trials used
- Almost all trials came from North America and Europe; the survival estimate should not be quoted as established in Indian settings
Why it matters
The evidence for referring early is now large and consistent; what blocks it is the word, not the data.
Don't overread it
The survival benefit is an exploratory secondary analysis, and generalisability to low- and middle-income settings is untested.
The statistics, in plain English
A standardised mean difference of 0.30 is a small-to-moderate effect — real and consistent, but not transformative for an individual patient. The survival hazard ratio of 0.84 is labelled exploratory because survival was not the primary outcome most of these trials were designed around, so it is a finding to act on cautiously and not one to quote as a treatment effect. Low heterogeneity (I² 15.4%) means the trials agreed with each other, which strengthens both estimates.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for oncology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free