Millions of older adults receive care in intensive care units (ICUs) annually. However, the quality and accessibility of ICU-based palliative care is highly variable across hospitals and clinicians, due in part to specialists' limited workforce and geographic inconsistency. To address these gaps, the investigators developed an innovative mobile app-based primary palliative care intervention called ICUconnect. ICUconnect facilitates families' and patients' self-report of actual palliative care needs across all core domains of palliative care quality, provides ICU clinicians with a scalable digital infrastructure for coordinating consistent and personalized needs-targeted care, and provides a variety of informational supports relevant to each user's role. In this RCT, the investigators will test ICUconnect vs. usual care control among 350 patient-family member dyads with elevated baseline levels of unmet palliative care need in a 4-site network serving a diverse population (Duke, Medical University of South Carolina, University of Alabama at Birmingham, Columbia).

The specific aims are to: (1) Test the efficacy of ICUconnect vs. usual care control in improving palliative care needs and other person-centered outcomes including psychological distress, (2) Determine participant characteristics associated with a greater treatment response using a heterogeneity of treatment effects approach, and (3) Ensure off-the-shelf intervention readiness for implementation using a mixed-methods integration of qualitative analysis of semi-structured trial participant interviews and quantitative RE-AIM implementation framework-informed trial data.

Millions of older adults receive care in intensive care units (ICUs) annually. However, the quality and accessibility of ICU-based palliative care is highly variable across hospitals and clinicians, due in part to specialists' limited workforce and geographic inconsistency. Furthermore, there are few evidence-based interventions designed to help ICU clinicians provide primary palliative care themselves-and even fewer interventions tested among participants who adequately reflect the racial and ethnic diversity of the US.

To address these gaps, the investigators developed an innovative mobile app-based primary palliative care intervention called ICUconnect. ICUconnect facilitates families' and patients' self-report of actual palliative care needs across all core domains of palliative care quality, provides ICU clinicians with a scalable digital infrastructure for coordinating consistent and personalized needs-targeted care, and provides a variety of informational supports relevant to each user's role. In a recent single-center cluster RCT with 111 patient-family member dyads (U54 MD012530), 42% of whom were Black, ICUconnect significantly improved unmet needs and goal concordance of care compared to usual care control. The investigators have since linguistically and culturally adapted the intervention to Latin American Spanish.

What is now needed is a multicenter RCT to test the intervention's efficacy in a population diverse in race, ethnicity, and geography that reflects the real-world experience of patients and family members. Therefore, the investigators will test ICUconnect vs. usual care control among 350 patient-family member dyads with elevated baseline levels of unmet palliative care need in a 4-site network serving a diverse population (Duke, Medical University of South Carolina, University of Alabama at Birmingham, Columbia).

The specific aims are to: (1) Test the efficacy of ICUconnect vs. usual care control in improving palliative care needs and other person-centered outcomes including psychological distress, (2) Determine participant characteristics associated with a greater treatment response using a heterogeneity of treatment effects approach, and (3) Ensure off-the-shelf intervention readiness for implementation using a mixed-methods integration of qualitative analysis of semi-structured trial participant interviews and quantitative RE-AIM implementation framework-informed trial data.

Inclusion Criteria:

Patients (who are not interviewed but whose characteristics prompt enrollment of family members)

1. Adult aged ≥18 years
2. Managed in an adult medical, cardiac, trauma, surgical, or neurological ICU
3. Serious acute illness associated with a need for invasive mechanical ventilation
4. ICU team expect patient to require mechanical ventilation for 2 or more days

Family members

1. Adult aged ≥18 years
2. Family member: self-described as the individual (related or unrelated) who provides the most support and with whom the eligible patient has a significant relationship (Society of Critical Care Medicine definition of family; Davidson J, et al. Crit Care Med, 45:103-128; 2017)

ICU clinicians

1. Adult aged ≥18 years
2. ICU clinician: ICU attending caring for the eligible patient on the day of family member informed consent

Exclusion Criteria:

Patients

1. Death or full comfort care plan expected within 24 hours by ICU attending or fellow physician Rationale: measurable intervention effect unlikely

2. Palliative care consultative team and/or palliative care specialists are currently involved in the patient's care.

   Rationale: this would dilute the intervention's effect.

3. ICU length of stay >4 days during current ICU admission. Rationale: this would dilute intervention effect by widening the timeframe of family-clinician interactions and limiting standardization.

6. Imprisoned person. Rationale: this is a vulnerable population.

Family members

1. Low palliative care need burden (i.e., NEST scale score <15) at baseline. Rationale: if there are very low baseline needs, the intervention cannot impact the primary outcome measure (i.e., NEST).

2. Lack of English or Spanish fluency, operationalized as need for a translator to understand medical forms or participate in medical discussions.

   Rationale: the app has not been translated into other languages; also, many outcome measures aren't validated in languages other than English and Spanish.

3. Endorsement of suicidal ideation (i.e., PHQ-9 suicidality item) at the time of baseline data collection.

   Rationale: this indicates a serious level of distress that will require professional assistance beyond that provided by the intervention (or control).

4. Patient regains decisional capacity after informed consent but before the first family meeting. Rationale: as in our past ICU-based studies, this circumstance (the entry of the patient as a decision maker), would substantially change the nature of the interactions of family members and clinicians.

5. The ICU clinician changes from a consented clinician to a non-consented clinician before the first family meeting. Rationale: in this circumstance the intervention cannot be fully deployed.

ICU clinicians None

Per NIH regulations

Per NIH regulations

To be determined

All outcomes are self-reported by participants directly via the intervention apps and so not seen by the study team. The site PIs are blinded to study group.