Skilled Nursing Facility Care at Home for Adults Discharged From the Hospital: A Pilot Randomized Controlled Trial

Abstract

Objectives

The ability to deliver SNF-level care at home is unclear. We sought to demonstrate the feasibility of rehabilitation at home (RAH) via a pilot randomized controlled trial.

Methods

Daily care from a home health aide, certified nursing assistant, nurse, and physician. Core technologies included remote PT, automated medication dispensing, and continuous monitoring. Primary outcome: episode cost.

Results

We randomized 10 patients. Home patients’ episode cost a median $8404 (IQR, $2697) versus $9215 (IQR, $5702). LOS for both was 14 days. Home patients’ ADLs improved between admission and 30-days post-discharge by median 4 (IQR, 5) versus 1 (IQR, 2). Home patients’ median Picker patient experience score was 12/14 (IQR, 2) versus 7/14 (IQR, 3).

Discussion

A RAH pilot compared favorably to traditional SNF, with trends toward lower cost, greater functional status improvement, and better patient experience. This delivery innovation could reimagine how we deliver post-acute care but requires replication.

Clinicaltrials.gov registration

NCT04048590.

Keywords

skilled nursing facilities home care services rehabilitation rehabilitation centers post-acute care

Introduction

About 40% of older adults discharged from acute care hospitals require post-acute care (PAC), and about half of these receive care in an inpatient skilled nursing facility (SNF) (Report to the Congress: Report to the Congress, 2019). After a qualified hospital stay, SNF care includes 1.5 hours of therapy 5 days per week, nursing care, assistance with functional limitations, medications, and semi-private room and board.

Care in a SNF is unfortunately of variable safety, quality, and experience (Wunderlich & Kohler, 2001). In 2014, the Inspector General found that 22% of Medicare beneficiaries experienced adverse events during their SNF stays; 59% of these events were clearly or likely preventable. Half who suffered an event returned to the hospital (Levinson, 2014), and almost one-fourth discharged from a SNF are readmitted within 30 days (Mor et al., 2010). Poor quality leads to fewer days spent at home following discharge (Simning et al., 2020). Great variability exists in the amount of care received from a physician or advanced practice provider such that about 10% of patients never see a provider and about 70% see one within 4 days (Ryskina et al., 2019).

Skilled nursing facility cost is high, accelerating, and variable. In 2017, Medicare spent $28.4 billion on SNF services, with a $18,121 median payment per stay (Report to the Congress: Report to the Congress, 2019). Spending has accelerated such that cost for many chronic conditions in the post-acute setting nearly equals that of the acute setting (Mechanic, 2014), with spending in post-acute care noted as the largest driver of variability in Medicare (Newhouse & Garber, 2013). While one common way to reduce SNF cost is to change how it is paid for (e.g., bundled payments, which likely incentivize decreases in SNF length of stay), another is to change the site of post-acute care.

We hypothesized that many of the concerns associated with SNF care (Chandra et al., 2013) could be ameliorated—primarily high cost and secondarily variable quality, safety, and experience—by shifting the site of care to the patient’s home, while deploying technology and personnel in the home. Certain technologies and personnel might enable automation, task shifting, and allow all team members to practice at the top of their skill level. These shifts would provide support tailored to one’s actual living circumstances, an environment that encourages earlier mobilization, support of and interaction with family and caregivers, and psychosocial benefits of being at home. SNF care at home is likely preferred by many patients rather than receiving such care in an institution. During pandemic conditions, this care model also brings the benefit of removing a patient from a congregate setting. We therefore piloted “rehabilitation-at-home” (RAH) as a care model that fully substituted for traditional SNF care after discharge from the hospital.

Methods

Design Overview

We performed a parallel-design, pilot randomized controlled trial in which participants were randomly allocated to RAH (intervention) or traditional SNF (control) care. We enrolled participants between August 8, 2019 and September 27, 2019. Patients, clinicians, and study staff were not blinded to allocation status. This investigator-initiated study was approved by the Mass General Brigham institutional review board and registered at clinicaltrials.gov (NCT04048590).

Setting and Participants

Adult participants were recruited from the general medical wards of Brigham and Women’s Hospital (an academic medical center) and Brigham and Women’s Faulkner Hospital (a community hospital). A research assistant screened all patients who were referred to SNF by the case management team, thus ensuring the patient required SNF-level care. Patients were included if they were likely to return to community dwelling status after SNF care (a subjective determination by the inpatient and study teams), had capacity to consent, and lived within 10 miles of the hospital. Importantly, patients were approached only if deemed needing SNF-level care by the inpatient physical therapist; as is often the case, they may have declined this care and therefore been ineligible for enrollment. Patients were excluded if they were undomiciled, lacked basic utilities, screened positive for domestic violence, or resided in a facility that did not allow on-site medical care. Patients were also ineligible if they had ventilator needs, end-stage renal disease on hemodialysis, required intravenous controlled substances, or required transfusion of blood products.

Randomization

Patients who met criteria and provided written informed consent were randomized to usual care in a SNF or to RAH. Randomization was stratified by functional status, obtained after consent (activities of daily living: 0, 1, 2–3, 4–6), with randomly selected block sizes between 4 and 6 with allocation concealment via sealed envelopes. An outside statistician generated the randomization using SAS (SAS Institute).

Intervention

On admission to their home, patients randomly assigned to RAH received a nurse-certified nurse assistant (CNA) dyad visit and a geriatric medicine attending physician visit. On subsequent days, the CNA visited daily, with the nurse available for as-needed virtual and in-person consultation. The CNA facilitated medication administration via an automated remotely programmable medication dispenser overseen remotely by the physician (MedaCube, PharmAdva). The physician was updated daily by the CNA via video, voice, and/or text. Tailored to patients’ needs, a home health aide (HHA) was present in the home up to 24 hr a day. A remote physical therapist first assessed a patient and then programmed a virtual therapy avatar for therapy sessions up to three times daily initiated by the patient and/or their in-home care team (Vera, Reflexion Health) (Prvu Bettger et al., 2020). The system used a depth camera to monitor patient movement and effectiveness and the session was facilitated, as needed, by in-home aides who had received 1 hour of training regarding the system. If necessary, an in-home physical therapist was available if the remote physical therapy was insufficient (this was not deemed necessary as therapy goals were consistently met). Patients were biometrically monitored continuously, including changes in their heart rate, respiratory rate, minutes of sleep, and step count, with the result that as 24/7 staffing was tapered down, the team would still be able to closely monitor patients (Health Tag, Spire Health).

Criteria for discharge were by design left to the discretion of the interdisciplinary team. We mandated no treatment pathways or algorithms. Follow-up after discharge was by design no different than usual care.

Patients randomized to receive usual care in a SNF were transported to a traditional SNF and were not monitored in any way.

Data Sources and Outcomes

For both groups, study staff interviewed patients on admission to PAC (discharge from hospital), at discharge from PAC, and at 30-days post-discharge from PAC. On admission, we obtained data on the hospital admission (length of stay, requirement for ICU, orthopedic involvement, and discharge diagnosis) and patients reported their sociodemographic characteristics and completed assessments of frailty (PRISMA; Raîche et al., 2008), cognitive impairment (Ascertain dementia-8; Galvin et al., 2005), depression (PHQ-2; Kroenke et al., 2003), emotional support (PROMIS; Ader, 2007), health literacy (BRIEF; Chew et al., 2004), health-related quality of life (EuroQol VAS; van Hout et al., 2012), and functional status (Lawton & Brody, 1969). Upon study enrollment, patients were asked to report their activities of daily living (ADLs) and instrumental activities of daily living (IADLs) both at present and 30 days before presentation to the hospital. We again asked ADLs and IADLs upon discharge from PAC and 30 days after discharge from PAC. Study staff supplemented patient-reported data with the hospitals’ electronic health record (EHR) for items such as insurance status.

Our primary outcome was direct cost of the post-acute care episode. For the intervention group, we calculated cost as the sum of non-physician labor, supplies, monitoring equipment, medications, labs, radiology, and transport directly attributed to the patient’s care using a previously accepted methodology (D. M. Levine et al., 2020). For patients in the control group who received care at an in-network SNF, we obtained the direct costs for the patient’s individual stay. For the patients in the control group who received care at an out-of-network SNF, we calculated total cost by multiplying length of stay by the facility-level average cost per day reported by a facility on their Centers for Medicare and Medicaid Services (CMS) cost report (Cost Reports, n.d.). Physician labor was excluded from direct cost calculations because this cost is billed separately from traditional facility cost.

We secondarily studied utilization, physical activity, patient experience, safety, and quality during the rehabilitation care episode. Utilization included IV medication, oxygen, nebulized therapy, lab orders, radiology studies, clinician visits, and length of stay. Physical activity included daily steps but was only available for the RAH group. Patient experience measures included the Care Transitions Measure (CTM)-3 (3-Item Care Transitions Measure (CTM-3), 2018), the Picker patient experience questionnaire-15 (Jenkinson, 2002), recommending the rehabilitation experience, and global experience, based on the 30-day post-discharge interview. Quality measures included days at home following hospital discharge (period of 30 days), days at home following SNF discharge (period of 30 days), pain scores (0–10), inappropriate medication use, urinary catheter use, and restraint use. Safety measures included unplanned mortality (obtained through interview and considered unplanned if the patient was not discharged on hospice), delirium (measured through the confusion assessment method), fall, DVT/PE, and new pressure ulcer. In the RAH group, all measures were derived from the EHR or patients, except for step count and nightly sleep, which was observed via the biometric sensor. In the control group, we derived all measures from a combination of the EHR, medical records request (when the patient went to a SNF outside of our system), or patients.

We additionally measured cost and utilization in the 30-day post-discharge period using the same cost accounting method. We tracked readmissions, distinct ED visits, presenting to primary care within 14 days, and unplanned mortality. In addition to EHR records to all Mass General Brigham facilities (the health system that includes Brigham and Women’s Hospital and Brigham and Women’s Faulkner Hospital) and the CareEverywhere system that joins all institutions that use the Epic EHR, we asked participants whether they received any health care outside of our health system during the 30-day post-discharge interview and added those to the cost estimates. None did. For the one patient in the control group who we were unable to reach 30 days after discharge, we used EHR data alone to estimate utilization and readmission rates and did not measure patient experience or functional status.

Statistical Analysis

This pilot study was constrained by funding, allowing for only a 2-month enrollment period. Given our small sample size, and in the case of cost, skewed data, we present medians and interquartile ranges (IQRs). We also do not show statistical comparisons given the proof-of-concept nature of our pilot and the small sample size. We performed all analyses in SAS v9.4 (SAS Institute).

Results

Patient Characteristics

Of the 51 patients who were screened for eligibility, 10 were enrolled and randomly assigned to a group (Figure 1). Exclusions occurred because patients declined to consider any facility-based PAC (24%), patients declined RAH (37%), or the caregiver declined RAH (15%), among others. Patients who declined were similar to those who enrolled (supplemental material eTable 1). All patients assigned received their allocated treatment.

Figure 1.

Participant Flow. The Consolidated Standards of Reporting Trials flow diagram. a: 1 patient declined to answer the discharge telephone questionnaire. b: 1 patient declined to answer the post-discharge telephone questionnaire; 1 patient passed away prior to 30 days. Not completing a discharge or 30-day post-discharge call required estimation of post-discharge utilization through the electronic health record and incurred missing values for functional assessment and patient experience measures. Note. RAH, rehabilitation at home; SNF, skilled nursing facility.

Both groups had similar characteristics. Patients were in general in their 80s, chronically ill, frail, well-educated, English-speaking, insured, and retired (Table 1). Patients utilized a significant amount of care, with a median of two hospitalizations in the last 6 months and had complex medication regimens (median of 18 outpatient medications). Most patients lived alone (RAH group, 60%; control group, 100%) and few used a home health aide prior to their hospitalization (RAH group, 20%; control group, 40%). Health-related quality of life was subpar at 54 out of 100 by the EuroQol visual analogue scale.

Table 1.

Baseline Patient Characteristics.

	Home (n = 5)	Control (n = 5)
Age, y, median (IQR)	88 (11)	80 (8)
Female, n (%)	4 (80)	3 (60)
Race/ethnicity, n (%)
White	4 (80)	4 (80)
Latin@	1 (20)	1 (20)
Partner status, n (%)
Partnered	2 (40)	1 (20)
Divorced	0	2 (40)
Widowed	2 (40)	0
Single, never partnered	1 (20)	2 (40)
Primary language, n (%)
English	4 (80)	4 (80)
Spanish	1 (20)	1 (20)
Insurance, n (%)
Medicare	4 (80)	4 (80)
Medicare+Medicaid	1 (20)	1 (20)
Education, n (%)
High school	1 (20)	0
<4-year college	1 (20)	2 (50)
4-year college	1 (20)	2 (50)
>4-year college	2 (40)	0
Employment, n (%)
Unemployed	1 (20)	0
Retired	4 (80)	5 (100)
Cigarette smoking, n (%)
Never	1 (20)	3 (60)
Current	0	1 (20)
Prior	4 (80)	1 (20)
BMI, median (IQR)	23 (3)	20 (5)
PRISMA frailty score (0–7), median (IQR)^a	4 (2)	4 (2)
Comorbidity count, median (IQR)^b	14 (4)	9 (5)
Number of outpatient medications, median (IQR)	19 (4)	17 (6)
Hospitalizations in last 6 months, median (IQR)	2 (1)	1 (1)
Emergency department visits in last 6 months, median (IQR)	1 (1)	2 (2)
Ascertain dementia-8 (0–8), median (IQR)^c	1 (0)	1 (2)
PHQ-2 (0–6), median (IQR)^d	1 (2)	2 (1)
PROMIS emotional support (4–20), median (IQR)^e	18 (6)	16 (4)
BRIEF health literacy tool (4–20), median (IQR)^f	19 (4)	16 (5)
EuroQol VAS (0–100), median (IQR)^g	54 (25)	55 (10)
Code status: Full code, n (%)	2 (40)	5 (100)
Prior home health aide, n (%)	1 (20)	2 (40)
Lives alone, n (%)	3 (60)	5 (100)
Yes, surprised if died in 1 year, n (%)	3 (60)	3 (60)

Note. ADLs = activities of daily living; CI = confidence interval; COPD = chronic obstructive pulmonary disease; IADLs = instrumental activities of daily living; IQR = interquartile range; PHQ-2 = Patient Health Questionnaire-2; PRISMA = Program of Research to Integrate Services for the Maintenance of Autonomy; PROMIS = Patient-Reported Outcomes Measurement Information System; UTI = urinary tract infection; VAS = visual analogue scale.

^a>2 indicates frailty.

^bCount of patient’s chronic comorbidities, out of the 20 conditions considered chronic by the Health and Human Services Office of the Assistant Secretary of Health.³⁵

^c>1 indicates cognitive impairment.

^d>2 indicates depression.

^e>17 indicates better than average emotional support.

^fBrief health literacy screener, 4–12: limited; 13–16: marginal; 17–20: adequate.

^gFor reference, the mean VAS score for individuals age 80+ is approximately 62 (Szende & Williams, 2004)

The hospitalizations for both groups were similar. Median length of stay was 8 days (IQR, 7) in the control group and 6 (IQR, 11) days in the RAH group. One patient in the control group and three patients in the RAH group had an orthopedic issue during hospitalization. One patient in the RAH group required ICU-level care during hospitalization. Discharge diagnoses were variable, including heart failure, pneumonia, fracture, and other infections.

Cost and Health Care Utilization

Control patients used 4 different SNFs. Median cost of the RAH episode was $8404 (IQR, $2697) compared to $9215 (IQR, $5702) for the control group (Table 2). Length of stay in the RAH group was 14.0 (IQR, 1.3) days and 13.8 (IQR 7.9) days in the control group. Utilization of IV medication (0 in both groups), oxygen (1 instance in RAH), and nebulized therapy (1 in control group) was less common. Imaging occurred in 60% of the RAH group versus 25% of the control group. A median of one lab order was drawn in the RAH group versus 11 in the control group.

Table 2.

Cost and Patient Utilization.

Measure	Home (n = 5)	Control (n = 5)
During rehab episode
Cost, median $ (IQR)	8404 (2697)	9215 (5702)
Length of stay, median days (IQR)	14.0 (1.3)	13.8 (7.9)
IV Medication during admission, n (%)	0	0
Oxygen during admission, n (%)	1 (20)	0
Nebulized therapy during admission, n (%)	0	1 (25)
Imaging during admission, n (%)	3 (60)	1 (25)
Lab orders per admission, median (IQR)	1 (4)	11 (20)
MD home visits, median # (IQR)	1 (2)	—
MD video visits, median # (IQR)	0 (2)	—
RN home visits, median # (IQR)	2 (1)	—
CNA home visits, median # (IQR)	14 (2)	—
PT video visits, median # (IQR)	1 (1)	—
PT avatar sessions, median # (IQR)	11 (1)
OT home visits, median # (IQR)	0 (1)	—
HHA hours, median hours (IQR)	212 (86)	—
Disposition, n (%)
Routine	2 (40)	2 (40)
Home health	2 (40)	3 (60)
Return to hospital	1 (20)	0
30 days after discharge
Primary care visit by 14 days post-discharge, n (%)	0	1 (20)
30-day readmission, n (%)	1 (20)	2 (40)
30-day ED presentation, n (%)	0	0
Days at home (from rehabilitation discharge), median (IQR)^a	30 (0)	30 (10)

Note. ED = emergency department; IQR = interquartile range; IV = intravenous; OT = occupational therapy; PT = physical therapy; “-,” indicates not applicable.

^aThe number of days starting with discharge from rehabilitation and ending 30 days later that a patient was home.

Throughout a RAH episode (including admission), patients received a median of 1 (IQR, 2) physician home visit, 0 (IQR, 2) physician video visits, 2 (IQR, 1) nurse home visits, 14 (IQR, 2) CNA visits, 1 (IQR, 1) live video PT visit, 11 (IQR, 1) PT avatar sessions, and 212 (IQR, 86) hours of home health aide care. Patients were discharged with home health (RAH, 40%; 60%, control), with no services (RAH, 40%; control, 40%), and one patient in the RAH group returned to the hospital prior to discharge due to orthopedic hardware malfunction (Table 2).

30 days after discharge from the rehabilitation episode, 20% of patients in the RAH group and 40% of patients in the control group had been readmitted (Table 2). RAH patients spent 30 days (IQR, 0) at home following discharge from rehabilitation; control patients spent a median of 30 days (IQR, 10) as well but with more variation. None presented to the emergency department. Almost no patients had seen their primary care physician within 14 days of discharge.

Functional Status and Health-Related Quality of Life

Activities of daily living were lower upon enrollment than 30 days prior to presentation to the hospital in both groups (Table 3). In the RAH group, ADLs increased during the rehabilitation episode (2 [IQR 0] on admission to 4.5 [IQR, 1.5] on discharge) and remained similarly increased 30 days after discharge (5 [IQR, 1.5]). In the control group, ADLs remained similar during the rehabilitation episode (3 [IQR two] on admission to 3.5 [IQR, 3.5] on discharge) and at 30-days post-discharge (4 [IQR 2]).

Table 3.

Functional Status and Health-Related Quality of Life.

Measure	Home (n = 5)	Control (n = 5)
ADLs (0–6), median (IQR)
–30	3 (4)	6 (1)
Admission	2 (0)	3 (2)
Discharge	4.5 (1.5)	3.5 (3.5)
+30	5 (4)	4 (2)
Admission->Discharge	2.5 (2.5)	0.5 (4)
Admission->+30	4 (5)	1 (4)
IADLs (0-8), median (IQR)
–30	7 (4)	5 (2)
Admission	3 (2)	4 (2)
Discharge	3.5 (3)	4 (3)
+30	3 (2)	4 (5)
Admission->Discharge	1 (3)	−1 (3)
Admission->+30	0 (4)	0 (4)
EQ-VAS, median (IQR)
Admission	54 (25)	55 (10)
Discharge	50 (15)	50 (25)
+30	55 (10)	65 (30)

Note. ADLs = activities of daily living; IADLs = instrumental activities of daily living; EQ-VAS = EuroQol-Visual Analogue Scale

Note: “−30” signifies 30 days prior to discharge from the hospital. “Admission” signifies day of discharge from the hospital and admission in RAH. “Discharge” signifies discharge from RAH. “+30” signifies 30 days after discharge from RAH.

IADLs similarly fell between 30 days prior to presentation and enrollment in both groups (Table 3). IADLs remained similar throughout the rehabilitation episode and 30 days following discharge in both the RAH group (3 [IQR, 2] -> 3.5 [IQR, 3] -> 3 [IQR, 2]) and control group (4 [IQR, 4] -> 4 [IQR, 3] -> 4 [IQR, 5]).

Health-related quality of life remained similar throughout the rehabilitation episode and 30 days following discharge (Table 4).

Table 4.

Quality, Safety, Physical Activity, and Experience.

Measure	Home (n = 5)	Control (n = 5)^a
Quality
Days at home (from hospital discharge), median (IQR)^b	30 (3)	16 (7)
Pain score (0–10), median (IQR)	5.2 (3.3)	1.78 (0)
Inappropriate medication use, n (%)^c	0	1 (25)
Urinary catheter use, n (%)	0	0
Restraint use, n (%)	0	0
Safety
Unplanned mortality during episode, n (%)	0	0
Delirium, n (%)	0	0
Fall, n (%)	1 (20)	0
DVT/PE, n (%)	0	1 (20)
New pressure ulcer, n (%)	1 (20)	0
30-day unplanned mortality, n (%)	0	1 (20)
Activity each day
Nightly sleep, median minutes (IQR)	349 (205)	-
Steps per day, median (IQR)	43 (194)	-
Patient experience
Care transitions measure-3 (3–12), median (IQR)	12 (4)	6 (5)
Picker questionnaire (0–15), median (IQR)	11.5 (2)	7 (3)
Global satisfaction (0–10), median (IQR)^d	8.5 (1)	7 (1.5)
Recommend hospital (0–4), median (IQR)^e	3 (1)	3.5 (2.5)

Note. DVT/PE = deep venous thromboembolism/pulmonary embolism; IQR = interquartile range; IV = intravenous

^aFor one control patient, data during the rehabilitation episode are missing.

^bThe number of days starting with discharge from the hospital and ending 30 days later that a patient was home.

^cUsing the updated Beers criteria.³⁶

^dScale: 0 = the worst possible rehab experience; 10 = the best possible rehab experience.

^eScale: 0 = definitely would not recommend; 4 = definitely would recommend.

Outcomes During Rehabilitation

From the time of hospital discharge to 30-days later, patients in RAH spent a median of 30 days (IQR, 3) at home versus 16 days (IQR, 7) at home for patients in the control group (Table 4). Pain scores were higher in RAH (5.2 [IQR, 3.3] vs. 1.8 [IQR, 0]), although this data point was driven substantially by a single patient. Isolated safety events occurred in both arms: 1 fall and 1 pressure ulcer occurred in the RAH group; 1 DVT and 1 unplanned death occurred in the control group.

In the RAH group, patients slept a median of 349 min (IQR, 205) each night. They took a median of 43 steps (IQR, 194) daily, although it is likely daily steps were erroneously measured due to impaired gait, use of a walker, or other aide.

Patient experience appeared higher in the RAH group. Patients reported higher ratings on the care transitions measure (12 [IQR, 4] vs. 6 [IQR, 5]), Picker questionnaire (11.5 [IQR, 2] vs. 7 [IQR, 3]), and global satisfaction 8.5 [IQR, 1] versus. 7 [IQR, 1.5]).

Discussion

In this small pilot RCT of RAH compared to traditional facility-based SNF PAC, we demonstrated signals toward improved patient experience, functional status, cost, and days at home with similar length of stay and quality. To our knowledge, this is the first US study to compare a home-based PAC model as a substitute for SNF-based PAC through a RCT. It is also likely the first to care for highly debilitated patients at home. Our model required an interprofessional team augmented by remote therapy, remote medication dispensing, and biometric monitoring.

Several factors related to the intervention may have contributed to our pilot’s favorable findings. First, the ability for a patient to rehabilitate in the comfort of their own familiar environment surrounded by their family and friends cannot be understated. The minimum clinically important difference in home time is about 18.6 days in a year (Lee et al., 2019). RAH achieved this rapidly. Prior qualitative work of acute care at home demonstrates reduced anxiety, improved locus of control, and increased movement when patients receive care at home (Dubois & Santos-Eggimann, 2001; Leff et al., 2006; D. Levine et al., 2021; D. M. Levine et al., 2018; Wilson et al., 2002). All of these benefits likely exist when receiving RAH and are likely magnified as length of stay is longer during a rehabilitation episode and the need to adapt one’s environment to new physical limitations is more important.

Second, RAH systematically leveraged technology. To optimize therapy, we deployed semi-structured remote physical therapy, allowing for highly tailored and frequent sessions. To enhance the time-consuming but important task of medication administration, we automated it. To improve monitoring of patients, particularly as in-home personnel became less frequent, we used wearable sensors. The task shifting allowed by technology is compelling from a scale and economic perspective but could be concerning if not evaluated effectively in future studies.

Third, RAH ensured all clinicians practiced at the top of their license. We placed CNAs in central roles with remote nurse backup and daily remote physician oversight. We tailored and tapered the 24/7 HHA presence as patient functional status improved. Importantly, HHA hours accounted for the overwhelming majority of cost for the intervention group, yet overall cost trended less than the control group. During a time when staffing is a major challenge, this model has pros and cons. On the one hand, it largely reduces the nursing and physician requirements compared to traditional SNF. On the other hand, it has increased need for HHAs, which cannot be shared among patients as they are in a SNF. Given the training and pipeline requirements for clinicians, we feel this model may prove to be a net benefit to the industry, but future studies will need to look carefully at the labor requirements for RAH compared with SNF, at the labor markets for the personnel required in both settings, and at the safety of care delivery with novel personnel. This study preliminarily demonstrates the ability of the right selection criteria, technology, and novel personnel to offset what is a requisite high-touch HHA caregiving asset.

Our study corroborates and adds to prior work. In the REACH-OUT trial, an iteration on their hospital-in-the-home work in Australia, Caplan and colleagues demonstrated in a RCT a lower risk of delirium, greater patient experience, lower cost, and more efficient hospital bed use with an “early discharge” approach (Caplan et al., 2006). Of note, participants had to be able to transfer independently and toilet themselves in REACH-OUT. Patients in our study performed far fewer ADLs (none in the intervention group could transfer or toilet independently for example). In the US, Augustine and colleagues described a retrospective single-arm implementation of RAH (Augustine et al., 2020). More than half of these patients required at least moderate assist with functional mobility. Each week, they received 1.8 medical visits, 1.7 nurse visits, and 5.2 PT visits. Only 27% received HHA visits, averaging 3.9 weekly visits. Compared to both prior studies, our care model provided much more intensive services (for example, 24-hour HHA care) for patients with considerably more debility. Our evaluation approach demonstrated the feasibility of a RCT in the US and lays an important foundation for a much-needed larger RCT.

Our study has limitations. First, our small sample size resulted in differences between the groups and an inability to adjust for these differences. We undoubtedly found important signals from this pilot and a larger RCT is planned. Second, this study has limited generalizability because we enrolled patients from two sites in a similar geographic area. Third, we could not measure cost on the patient level for patients who were cared for at an out-of-network SNF and had to use average cost per day for these patients. Given the wide range in length of stay of our patients, it is unclear in which direction this may have biased our result. Fourth, about 63% of patients or their caregivers declined to participate, approximately the inverse of prior work, which resulted in about half the sample size than intended (Augustine et al., 2020; Caplan et al., 2006). Three reasons likely explain this. First, nearly one quarter of patients declined transfer to SNF. Due to our robust randomization schema, we could not enroll them. Second, patients or their caregivers may have been concerned about the reduction in in-home care by skilled professionals such as therapists or nurses that this model incorporates, even with a HHA present 24/7. Third, we did not have the RAH physician available in-person (only by video) to assist with recruitment, both for the patient, their family, and the primary hospital team. This left the research assistant to maneuver through the complex care coordination system. From prior work on acute care at home, we feel that many of the declines involving the patient, family, and care coordinator could have been ameliorated with a greater physician presence (D. M. Levine et al., 2021). Learnings from recruitment challenges will change how we recruit for a follow-on RCT.

Our study reinforces some interesting questions raised in recent years around the “true” value of SNF care. The largest single contributor to savings within Medicare’s largest accountable care organization (ACO) model and different bundled payment models has come through reductions made in SNF post-acute care without significant short-term differences in outcomes or patient satisfaction (Barnett et al., 2019). Between 2006 and 2011, nearly one-third of Medicare decedents used their SNF benefit for “rehab” in their last 6 months of life, suggesting little is known about which patients will actually benefit from facility-based SNF care (Aragon et al., 2012). Finally, the COVID-19 pandemic saw rapid dramatic declines in facility-based SNF PAC and a concurrent increase in home-based care utilization for patients who typically would have been referred to inpatient SNF PAC (Flynn et al., 2020). These and other observations highlight the short-comings of past SNF PAC policies and leaves room for the potential positive impact of revisioning the location, personnel, and technology involved in delivering high-value PAC.

Conclusions and Implications

Rehabilitation at home as a substitute for a traditional SNF has favorable signals in patient experience, functional status, cost, and days at home. Following replication in a larger trial, RAH could become the standard for high-value SNF PAC.

Supplemental Material

sj-pdf-1-jag-10.1177_07334648221077092 – Supplemental Material for Skilled Nursing Facility Care at Home for Adults Discharged From the Hospital: A Pilot Randomized Controlled Trial

Supplemental Material, sj-pdf-1-jag-10.1177_07334648221077092 for Skilled Nursing Facility Care at Home for Adults Discharged From the Hospital: A Pilot Randomized Controlled Trial by David M. Levine, Mary A. Cueva1, Sandra Shi, Idriz Limaj, Bessey Wambolt, David C. Grabowski, Jeffrey L. Schnipper and Charles T. Pu for Journal of Applied Gerontology

Footnotes

Declaration of Conflicting Interests

Mary A Cueva: None, Sandra Shi: None, David Grabowski: research grants from the National Institute on Aging, the Agency for Healthcare Research and Quality, Donaghue Foundation, Warren Alpert Foundation, and the Arnold Foundation. He received personal fees from naviHealth, Medicare Payment Advisory Commission, Compass Lexecon, Analysis Group, Abt Associates, and the Research Triangle Institute. Charles T Pu: former advisor and current shareholder at MDOps Corporation, David Levine: PI-initiated grants from IBM and Biofourmis, separate from the present work. Jeffrey Schnipper: funding from Mallinckrodt Pharmaceuticals for an investigator-initiated study of opioid related adverse drug events in post-operative hospitalized patients, separate from the present work. Bessey Wamboldt: None

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Department of Internal Medicine, Brigham and Women’s Hospital; Philanthropic gift from the late Mrs. Henry Ford II (Kate)

IRB Review

Mass General Brigham IRB (formerly Partners HealthCare), 2019P001231

ORCID iD

David M. Levine

Supplemental Material

Supplemental material for this article is available online.

References

3-Item Care Transitions Measure (CTM-3) (2018, August 23). CMS measures inventory tool. https://cmit.cms.gov/CMIT_public/ViewMeasure?MeasureId=1700

Ader

D. N.

(2007). Developing the patient-reported outcomes measurement information system (PROMIS). Medical Care, 45(Suppl 1), S1–S2. https://doi.org/10.1097/01.mlr.0000260537.45076.74

Aragon

Covinsky

Miao

Boscardin

W. J.

Flint

Smith

A. K.

(2012). Use of the medicare posthospitalization skilled nursing benefit in the last 6 months of life. Archives of Internal Medicine, 172(20), 1573–1579. https://doi.org/10.1001/archinternmed.2012.4451

Augustine

M. R.

Davenport

Ornstein

K. A.

Cuan

Saenger

Lubetsky

Federman

DeCherrie

L. V.

Leff

Siu

A. L.

(2020). Implementation of post-acute rehabilitation at home: A skilled nursing facility-substitutive model. Journal of the American Geriatrics Society, 68(7), 1584–1593. https://doi.org/https://doi.org/10.1111/jgs.16474

Barnett

M. L.

Mehrotra

Grabowski

D. C.

(2019). Postacute care - the piggy bank for savings in alternative payment models? The New England Journal of Medicine, 381(4), 302–303. https://doi.org/10.1056/NEJMp1901896

Caplan

G. A.

Coconis

Board

Sayers

Woods

(2006). Does home treatment affect delirium? A randomised controlled trial of rehabilitation of elderly and care at home or usual treatment (The REACH-OUT trial). Age and Ageing, 35(1), 53-60. https://doi.org/10.1093/ageing/afi206

Chandra

Dalton

M. A.

Holmes

(2013). Large increases in spending on postacute care in medicare point to the potential for cost savings in these settings. Health Affairs (Project Hope), 32(5), 864–872. https://doi.org/10.1377/hlthaff.2012.1262

Chew

L. D.

Bradley

K. A.

Boyko

E. J.

(2004). Brief questions to identify patients with inadequate health literacy. Family Medicine, 36(8), 588–594. http://www.ncbi.nlm.nih.gov/pubmed/15343421

Cost Reports . (n.d.). Retrieved August 1, 2020, from. https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/Cost-Reports

10.

Dubois

Santos-Eggimann

(2001). Evaluation of Patients’ Satisfaction with Hospital-At-Home Care. Evaluation & the Health Professions, 24(1), 84–98. https://doi.org/10.1177/01632780122034812

11.

Flynn

Morley

Bentley

(2020). Hospital Discharges to home health rebounding, but SNF volumes lag. Avalere. https://avalere.com/press-releases/hospital-discharges-to-home-health-rebound-while-snf-volumes-lag

12.

Galvin

J. E.

Roe

C. M.

Powlishta

K. K.

Coats

M. A.

Muich

S. J.

Grant

Miller

J. P.

Storandt

Morris

J. C.

(2005). The AD8: A brief informant interview to detect dementia. Neurology, 65(4), 559–564. https://doi.org/10.1212/01.wnl.0000172958.95282.2a

13.

Jenkinson

(2002). The picker patient experience questionnaire: Development and validation using data from in-patient surveys in five countries. International Journal for Quality in Health Care, 14(5), 353–358. https://doi.org/10.1093/intqhc/14.5.353

14.

Kroenke

Spitzer

R. L.

Williams

J. B. W.

(2003). The patient health questionnaire-2: Validity of a two-item depression screener. Medical Care, 41(11), 1284–1292. https://doi.org/10.1097/01.MLR.0000093487.78664.3C

15.

Lawton

M. P.

Brody

E. M.

(1969). Assessment of older people: self-maintaining and instrumental activities of daily living. The Gerontologist, 9(3), 179–186. http://www.ncbi.nlm.nih.gov/pubmed/5349366

16.

Lee

Shi

S. M.

Kim

D. H.

(2019). Home time as a patient-centered outcome in administrative claims data. Journal of the American Geriatrics Society, 67(2), 347–351. https://doi.org/https://doi.org/10.1111/jgs.15705

17.

Leff

Burton

Mader

Naughton

Burl

Clark

Greenough

W. B.

Guido

Steinwachs

Burton

J. R.

(2006). Satisfaction with hospital at home care. Journal of the American Geriatrics Society, 54(9), 1355-1363. https://doi.org/10.1111/j.1532-5415.2006.00855.x

18.

Levine

D. M.

Ouchi

Blanchfield

Diamond

Licurse

C. T.

Schnipper

J. L.

(2018). Hospital-level care at home for acutely ill adults: A pilot randomized controlled trial. Journal of General Internal Medicine, 33(5), 1–8. https://doi.org/10.1007/s11606-018-4307-z

19.

Levine

D. M.

Ouchi

Blanchfield

Saenz

Burke

Paz

Diamond

C. T.

Schnipper

J. L.

(2020). Hospital-level care at home for acutely ill adults a randomized controlled trial. Annals of Internal Medicine, 172(2), 77–85. https://doi.org/10.7326/M19-0600

20.

Levine

D. M.

Paz

Burke

Schnipper

J. L.

(2021). Predictors and reasons why patients decline to participate in home hospital: A mixed methods analysis of a randomized controlled trial. Journal of General Internal Medicine. Advance online publication. https://doi.org/10.1007/s11606-021-06833-2

21.

Levine

Pian

Mahendrakumar

Patel

Saenz

Schnipper

(2021). Hospital-Level Care at Home for Acutely Ill Adults: A Qualitative Evaluation of a Randomized Controlled Trial. Journal of General Internal Medicine, 36(7), 1965–1973. https://doi.org/10.1007/s11606-020-06416-7

22.

Levinson

(2014). Adverse events IN skilled nursing facilities: National incidence among medicare beneficiaries. OIG. https://oig.hhs.gov/oei/reports/oei-06-11-00370.pdf

23.

Mechanic

(2014). Post-acute care — the next frontier for controlling medicare spending. New England Journal of Medicine, 370(8), 692–694. https://doi.org/10.1056/NEJMp1315607

24.

Mor

Intrator

Feng

Grabowski

D. C.

(2010). The revolving door of rehospitalization from skilled nursing facilities. Health Affairs, 29(1), 57–64. https://doi.org/10.1377/hlthaff.2009.0629

25.

Newhouse

J. P.

Garber

A. M.

(2013). Geographic variation in Medicare services. The New England Journal of Medicine, 368(16), 1465–1468. https://doi.org/10.1056/NEJMp1302981

26.

Prvu Bettger

Green

C. L.

Holmes

D. N.

Chokshi

Mather

R. C.

Hoch

B. T.

de Leon

A. J.

Aluisio

Seyler

T. M.

Del Gaizo

D. J.

Chiavetta

Webb

Miller

Smith

J. M.

Peterson

E. D.

(2020). Effects of virtual exercise rehabilitation in-home therapy compared with traditional care after total knee arthroplasty: veritas, a randomized controlled trial. JBJS, 102(2), 101–109. https://doi.org/10.2106/JBJS.19.00695

27.

Raîche

Hébert

Dubois

M.-F.

(2008). PRISMA-7: A case-finding tool to identify older adults with moderate to severe disabilities. Archives of Gerontology and Geriatrics, 47(1), 9–18. https://doi.org/10.1016/j.archger.2007.06.004

28.

Report to the Congress . (2019). Medicare Payment Policy. https://www.medpac.gov/document/march-2019-report-to-the-congress-medicare-payment-policy/

29.

Ryskina

K. L.

Yuan

Teng

Burke

(2019). Assessing first visits by physicians to medicare patients discharged to skilled nursing facilities. Health Affairs, 38(4), 528–536. https://doi.org/10.1377/hlthaff.2018.05458

30.

Simning

Orth

Temkin-Greener

(2020). Patients discharged from higher-quality skilled nursing facilities spend more days at home. Health Services Research. https://doi.org/https://doi.org/10.1111/1475-6773.13543

31.

Szende

Williams

(Eds.), (2004). Measuring Self-Reported Population Health: An International Perspective based on EQ-5D. EuroQol Group. https://euroqol.org/wp-content/uploads/2016/10/Measuring_Self-Reported_Population_Health_-_An_International_Perspective_based_on_EQ-5D.pdf

32.

van Hout

Janssen

M. F.

Feng

Y.-S.

Kohlmann

Busschbach

Golicki

Lloyd

Scalone

Kind

Pickard

A. S.

(2012). Interim scoring for the EQ-5D-5L: Mapping the EQ-5D-5L to EQ-5D-3L value sets. Value in Health: The Journal of the International Society for Pharmacoeconomics and Outcomes Research, 15(5), 708–715. https://doi.org/10.1016/j.jval.2012.02.008

33.

Wilson

Wynn

Parker

(2002). Patient and carer satisfaction with “hospital at home”: Quantitative and qualitative results from a randomised controlled trial. The British Journal of General Practice: The Journal of the Royal College of General Practitioners, 52(474), 9–13. http://www.ncbi.nlm.nih.gov/pubmed/11791829

34.

Wunderlich

Kohler

(Eds.), (2001). Improving the Quality of Long-Term Care. National Academies Press. https://www.ncbi.nlm.nih.gov/books/NBK224500/

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.01 MB