Occurrence and timing of advanced care discussions in recurrent ovarian cancer patients participating in clinical trials remain to be optimized

Abstract

<h3>Objectives:</h3> In women with recurrent ovarian cancer (OC), advanced care planning (ACP) such as advanced directives (AD), code status, and timely hospice referral should be addressed. In a high-volume, clinical trial focused cancer center, treatment with novel potentially life-prolonging therapies may alter timing of discussions. Our study compares patterns of ACP between trial and non-trial recurrent OC patients. <h3>Methods:</h3> All patients ≥18 years who were treated at a single institution for the diagnosis of OC during the year of 2015 and had ever recurred were reviewed. Patients who ever (n=84) versus never (n=41) participated in a therapeutic clinical trial (CT) were compared. Chi-square or Fisher's exact tests and 2 sided t-tests or Wilcoxon Rank-Sum tests compared demographic data and ACP variables of interest using an α = 0.05. Multivariable logistic regression estimated adjusted odds ratios (aOR) adjusted by CT participation, age, and Charlson comorbidity index. <h3>Results:</h3> A total of 125 patients were identified, and 84 (67%) participated in CTs. Cohorts were similar in age, BMI, insurance status, and histopathologic characteristics. Median time to follow up after first recurrence was 856 days in trial patients vs 308 days in non-trial patients (p<0.0001). Caucasian patients comprised 95% of the CT cohort compared to 80% of non-trial patients (p=0.0205). Unadjusted analyses showed CT participants more frequently discussed AD (36% vs 17% non-trial patients, p=0.0321). Rates medical power of attorney (MPOA) discussion (54% in both), code status discussion (43% vs 27% in non-trial patients), and palliative care referral (49% vs 37% in non-trial patients) did not significantly differ. Median time between first recurrence and code status discussion was significantly longer in CT versus non-trial participants (731 days (IQR 102-1376) versus 57 days ((IQR 38-565), p=.0379). Of the 81 deceased patients, though rates of hospice enrollment were similar (74% CT versus 71% non-trial), CT patients more frequently died in the hospital (22.8% vs 8.7%) or a care facility (14% vs 0%) (p=0.0311). ACP discussions and palliative care referrals tended to occur either during or after the trial (Figure 1). In adjusted analyses, palliative care referral remained the only significant predictor of code status discussions (aOR 3.69, 95% CI 1.67-8.152), AD discussion (aOR 14.46, 95% CI 5.04-41.49), and MPOA discussion (aOR 8.71, 95% CI 3.66-20.73). Participation in neither late phase nor phase 1 trials significantly predicted the odds of ACP. <h3>Conclusions:</h3> The time between the recurrence and code status discussions was significantly longer in CT participants, and ACP discussions occurred typically during or after a trial. ACP occurred more frequently when patients were referred to palliative or supportive care referral, independent of participation in either late phase or phase 1 CT. Prioritizing ACP and supportive care referral, especially in CT participants, may improve these rates and optimize end of life care.