Feasibility of a Home-Based Palliative Care Intervention for Elderly Multimorbid Survivors of Critical Illness

Each year, 2 million elderly Americans are treated in intensive care units (ICUs).¹  Long-term outcomes after critical illness vary, but one-third of patients who survive to hospital discharge die during the following year, and one-sixth have severe persistent impairments for which they require ongoing medical care. Most ICU survivors who are discharged home return to their primary care physicians and subspecialists for aftercare.

Alternative care models exist, such as ICU survivor clinics that provide integrated care at a single location focused on treating sequelae of critical illness.²  However, both of these care models are clinic-based and require patients to travel, which may be burdensome for some patients or unfeasible for patients who are debilitated or have poor social support. In addition, these aftercare models often do not include palliative care interventions. Home-based care is an attractive alternative after hospital discharge, a period when patients are vulnerable and caregivers feel unprepared and overwhelmed.³  The feasibility of implementing a home-based care model in critical illness survivors has not been studied. Therefore, we conducted a quality improvement trial to test the operational feasibility of a 90-day, home-based palliative care intervention in elderly veteran ICU survivors.

We assessed operational feasibility by conducting a prospective study of a multidisciplinary, home-based palliative care intervention in 15 patients who were discharged home after treatment in the ICU for community-acquired pneumonia, heart failure exacerbation, chronic obstructive pulmonary disease (COPD), or sepsis. This project was reviewed and approved by the institutional review board at VA Pittsburgh Healthcare System as a quality improvement project.

Our intervention included at least 2 scheduled in-home visits and 2 follow-up telephone visits in 90 days, delivered by a multidisciplinary team of palliative care, geriatric, and critical care nurses, nurse practitioners, and physicians; physical therapists; and social workers (see Figure, part A; home visit data collection form: Supplement 1). Eligible veterans were identified during hospitalization. The first home visit was scheduled within 72 hours of hospital discharge, and the final visit was scheduled 90 days after hospital discharge. Scripted telephone calls were planned at weeks 4 and 8 (telephone visit data collection form: Supplement 2). The team met weekly to discuss the patients’ progress and to develop, monitor, and refine care plans.

The care team assessed patients’ physical, medical, psychological, social, and spiritual needs during the 2 home visits with a multidisciplinary approach involving physicians, nurse practitioners, nurses, and social workers. The team assessed baseline physical function and symptom burden with the Palliative Performance Scale (PPS) and simplified Edmonton Symptom Assessment System (sESAS; Supplemental Tables 1–3), which have been extensively validated in cancer and noncancer populations.^4–6  The participants completed a perception-of-care questionnaire after the final home visit (Supplement 3). Telephone visits were conducted to provide scheduled updates on new health events including changes in symptoms, new medical problems, unplanned rehospitalizations, and emergency department visits. A team member was available 24-7 via an emergency telephone number to help manage new symptoms that required immediate medical attention.

The primary outcome was the operational feasibility of conducting this trial.⁷  We aimed to enroll 1 to 2 patients per month for 12 months and complete at least 70% of all prespecified home and telephone visits within the proposed time windows. Secondary outcomes included (1) PPS score at day 90, (2) sESAS score at day 90, (3) patient/caregiver satisfaction, and (4) number of unplanned hospital readmissions and emergency department visits within 90 days.

We present categorical data as count and percentage and continuous data as median and inter-quartile range (IQR). We refrained from formal statistical testing because of the limited sample size. We estimated resource utilization by applying current procedural terminology billing codes for in-home, telephone, and care coordination visits (Supplemental Table 4). We compared the number of unplanned readmissions and emergency department visits during the study period with the number in the 3-month period preceding the index hospitalization. Data management and analyses were performed with Stata/SE 15.1 (Stata Corp).

Patient characteristics are summarized in the Table. Median (IQR) age was 76 (69-87) years, and all participants were male and multimorbid (median [IQR] number of chronic health conditions, 8 [7-8]). Patients were moderately debilitated at baseline (median [IQR] PPS score, 60 [60-70]; median (IQR) sESAS score, 4 [2-7]). The reasons for the index hospitalization were heart failure exacerbation (67%), community-acquired pneumonia (20%), sepsis (7%), and COPD exacerbation (7%). Ten patients required organ support, 8 received noninvasive or invasive positive pressure ventilation, and 3 received vasopressor support. The median (IQR) APACHE score at admission was 15 (13-18). The most common reasons for nonenrollment were a dependent living situation before hos-pitalization, transfer from an outside hospital, or residence more than 25 miles from our institution.

We enrolled 15 patients in 12 months. Ninety-day follow-up was available in all 15 patients; 11 patients (73%) completed all study visits (see Table). Two patients (13%) died before study completion, and 2 patients did not complete all visits because of prolonged rehospitalization.

The median (IQR) time to first study visit was 8 (5-12) days. Each patient received a median (IQR) of 8 (5-11) in-home visits during the 90-day study period: 4 (3-6) visits by nurses, 3 (2-4) visits by physicians, and 1 (1-1) visit by a social worker. In addition to home visits, patients had a median (IQR) of 6 (3-7) telephone encounters during the study period (see Table).

Each patient was visited at home by nurses for a median (IQR) total of 330 (240-585) minutes and was visited by telephone for a median (IQR) cumulative time of 30 (10-70) minutes (see Table). The median (IQR) time for each home visit was 90 (75-90) minutes. We estimated the median (IQR) cost per patient to be $2321 ($1901-$3331), which is in stark contrast to the average cost of $31 679 per ICU admission for the very elderly.⁸  We refrained from formal cost-benefit analysis at this early stage of feasibility testing, but we estimate that the intervention was cost-neutral in our setting.

Median PPS scores at day 90 (n = 11; median [IQR], 70% [50%-80%]) were slightly higher than the initial PPS scores (n = 15; median [IQR], 60% [60%-70%]; see Table). Consistently, median sESAS scores were slightly lower at day 90 (n = 11; median [IQR], 3 [2-4]) than the initial scores (n = 15; median [IQR], 4 [2-7]). Of the 11 patients who completed all visits, 3 (27%) were readmitted within 30 days and 7 (64%) were readmitted within 90 days. The median number of unplanned hospital admissions during the project period (median [IQR], 1 [0-2]) was similar to that of the 3-month period preceding index hospitalization (median [IQR], 1 [0-3]), whereas unplanned emergency department visits were lower during the project period (median [IQR], 1 [0-2]) than they were during the 3-month period preceding index hospitalization (median [IQR], 2 [2-4]); see Figure, part B). Patients rated their overall satisfaction with this intervention with a median (IQR) score of 4.5 (4-5) (n = 11) on a 5-point Likert scale.

We demonstrated operational feasibility of a home-based multidisciplinary intervention in multimorbid ICU survivors and reached our predefined goals of enrolling 15 patients within 12 months and completing at least 70% of prespecified visits. The importance of palliative care is highlighted by the fact that 1 patient transitioned to hospice during during the 90-day study period and 2 additional patients transitioned to hospice shortly after the 90-day study period. We observed improvements in physical functioning, quality of life, and symptom control, which contributed to high patient satisfaction. This intervention was time intensive and required multiple in-person and telephone visits by nurses, physicians, and social workers. Different models of care may be appropriate for different ICU survivors, and a resource-intensive intervention like this may be appropriate only for a subset of high-risk patients. On the basis of this experience, we recommend that future studies test efficacy, generalizability, and scalability; optimize efficiency by incorporating novel technology (telehealth visits); test alternative follow-up periods (eg, 60 days vs 90 days); and identify subgroups of ICU survivors who are most likely to benefit from home-based care interventions.

Our project has several limitations. First, it was designed as a quality improvement study to test the operational feasibility of implementing comprehensive home-based palliative care in multimorbid veterans. Although we were able to meet our enrollment criteria, the scalability of interventions like this will depend on the ability to automatize study procedures, for example, through EHR-prompted screening. Second, the median time to the first home visit was 8 days although it was planned to be 72 hours. Future studies of home-based interventions should aim to complete initial home visits as early as possible because the period immediately after hospital discharge is the most stressful and vulnerable time for patients and caretakers. Third, our project was conducted in the unique environment of the largest integrated health care system, and our findings may not generalizable to other more diverse settings. Fourth, we did not evaluate caregiver satisfaction during this project.

In conclusion, we successfully demonstrated the operational feasibility of a comprehensive home-based palliative care intervention in multimorbid elderly veteran survivors of critical illness at high risk of hospital readmission.

The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.