The quality of intensive care unit (ICU)-based palliative care is highly variable, particularly for the 2 million older adults admitted annually to ICUs. To address these care delivery barriers among older ICU patients, a mobile app platform called PCplanner (Palliative Care planner) was developed. PCplanner automates the identification of high-risk patients (e.g., dementia, declining health status, poor functioning) by directly capturing data from electronic health record (EHR) systems, cultivates family engagement with supportive information and a digital system for self-report of actual needs, and facilitates the delivery of care to those with a high burden of need by coordinating collaboration between ICU teams and palliative care specialists.

150 patients, 150 family caregivers, and 75 physicians from academic and community settings will be enrolled in a RCT designed to test the efficacy of PCplanner-augmented collaborative palliative care vs usual care. Family caregiver and clinician experiences will be explored using mixed methods to understand intervention mechanisms as well as implementation barriers within diverse case contexts. The key hypothesis is that compared to usual care, PCplanner will reduce family caregivers' unmet needs and psychological distress, increase the frequency of goal concordant treatment among older adult patients, and reduce hospital length of stay.

The quality of intensive care unit (ICU)-based palliative care is highly variable, particularly for the 2 million older adults admitted annually to ICUs. However, improving care quality on a broad scale with the efficient delivery of patient-centered, need-targeted palliative care is challenging because of logistical and technological barriers. To address these care delivery barriers among older ICU patients, a mobile app platform called PCplanner (Palliative Care planner) was developed. PCplanner automates the identification of high-risk patients (e.g., dementia, declining health status, poor functioning) by directly capturing data from electronic health record (EHR) systems, cultivates family engagement with supportive information and a digital system for self-report of actual needs, and facilitates the delivery of care to those with a high burden of need by coordinating collaboration between ICU teams and palliative care specialists. In pilot comparison to a standard palliative care control, the intervention reduced unmet needs, psychological distress, and length of stay and increased goal concordant care, communication, and hospice utilization.

While these data are compelling, an efficacy evaluation of PCplanner is needed. Therefore, we plan to enroll approximately 150 patients, 150 family caregivers, and 75 physicians will be enrolled from academic and community settings in a project with 2 key aims: (1) Test the efficacy of PCplanner-augmented collaborative palliative care vs usual care in a randomized clinical trial (RCT) with 3-month follow up, and (2) Explore family caregiver and clinician experiences using mixed methods to understand intervention mechanisms as well as implementation barriers within diverse case contexts. The key hypothesis is that compared to usual care, PCplanner will reduce family caregivers' unmet needs and psychological distress, increase the frequency of goal concordant treatment among older adult patients, and reduce hospital length of stay.

PATIENTS

Inclusion Criteria:

• ≥50 years of age
• Receive care in a study ICU for ≥24 hours
• Meets ≥1 of 9 high risk phenotypes a. Dementia (e.g., Alzheimer's, multi-infarct, other dementia etiology) b. Declining health status defined by EITHER: i. ≥2 hospital admissions in 3 months preceding current admission OR ii. >1 ICU admission in 3 months preceding current admission c. Poor functional status defined by EITHER: i. admit from Skilled Nursing Facility (SNF) or Long-Term Acute Care (LTAC) facility OR ii. ≥3 activities of daily living (ADL) limitations at admission d. Severe acute illness defined by EITHER: i. cardiac arrest OR ii. multisystem organ failure (≥3 of: lung, kidney, hematological, brain, cardiac, liver) that has worsened over 48 hours (i.e., Sequential Organ Failure Assessment [SOFA] score increase) e. Severe acute stroke (e.g., acute intracranial hemorrhage, ischemic stroke, or traumatic brain injury) f. Acute respiratory failure (ventilation or high oxygen support for ≥24 hours) g. Acute renal failure (new hemodialysis or continuous venovenous hemodiafiltration for ≥1 hour) h. Advanced cancer (Advanced / metastatic cancer diagnosis) i. Shock (use of vasopressor or inotrope for ≥4 hours)

Exclusion Criteria (pre-consent):

• Palliative care consultation performed during the hospitalization before eligibility determination
• Current admission to ICU at the index hospital ≥8 days
• Imprisoned
• No known family or surrogate decision maker
• Death expected within 24 hours

Exclusion Criteria (post-consent):

- Patient dies before T2

FAMILY MEMBER

Inclusion Criteria:

• ≥18 years of age
• Self-described as the individual (related or unrelated) who provides the most support and with whom the patient has a significant relationship (per definition of 'family' described in the Society of Critical Care Medicine 2016 Guidelines for Family-Centered Care in the Neonatal, Pediatric, and Adult ICU)

Exclusion Criteria (pre-consent):

• Lack a knowledge of English such that the potential participant is not confident that they could complete study tasks (app viewing, surveys)
• Imprisoned
• Unable to complete surveys for any reason

Exclusion Criteria (post-consent):

- Low need burden (NEST score <10) at baseline

ICU PHYSICIANS

Inclusion Criteria:

• ≥18 years of age
• Attending or fellow physician in a study ICU

Exclusion Criteria:

- None

Per NIH and Duke University policies after trial and analyses completed.

2025

review by study team

ONLY FAMILY MEMBERS have survey outcomes reported (i.e., NEST). Physicians and patients DID NOT complete surveys.

The ONLY patient-level data reported are for demographics, length of stay, and AEs.

3 months

Adverse event reporting by family members and patients are as follows. Numbers reported for PHQ-9 item 9 (suicidality) reflect FAMILY MEMBERS. Numbers reported for deaths reflect PATIENTS.

Deaths reported reflect patient, not family member, events.

All masking done via differential user accesses allowed by a mobile app platform.