Weekly symptom monitoring in lung cancer tracked with five months more survival

Design

multicentre stepped-wedge cluster-randomised trial, post-hoc long-term survival analysis, Cox mixed-effects model

Population

446 analysable patients with stage I–IV lung cancer across 14 Dutch centres, median follow-up 50 months

Primary outcome

overall survival (post hoc; the trial's original primary outcome was quality of life)

Effect

median overall survival 26 vs 21 months, adjusted hazard ratio 0.80 (95% CI 0.63–1.00), p=0.049

Patient-reported outcome monitoring has improved survival in resource-intensive trials of advanced cancer, and the open question was whether it does anything in ordinary practice across all stages. The SYMPRO-Lung trial was a stepped-wedge cluster-randomised trial across 14 Dutch centres: every hospital began in a control period and crossed over at a predefined point to an intervention period, in which patients with stage I–IV lung cancer completed weekly online symptom assessments for a year. Quality of life was the original primary endpoint; this is a post-hoc analysis of long-term survival.

Of 515 patients enrolled, 446 were analysable at a median follow-up of 50 months — 235 intervention, 211 control. Median overall survival was 26 months in the intervention group (95% CI 21–33) against 21 months (17–26), with an adjusted hazard ratio of 0.80 (0.63–1.00, p=0.049). Progression-free survival did not differ significantly: 0.86 (0.70–1.07, p=0.18).

Two qualifications matter, and neither cancels the finding. This is post hoc, and the p value of 0.049 sits exactly on the conventional threshold — a finding that would not survive much scrutiny on its own. But it points the same way as the earlier advanced-disease trials, in a broader population and a real-world setting, and the intervention is a weekly questionnaire. That is the calculation for a service: the evidence is not definitive, and the intervention is close to harmless and inexpensive. The difficult part is not the questionnaire but who reads the alerts it generates, and what they are empowered to do about them.

• Decide who monitors the responses and what their escalation authority is before starting any programme
• Include early-stage patients — this trial covered stage I to IV, unlike its predecessors
• Note progression-free survival did not differ, so the mechanism is not earlier detection of progression
• Plan for a year of weekly reporting, which is what was tested
• Consider language and literacy barriers to online symptom reporting in Indian practice before assuming uptake

Why it matters

The survival benefit appears to come from responding to symptoms sooner, not from detecting progression earlier.

Don't overread it

This is a post-hoc survival analysis of a trial whose primary endpoint was quality of life, with a p value of 0.049.

The statistics, in plain English

A hazard ratio of 0.80 with an upper confidence limit of exactly 1.00 and p=0.049 is the weakest form of a positive result — move one patient between groups and it becomes non-significant. The five-month difference in median survival is the more tangible figure, though medians are sensitive to where the curves happen to cross. That progression-free survival showed no significant difference while overall survival did is worth noting: it argues the benefit came from better symptom management rather than from catching progression sooner.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.