Emulation of a Target Trial to Evaluate the Causal Effect of Palliative Care Consultation on the Survival Time of Patients with Hepatocellular Carcinoma

Abstract

Palliative care has the potential to improve the quality of life of patients with incurable diseases or cancer, such as hepatocellular carcinoma (HCC). A common misconception of palliative care with respect to the patient's survival remains a significant barrier to the discipline. This study aimed to provide causal evidence for the effect of palliative care consultation on the survival time after diagnosis among HCC patients. An emulation of a target trial was conducted on a retrospective cohort of HCC patients from January 2017 to August 2019. The primary endpoint was the restricted mean survival time (RMST) at 12 months after HCC diagnosis. We used the clone-censor-weight approach to account for potential immortal time bias. In this study, 86 patients with palliative care consultation and 71 patients without palliative care consultation were included. The adjusted RMST difference was -29.7 (95% confidence interval (CI): -81.7, 22.3; p-value = 0.263) days in favor of no palliative care consultation. However, palliative care consultation was associated with an increase in the prescription of symptom control medications, as well as a reduction in life-sustaining interventions and healthcare costs. Our findings suggest that palliative care consultation was associated with neither additional survival benefit nor harm in HCC patients. The misconception that it significantly accelerates the dying process should be disregarded.

Keywords: hepatocellular carcinoma; mean survival time; palliative care; supportive care.

Figure 1

Balance in prognostic factors at baseline for approach A (conventional cohort analysis) based on absolute standardized difference in the unweighted and weighted samples.

Figure 2

Balance in prognostic factors at 12 months after diagnosis for approach B (emulated cohort) based on absolute standardized difference in the unweighted and weighted clone samples.

Figure 3

Predicted survival curve based on flexible parametric survival regression for approach A and approach B (limiting the analysis at 1 year after HCC diagnosis).