A randomised controlled trial of an advance care planning intervention for patients with incurable cancer

Abstract

We modified and evaluated an advance care planning (ACP) intervention, which had been shown to improve compliance with patient's end of life (EoL) wishes, in a different patient population. Patients with incurable cancer, and a Family Member (FM), were randomised one-to-one to usual care or usual care plus an ACP intervention, between April 2014 and January 2017. Oncologists and participants were non-blinded. ACP was based on the Respecting Patient Choices model, with an offer to provide individualised ranges for typical, best-case and worst-case scenarios for survival time. Seven facilitators (two oncology nurses, two nurses and three allied health professionals) delivered the intervention within 2 weeks of study enrolment. The primary outcome measure, assessed by interviewing the FM 3 months after patient death, was the FM perception that the patient's wishes were discussed, and met. Six hundred and sixty-five patients from seven Australian metropolitan oncology centres were referred for consideration by their oncologists, 444 (67%) met the study inclusion criteria and were approached by a study researcher. Two hundred and eight patients (47%) and their FM entered the trial as dyads. Fifty-three (46%) dyads in the ACP group and 63 (54%) dyads in the usual-care group had complete primary outcome data (p = 0.16). Seventy-nine patients and 53 FMs attended an ACP discussion. Mean length of discussion was 57 min. FMs from 23 (43%) dyads allocated to ACP and 21 (33%) dyads allocated usual care reported the patient's EoL wishes were discussed and met (difference 10%, 95% CI: -2 to 8, p = 0.27). There were no differences in EoL care received, patient satisfaction with care; FM satisfaction with care or with death; or FM well being. Rates of palliative care referral were high in both groups (97% vs 96%). A formal ACP intervention did not increase the likelihood that EoL care was consistent with patients' preferences.