Conversational Agents to Improve Quality of Life in Palliative Care (ECA-PAL)

Boston Medical Center (BMC)

Status: Completed

Conditions: Palliative Care

Treatments: Behavioral: Embodied Conversational Agent (ECA)

Study type: Interventional

Funder types: Other; NIH

Identifiers

NCT02750865

1R01NR016131 (U.S. NIH Grant/Contract)

H-34877

Details and patient eligibility

About

In this study the investigators will advance research on the development of easy to use technologies to empower patients. This is a scalable approach that has a significant potential to reduce suffering for palliative care patients and their caregivers. The investigators will adapt existing tested empathic conversational agents (ECA) for home-based cancer care management and inpatient bedside counseling to provide the following functions:

1. medication counseling;
2. physical activity promotion;
3. symptom management and continual screening for adverse events; and
4. alleviation of stress and anxiety
5. spiritual needs assessment; and
6. advanced care planning.

Data from the system will be monitored by a health professional, who communicates with members of the patient's care team. In this project, this activity will include facilitating referral for palliative care services.

The investigators will conduct pilot studies both at Boston Medical Center and at Northeastern University that will test system usability (Northeastern University and BMC) as well as interview burden (BMC only).

The investigators will conduct a Randomized Control Trial (RCT) to evaluate the agent technology, comparing usual care (UC) versus usual care plus the agent (UC+ECA) for patients 21 or older, with a life expectancy of < 1 year, from outpatient clinics at Boston Medical Center (BMC). Each subject will be enrolled along with a caregiver surrogate subject. In addition to baseline data collection, there will be monthly phone surveys for six months; intervention subjects will use the system for a six-month period of time. In addition, the investigators will conduct a randomized trial within the intervention group to compare subjects getting functions 1-4 (above) versus subjects getting the augmented intervention with all six functions. The intervention will include a nurse management interface to monitor clinical alerts generated by the system to stimulate interventions by the clinical staff.

Full description

Pilot Study The pilot study will determine usability of the ECA tablet system. Healthy volunteers will be recruited at Northeastern University, where the system is being developed. Subjects will spend approximately 60 minutes with the tablet system and provide usability feedback. Twelve healthy patient volunteers will be recruited to review the system as well as provide feedback on the relative interview burden (was the interview too long, too invasive, etc). No identifiable data will be collected from pilot study subjects.

Main Study The study will consist of adults with advanced illness (health care provider endorses the existence of a potentially life-limiting condition and endorses that "it would not be a surprise if the patient died within 12 months") who will be randomized to intervention or usual care control groups. Each person with an advanced illness will be enrolled along with a caregiver (i.e., family, friend) who will serve as a surrogate subject. The caregiver will be consented at baseline and interviewed at that point and at the final, 6-month study visit. In addition, if, at any point during their enrollment in the study, the patient-subject becomes unable, or decisionally impaired, to answer interview questions, the caregiver will be asked to answer on behalf of the subject.

The patient can be found to be decisionally impaired in any of 3 ways:

1. the patient declares themselves to be unable to answer questions appropriately
2. the patient is unable to pass a brief cognitive assessment that will be administered at the beginning of each telephone follow-up (per the Brief Screen for Cognitive Impairment)
3. the patient is physically unable to answer questions on the phone (due to, for example, serious health decline or death)

Baseline data will be collected after enrollment and before randomization. Baseline data includes:

1. sociodemographic data
2. Quality of life as measured by the European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire-Palliative Care
3. Patient Reported Outcomes Measurement Information System (PROMIS) Mental Health
4. Patient Reported Outcomes Measurement Information System (PROMIS) Social Health
5. Patient Activation Measure (PAM)
6. Patient Medication Adherence Questionnaire (PMAQ)
7. Spiritual Needs Assessment for Patients (SNAP)
8. Community Healthy Activity Model Program for Seniors (CHAMPS)
9. a survey of Advance Care Planning (ACP) communication and preferences
10. BSCI (This cognitive assessment is non-exclusionary, but, if the patient scores "0" on the word recall, the designated caregiver/surrogate will be asked to act as the research proxy and answer questions for the patient-subject.)

The investigators will also interview the surrogate subject at baseline regarding their level of caregiver stress with:

1. the CarerQol-7D
2. the caregiver's level of happiness (CarerQol-VAS)
3. self-rated burden (SRB) of care giving After the initial visit, study staff will administer 5 monthly phone surveys (months 1, 2, 3, 4, and 5), where the EORTC QLQ-C15-PAL will be collected. At the beginning of each data collection point study staff will administer the Brief Survey of Cognitive Impairment (BSCI) to determine if the patient is cognitively able to answer questions, and also inquire about unplanned health care utilization (i.e., urgent care, emergency room, hospital observation visits, and hospital admissions).

At the final, 6-month visit, the patient-subject will be interviewed in person, with the surrogate, and the full list of instruments will be administered. In addition, study staff will prepare a summary data report for the patient's clinical team. The report will include obtaining information from the Palliative Care service. The research staff will ask the subject which clinicians of theirs they would like to have receive this information.

Subjects randomized to the intervention will be given a mini-tablet computer configured for the purposes of this study. A research assistant will set up the tablet, enroll subjects into the study database, and train these subjects in the use of the system. Subjects who do not have existing WiFi at home will be provided with a cellular-based router. In such cases, study staff will arrange the WiFi service plan, install the router in the home, and confirm connectivity.

The investigators estimate that in many instances, when a clinician endorses that they would not be surprised if their patient died within 12 months, that this would be the first time they solidified a prognostic outlook in their minds for that patient. Stimulating clinicians to consider their patient's life-expectancy is itself a form of intervention, as this may instigate communication about prognosis with the patient, a re-evaluation of the treatment course, assessment of goals of care, and an improved level of attention to symptoms. The investigators will conduct informational sessions about the study and provide material about palliative care services in each outpatient clinical environment from which subjects will be enrolled. Information will be provided on how clinicians can contact the Palliative Care service to obtain consultation. Clinicians will be encouraged to call for assistance on palliative care needs for intervention subjects, usual care subjects, and non-subjects.

The same measures collected at baseline will also be collected at the final 6-month interview. At the end of study participation the investigators will also prepare a summary data report for the patient's clinical team. The report will include information on obtaining consultation from the Palliative Care service. The research nurse will ask the subject which clinicians of theirs they would like to have receive this information.

Enrollment: 367 patients

Sex: All

Ages: 21+ years old

Volunteers: No Healthy Volunteers

Inclusion criteria

1. 21 years of age or older
2. English speaking
3. Able to independently consent to be in the study
4. Has adequate corrected vision to be able to use the ECA system (based on a 1-minute ECA functional screener)
5. Has adequate hearing to be able to use the ECA system (headphones will be available)
6. Has a health care provider who endorses the existence of a potentially life-limiting condition and endorses that it would not be a surprise if the patient died within 12 months (PATIENT-SUBJECTS only)

Exclusion criteria

1. Enrolled in hospice
2. Already being followed by the Palliative Care team
3. Suicidal or homicidal
4. In police custody
5. Do not live in the Boston area
6. Plan on leaving the Boston area for more than 4 weeks in the next 6 months
7. Not able to use the ECA tablet computer

Trial design

Primary purpose: Supportive Care

Allocation: Randomized

Interventional model: Parallel Assignment

Masking: None (Open label)

---

367 participants in 2 patient groups

Control Usual Care

No Intervention group

Description:

In this arm the subject receives usual care and just completes the baseline interview, monthly telephone surveys about their quality of life, and the exit interview at 6 months after enrollment.

Embodied Conversational Agent (ECA)

Experimental group

Description:

In this arm the subject is trained how to use a tablet device which they take home for the duration of the study. These subjects complete the baseline interview, monthly telephone surveys about their quality of life, and the exit interview at 6 months after enrollment.

Treatment:

Behavioral: Embodied Conversational Agent (ECA)