Cluster-randomized trial of early palliative care for patients with metastatic cancer

Reporter: Gita Suneja, MD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 4, 2012

Presenter: Camilla Zimmermann
Affiliation: Princess Margaret Hospital, University Health Network, Toronto, ON, Canada


  • Patients with metastatic cancer have compromised quality of life (QOL), which tends to worsen towards the end of life.
  • Several recent studies have demonstrated an improvement of quality of life with palliative care intervention.
  • Bakitas, et al demonstrated higher QOL and mood with nurse-led palliative care-focused interventions, but no improvement in symptom intensity scores (Bakitas, JAMA, 2009).
  • Temel, et al demonstrated that patients with metastatic non-small cell lung cancer receiving early palliative intervention had improved QOL, mood, and survival.
  • After a pilot study testing the feasibility of the trial in 150 patients (Follwell, JCO, 2009), the presenting authors conducted a cluster-randomized trial of early versus routine palliative care in patients with metastatic cancer.
  • The primary objective was to assess the impact of intervention on QOL, symptom control and satisfaction with care.


  • Twenty-four medical oncology clinics were stratified by tumor site (4 lung clinics, 8 gastrointestinal, 4 genitourinary, 6 breast, 2 gynecological), then randomized to intervention by a palliative care team (physician and nurse) with monthly follow-up, or to routine cancer care.
  • Early intervention was defined as intervention within 4 weeks of diagnosis.
  • Eligible patients were > 18 years of age, had stage IV cancer (or stage III for poor prognosis cancers), ECOG performance status 0-2 and a clinical prognosis of 6 months to 2 years.
  • Patients completed the following measures of QOL at baseline and monthly for 4 months:
    • FACIT-Sp, including physical, social, emotional, functional and spiritual well-being; range 0-156, with higher scores indicating better QOL
    • Edmonton Symptom Assessment System, including symptoms severity; range 0-90, with higher scores indicating worse symptom severity
    • FAMCARE-P16, including satisfaction with care; range 16-80
  • The primary outcome was the change in FACIT-Sp at 3 months.
  • Secondary outcomes included symptom severity and satisfaction
  • The planned sample size was 225 patients per arm, assuming 80% power and a 2-sided significance level of 0.05.
  • The control arm received palliative care intervention upon request only.
  • The unit of analysis was the individual rather than the cluster. Intention to treat analysis was conducted with adjustment for cancer site, age, chemotherapy administration, and ECOG performance status.
  • The impact of missing data was assessed with sensitivity analysis.


  • From December 2006 to September 2010, 461 patients completed baseline measures (228 intervention, 233 control), and 442 patients completed at least one follow-up assessment.
  • The clusters were well-balanced between the intervention and control groups for demographics (age, sex, income, education, employment) and medical characteristics (ECOG, chemotherapy).
  • However, the control group had significantly better quality of life prior to randomization due to selection bias associated with palliative intervention.
  • At 3 months, patients in the intervention group had marginally improved QOL (mean change in intervention vs. control, 1.6±14.5 vs. -2.0±13.6, p=0.07) but not symptom severity (0.1±16.9 vs. 2.1±13.9, p=0.34).
  • At 4 months the change in QOL was more marked (2.5±15.5 vs. -4.0±14.2, p=0.008) and symptom severity was marginally better (-1.3±16.0 vs. 3.2±13.9, p=0.05).
  • Improvement in satisfaction with care was evident at one month (p=0.001) and remained significant at both 3 months (2.3±9.1 vs.-1.8±8.2, p=0.001) and 4 months (3.7±8.6 vs. -2.4±8.3, p<0.001).

Author's Conclusions

  • In patients with metastatic cancer, early palliative care intervention immediately improved satisfaction with care, while QOL and symptom control improved later.
  • Early referral improved satisfaction and prevented usual decline in QOL.


  • The authors present a follow-up study to an initial phase II pilot study examining the impact of early palliative care intervention on quality of life, symptoms severity, and patient satisfaction for patients with advanced cancer.
  • Their pilot data demonstrated improvement in symptom severity and satisfaction at both 1 week and 1 month. The present study better defined quality of life measures and extended the follow-up time to 4 months.
  • A cluster-randomized design is a trial in which individuals are randomized in groups (i.e. the group is randomized, not the individual). This design was chosen to enhance recruitment, as other trials in palliative care often have difficulty with recruitment, particularly early in the study period.
  • Benefits of the cluster-randomized design include intervention on a group level, avoidance of contamination of control and cross over groups, and enhanced recruitment. Disadvantages are the larger sample size required, and the need to account for clustering effects and selection bias.
  • Of note, the selection bias in this study was such that patients in the intervention arm had worse QOL and symptom severity at the onset of study, and this favored the control arm. Results would have been stronger without the selection bias.
  • Limitations of this study include that it was performed at a single institution tertiary care center with a well-established palliative care team. The results may not be reproducible at a smaller center with less palliative care resources.
  • Furthermore, the intervention and control groups were not blinded, and knowledge of receiving intervention may have had a placebo effect in terms of symptom reporting and patient satisfaction.
  • Nonetheless, the study is one a very few phase III clinical trials studying the impact of palliative care on patient quality of life.
  • The strong evidence from studies prior to this suggest that patients have improved quality of life, quality of care, and no evidence of harm from early palliative care interventions.
  • Furthermore, numerous studies have found palliative care services to be affordable, and in many cases, cost-saving interventions. Data from Kaiser Permanente showed savings of $5-7000 per patient, as well as improved communication, and equivalent survival for patients receiving early palliative care interventions. A similar study by Aetna showed reduction in inpatient and intensive care unit days, with savings of 22% in the last 40 days of life with palliative care interventions.
  • These findings of these prior studies have led ASCO to issue Provisional Clinical Opinion recommending integration of palliative care service into standard oncology care for all patients with metastatic non-small cell lung cancer at the time of initial diagnosis.
  • Further studies are necessary to address the reluctance of early referral to palliative care, such that solutions can be devised and more patients can have access to early palliative care intervention.