The impact of a hybrid hospital at home program in reducing subacute rehabilitation referrals

Introduction

Healthcare digital transformation and telemedicine have noticeably increased during the past few years, especially with the COVID-19 pandemic^1–4. Hospital at Home services, inpatient level of care in the home, began at Johns Hopkins University in the mid-1990s as a way to avoid the decline that many seniors experience as a component of care within hospital walls^5,6. Throughout the years, several health systems adopted the model, but a broad payment structure did not exist making it financially challenging for more systems across the United States to fully implement similar programs⁷. In November 2020, the Centers for Medicare & Medicaid Services (CMS) launched the Acute Hospital Care at Home program to provide hospitals expanded flexibility to care for patients in their homes⁸. This unique waiver program was an expansion of the CMS Hospitals Without Walls legislation, as an effort to provide health systems with avenues to assist in treating the impending COVID-19 surges⁹. The result of this program was the first example of payment for this level of care at home in the United States¹⁰.

While the adoption of Hospital at Home programs across the United States had been slow prior to the CMS waiver program, the concept is not new internationally. For example, in Victoria, Australia upwards of six percent of all hospital bed-days are provided in the home⁷. In general, Hospital at Home programs have shown significant improvements in the quality of care including reduction in total length of stay, reduction in 30 day readmissions, reduced hospital infections and hospital induced delirium^11–13(but see ^14,15), as well as a lower total cost of care^13,16–19(but see ²⁰) and greatly improved patient satisfaction^11,15,16.

Healthcare systems interested in starting virtual hospital programs are tasked with developing and deploying a model of care to meet patient needs, often involving frequent telehealth calls, at-home monitoring tools, and some in-home clinical care. Atrium Health, in the southeastern United States, described how they engaged multiple departments (hospital medicine, primary care, nursing, care management, research, administration, etc.) to develop a workflow that mimicked traditional patient care²¹. Patients in the program who qualified for traditional inpatient care received daily telehealth calls with a physician and twice daily telehealth calls from a nurse. Atrium Health also expanded community care to meet the higher demand for clinicians entering homes and delivered in-home monitoring kits to Hospital at Home patients.

Virtua Health applied for and was approved for the CMS waiver in September 2021 and launched its Hospital at Home program in January 2022 in response to a significant surge in hospital patients following the spread of the Omicron COVID-19 variant. Virtua’s Hospital at Home program began in one of its five hospitals and focused on transferring patients diagnosed with COVID-19 from a hospital floor to complete their care in a home bed. Today, Virtua’s Hospital at Home Program is available at all five hospitals with patients being transferred into the program from a hospital floor or directly from the emergency room after meeting admission criteria. The program includes more than 60 DRGs across several main DRG category codes including community acquired pneumonia, congestive obstructive pulmonary disease, congestive heart failure, COVID-19, cellulitis, and urinary tract infections. In 2022, Virtua’s Hospital at Home program treated more than 270 patients, with nearly 1,000 total inpatient days completed in a home bed.

Patients in Virtua’s Hospital at Home program receive support in the home from a well-integrated care team including physicians, nurses, and paramedics. Every Virtua Health Hospital at Home patient receives a remote patient monitoring kit which includes a tablet for virtual care visits via video as well as an armband continuously measuring heart rate, respiratory rate, body temperature, pulse oximetry, and step count. Patients also receive integrated peripheral devices depending on the diagnosis, such as a spirometer and weight scale. Patients are monitored and supported by Virtua’s 24/7 command center staffed by registered nurses. In addition, every patient receives two in-person nursing visits per day, a virtual physician visit from the attending hospitalists as well as six additional virtual rounding sessions with the command center registered nurses. Virtua’s program can support patients at home with virtual specialist consultations, oxygen concentrators, physical and occupational therapy, daily blood work and laboratory testing, mobile imaging including x-ray and ultrasound, oral and IV medications and fluids, as well as easy to heat nutritious meals.

The objective of this study was to validate the home as an appropriate care option for patients within certain DRGs and acuity levels. We wondered if care in the home would lead to overall improved health outcomes, possibly due to increased mobility and comfort in the home. We observed a significant difference in subacute rehabilitation (SAR) referrals between in-hospital and at-home patients, and suggest that this result supports future investigations into this alternative care model.

Material and methods

Data collection

Patient data was recorded in an electronic medical record (EMR) using Epic (Epic, Epic Systems Corporation, Verona, Wisconsin, USA). To collect EMRs, we first queried Epic for patients in the Hospital at Home program using SQL. We then queried the database for EMRs for patients who did not participate in the program and with a DRG matching one of the DRGs in the Hospital at Home group. The following variables were extracted from the EMRs and saved in Excel files for use in further analysis: clinical note contents, discharge disposition, diagnosis related group, date of service, age (years) at admission, sex, and insurance payor. Information from clinical notes was extracted using natural language processing (NLP) and combined with tabular (structured) data for further analysis (Figure 1).

• Download figure
• Open in new tab

Figure 1.

Flow diagram depicting the combination of clinical notes (unstructured data) and tabular (structured) data from EMRs used for analysis in this study. Clinical notes for each patient were collected from the EMR. The free text in the notes was processed (uniform capitalization, special characters removed) before each note was searched for keywords (see Methods). In the final stage of language processing, a clinical BERT model was used to extract the sentiment in the note eg, identify if the note asserts SAR recommendation or negates SAR recommendation. This information was converted to structured data and combined with other structured data from the patient for further analysis and interpretation.

Results

We aimed to determine if outcomes for patients in Virtua’s Hospital at Home program were different from patients who did not participate in the program and remained in the hospital. Towards this goal, we extracted EMRs from patients who participated in the Hospital at Home program in 2022 (N = 272; 114 male, 158 female; age (years), m: 77.82, std: 9.95). We found that there were 43 observed unique diagnoses within this group, based on the MS_DRG codes in the records. In order to compare the Hospital at Home group to a comparable population, we limited the control group to inpatients at a Virtua hospital in 2022 who had a diagnosis (MS_DRG) that matched one of the unique diagnoses in the Hospital at Home group (N = 13879; 6366 male, 7513 female; age (years), m: 68.16, std: 20.29).

We used recommendation for SAR and discharge to SAR as metrics for outcomes following discharge from inpatient care. These metrics were chosen because they can be tracked within the Virtua Health system (for example, it would be more difficult to track re-admissions if a patient was re-admitted to a hospital outside the Virtua Health system). For each clinical note and for each patient, we used text mining and Clinical Assertion/Negation Classification BERT model²² (see Methods). After labeling each note as a recommendation to SAR (or not), we summed the number of patients for which there was at least one mention of a SAR recommendation. For the Hospital at Home group, only 40 of the 272 patients (14.7%) were recommended for SAR (Table 1). In contrast, 5012 of 13879 patients (36.1%) in the control group were recommended for SAR. This difference in the proportion of patients recommended for SAR across the two groups was significantly different (Figure 2; p < 0.001, Z = -7.30, two-tailed two proportion Z-test), with relatively far fewer Hospital at Home patients recommended for SAR.

• Download figure
• Open in new tab

Figure 2.

Proportion of patients in the Hospital at Home program that were recommended for SAR (0.147; N = 40 of 272) was significantly different from that in the control group (0.361; N = 5012 of 13879) (p < 0.001, Z = -7.30, two-tailed two proportion Z-test).

We were interested in knowing whether there was a significant difference in the final disposition of patients who had a recommendation for SAR between the Hospital at Home and control groups. To address this question, we searched the Discharge Disposition note for indication that the patient was transferred to SAR. Of the 40 patients in the Hospital at Home group who had a recommendation for SAR in their EMR, only 1 (proportion = 0.025) was eventually discharged to SAR (Table 2). In contrast, 2499 of the 5012 patients (proportion = 0.499) recommended for SAR were eventually discharged to SAR (Figure 3). Because only 1 patient in the Hospital at Home group was discharged to SAR, we did not perform a statistical test (N too low). However, the striking difference in the proportions of patients transferred to SAR between the groups suggests that Hospital at Home patients fared better than their in-hospital counterparts.

• Download figure
• Open in new tab

Figure 3.

The proportion of patients in the Hospital at Home program that were both recommended and discharged for SAR (0.026; N = 1 of 40) was lower than that in the control group (0.499; N = 2499 of 5012). Because only one patient was discharged to SAR from the Hospital at Home group, this comparison did not meet the requirements for the two proportions Z-test.

Discussion

Virtua’s Hospital at Home program admits patients from a variety of DRGs, with the most common diagnoses being pneumonia, congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), asthma, cellulitis, urinary tract infection (UTI), and COVID. Currently, all patients admitted to the Hospital at Home program are Medicare patients. Prior to being transferred home, all patients have a baseline physical therapy evaluation. Recommendations following this evaluation range from no needs to home with assistive devices, home physical therapy, or transfer to a rehabilitation facility. Of the Hospital at Home patients included in this study, 14.7% had a baseline recommendation for SAR on discharge and only 1 patient was eventually discharged to SAR. In contrast, 36.1% of patients treated in the brick-and-mortar hospital setting had a baseline recommendation for SAR, and 49% of those patients with a recommendation went to SAR on discharge.

One possible reason for the lower rate of patients going to SAR in the Hospital at Home group compared to the brick-and-mortar inpatient group may be due to differences in mobility. We notice that patients in the Hospital at Home program benefit from increased and early mobility, as assessed from remote monitoring kits tracking step counts. While at home, patients can liberalize their ambulation. They are not restricted to the hospital setting, which in most cases means staying in a semi-private room. Being at home also correlates with decreased hospital-acquired delirium as the patients are familiar with the layout and nuances of their home. Furthermore, at home, patients can be with their loved ones and pets, and sleep in their own beds. In addition to these liberties, patients likely feel more secure knowing that we are monitoring them 24/7 with remote monitoring kits, providing us with valuable information.

We believe that patients in the Hospital at Home program experience early mobility, better sleep cycles and decreased delirium, all of which translate to less deconditioning and subsequently no need for SAR on discharge.

Currently we are limited to admitting only Medicare patients to our program. We have no reason to believe that these outcomes are specific to Medicare patients and predict that these outcomes would hold if expanded to a broader patient population.

Conclusion

We observed that patients in Virtua’s Hospital at Home program were recommended for SAR (14.7%) at less than half of the rate of inpatients in the traditional brick-and-mortar hospital (36.1%). This reduced need for SAR for the Hospital at Home patients is likely due to relatively increased mobility, better sleep, and decreased rate of delirium in the home compared to the brick-and-mortar hospital. The results from this study suggest that the Hospital at Home program is an attractive alternative to traditional inpatient care for patients who meet certain clinical criteria.

Disclosure

All authors work for a healthcare system.