A Pilot Trial of Decision Aids to Give Truthful Prognostic and Treatment Information to Chemotherapy Patients with Advanced Cancer

Patients with incurable disease state that they want truthful information about their diagnosis, treatment options, and course even if the outlook is poor;¹ but most patients never receive information from their physicians about prognosis² or even imminent death.³ U.S. physicians do not disclose prognosis at least half the time and feel unprepared to have these discussions.⁴ Not having a discussion about imminent death is associated with worse quality of care, worse quality of life, worse caregiver quality of life,⁵ and over $1,000 more in medical care cost in the last week of life.⁶ Physicians are reluctant to give people poor prognostic information⁷ for fear of dashing hope,⁸ and Web sites such as www.cancer.gov do not contain detailed information about prognosis, survival, palliative care options, or hospice referrals.

We designed decision aids for patients with incurable cancer and attempted to determine if people would opt for full disclosure about prognosis and treatment. If they opted for full disclosure, we assessed current knowledge about chance of cure, survival, disease response rates, and symptom control, before and after. This pilot trial was done to see if patients would complete a decision aid about their advanced cancer, even if it contained truthful information about their limited prognosis and treatment benefits.

Methods

We created state-of-the-art tables of information for patients with advanced breast, lung, colon, and hormone-refractory prostate cancers, based on expert review, external review, and comparison with Up To Date^© (available from the authors). The information was approved by all three oncologists involved. We used bar graphs to illustrate benefit, developed for patient education graphs for a randomized study of insurance types and treatment choices⁹ and in common use on the Web site Adjuvant Online (www.adjuvantonline.org).^{[10] and [11]} It is similar to what we do with the written medical record, a concise review of diagnosis, prognosis, treatment options, side effects, and when to call the doctor.¹²

We tested the intervention in a heterogeneous sample of 27 patients recruited through the Dalton Oncology Clinic, which serves a mix of patients from the most discerning third-opinion clinical trial patient to the community cancer patient and provides most of the indigent cancer care in the central Virginia area. The study was done within 3 months in early 2009.

Our primary outcome was the number of patients who would opt for full disclosure once they viewed the decision aid. Our secondary outcomes included the following: the amount of information patients have about cure, response rates, and symptom control; the impact of truthful information on hope, as measured by the Herth Hope Index^©13 (HHI) used to assess hope in clinical studies of adults;¹⁴ whether the information was deemed helpful to the patient; and whether the patient intended to share the information with a doctor.

Patients were accrued by reviewing the daily clinic list to find patients on treatment for incurable breast, colorectal, non-small-cell lung, or hormone-refractory prostate cancer. Treating oncologists were made aware of the study through e-mail, announcements, the Massey Cancer Center Web site, and individual meetings. All oncologists approached agreed to their chemotherapy patients participating in the study in general, and the primary nurse or treating oncologist was contacted about each eligible patient. Eligible patients were not contacted about the study when the treating oncologist or primary oncology nurse determined that a patient was experiencing significant distress or had significant psychiatric problems or difficulty with adjustment to illness or believed the patient would have great emotional difficulty with the information. The number of patients excluded by each oncologist due to concern about distress was estimated to be less than 10% of the total available but was not measured. Since these patients were not enrolled in the study, we did not collect information about them. A clinical psychologist and a chaplain were available to any patient who experienced distress during or after the interview process. The interview questions and intervention were administered by a member of the study team who was not the patient's oncologist or involved in his or her care. The interview team included a graduate student who was also a minister and chaplain (E. A. V.), a medical student with special training in empathic communication (L. A. D.), and/or the principal investigator (T. J. S.); usually one interviewer was present (L. A. D. or E. A. V.).

The interview sequence included screening questions to ensure that the patient wanted full information, sociodemographic questions, a pretest about the chance of cure and treatment effect for a patient with their illness, and the HHI. Next, the decision aid was administered. Immediately afterward, the patient completed a posttest, the HHI, and information about how he or she would use the information.

We modeled this approach on the Ottawa Decision Support Framework, a clinically tested decision-making tool designed to inform decisional conflict,^{[15] and [16]} defined as uncertainty about which course of action to take when the choice involves balancing gain, risk, loss, regrets, or challenges to personal life.¹⁷

Our study was approved by the Massey Cancer Center Protocol Review and Monitoring System and the VCU Institutional Review Board for the Conduct of Human Research. Because it was not a clinical trial, no clinical trial registration was required.

Results

The patients were typical for our urban, tertiary referral, and safety net hospital and National Cancer Institute–designated cancer center, as shown in Table 1.