1592MO - Anticancer therapy at the end-of-life: A cluster-randomized trial

Background

Early palliative care (PC) and regular use of patient self-reported symptoms at End-of-Life (EoL) have shown beneficial results on diverse quality of life (QoL) outcomes and survival. Effect on anticancer therapy (ACT) use is rarely investigated. The primary outcome in this cluster-randomized trial was use of ACT at EoL, i.e., last 3 months of life following a complex intervention in one arm (INT) versus conventional care in the other (CTR).

Methods

Hospitals were stratified on size prior to randomization. Patient eligibility was age ≥18, advanced stage solid tumours, starting last ACT line, estimated life expectancy ≤12 months. The intervention consisted of a physician educational program, compulsory PC referrals at inclusion, and patient symptom reporting (paper/digitally) prior to all consultations. Clinical data were recorded, and all patients completed postal QoL forms bimonthly. Follow-up was 1 year/till death, whichever came first.

Results

616 (93%) patients were eligible (INT:309/CTR:307); 63% male, median age 69 (IQR 62-75), 71% with Karnofsky score (KPS) 80-100. 77% had gastrointestinal (GI) cancers. Median survival time from inclusion was 8 (IQR 3-14) and 7 months (IQR 3-12) (INT/CTR). Median number of days between ACT start and death was 188 (OQR 77-370). Median number of days between diagnosis, ACT start or stop, and death did not differ by group. 177 (28.7 %) patients received ACT in the last 3 months, 75 (12.1%) in the last month. Adjusted model for the time from ACT cessation to death was longer in INT (p=0.25; HR=0.81 95% CI 0.57-1.15) and in patients with high KPS at inclusion (p<0.001). No QoL differences across groups was found for those with complete measures at 2, 4 and 6 months (n=215). Physical function was significantly lower at 6 months, while insomnia improved (p <.05).

Conclusions

The intervention did not have the desired effect. This may be due to a relatively low national use of ACT at EoL in Norway. A lower proportion of GI cancers had been desirable. Additionally, the intervention was too general and maybe too focused on communication skills to exert a substantial influence on conventional clinical practice. We recommend close, continuous monitoring to secure adherence to study procedures.

Clinical trial identification

NCT01362816.

Editorial acknowledgement

Legal entity responsible for the study

M. J. Hjermstad, N. Aass, S. Kaasa.

Funding

Helseforsk through the Norwegian Research Council.

Disclosure

All authors have declared no conflicts of interest.