Evaluation of a novel sub-acute Hospital in the Home model for providing inpatient geriatric and rehabilitation services

Geriatric Evaluation and Management Rehabilitation Hospital in the Home model

This GEMRHITH expanded traditional acute HITH by adding a service lead consultant geriatrician and medical registrar. The medical model of care consisted of GEMRHITH admission assessment, attendance at twice daily handover meetings, bi-weekly case conferences and frequent telephone and face-to-face patient consults (in-home, clinic or virtual) on an as-required basis. All staff disciplines (medical, nursing and allied health) participated in all methods of patient consults. Allied health staff consisted of those from the disciplines of physiotherapy, occupational therapy, speech pathology, nutrition and dietetics, pharmacy, social work, and psychology, along with an allied health assistant. Medical and allied health staff worked weekdays, while nursing staff provided a 7 day service, shared with an acute HITH. Each patient received a minimum of a daily home visit from one discipline of practitioner. Medical and nursing also provided a nurse-led 24 h remote on-call service as part of the standard HITH after-hours call service, and no GEMRHITH-specific data were logged for reporting. Personal care workers were available if required. Referrals were accepted from inpatient medical and surgical wards, ED and geriatrician outpatient clinics. Box 1 details patient inclusion and exclusion criteria. Patients requiring medical investigations or use of medical equipment unable to be transported (i.e. X-ray, computed tomography) were conveyed to the hospital-based HITH clinic for medical officer review. Transport was arranged if required. Discharge was determined by medical status, functional ability, and availability and capacity of service providers and outpatient/community rehabilitation services to continue patients’ rehabilitation journey as required. Fig. 1 shows the GEMRHITH patient journey.

Fig. 1.

GEMRHITH: a patient’s journey (adapted from figure by Dr Abrams, Mid-West Toronto Sub-Region Hospital at Home Program). GEDI, Geriatric Emergency Department Initiative; NGO, non-government organisation; TCP, Transitional Care Program.

Data collection

All patients discharged from the GEMRHITH service from 1 October 2019 to 30 September 2021 were included. Demographic and clinical data were extracted from the electronic medical record including age, sex, country of birth, Aboriginal and Torres Strait Islander status and cultural and linguistically diverse (CALD) status, referral source, diagnosis-related group (DRG), admission and discharge dates, discharge destination and Sub- and Non-Acute Patient (SNAP) activity class. Clinical outcome measures were also obtained including the Clinical Frailty Scale²⁶ (CFS), Malnutrition Screening Tool²⁷ (MST) and change from admission to discharge measured using the Functional Independence Measure²⁸ (FIM) (cognitive, motor and total scores) and Modified Rankin Scale²⁹ (MRS). These measures were collected by nursing and/or allied health staff through direct observation or, when missed, through medical record review. Service utilisation data were collected from centralised hospital datasets with additional data sourced from the internal database including patient average LoS, number of discharges and bed occupancy. Specific occasions of service data were not available on any databases; therefore a sample was selected for chart review to determine occasions of service. Eighty patients (13% of the sample) were drawn at random from the GEMRHITH database. One of the authors who was involved in care delivery on the service extracted occasions of service data for each patient’s GEMRHITH stay by date and health profession.

Data analysis

Descriptive statistics are used to describe the patient population. Student's t-test was used for comparative data including pre/post analyses of functional measures. Statistical significance was defined a priori as P < 0.05.

Ethics

Approval was obtained from the Metro South Health Human Research Ethics Committee including a waiver of consent to retrospectively access GEMRHITH service data previously collected (approval no.: HREC/2021/QMS/69168).

Reach

During the first 2 years of GEMRHITH operation, 617 patients were admitted. A total of 60% of patients were female (n = 373) with a median age of 82 years (range, 32–102 years). Languages other than English were spoken by 110 patients (17.9%). The three most predominant non-English languages were Mandarin (n = 14, 2.2%), Vietnamese (n = 14, 2.2%) and Cantonese (n = 10, 1.6%). Six patients (0.9%) identified as Aboriginal and Torres Strait Islander peoples.

Most patients (n = 534, 86.5%) were referred to GEMRHITH from inpatient wards, with 82 (13.5%) admitted directly from the ED. Table 2 lists the referring sources, with 79.6% (n = 491) referred from a medical specialty (mainly neurology, n = 178) and 20.4% (n = 126) from a surgical speciality (mainly orthopaedics, n = 95). The DRG and the SNAP code and class provide additional detail regarding reason for admission and patient care needs (Table 3). The most common DRGs were nervous system, musculoskeletal and cardiovascular. The sub-acute class allocated to each patient on admission to GEMRHITH included rehabilitation in 61.1% and GEM in 38.1% of patients. The most common rehabilitation SNAP codes were debility, orthopaedic fractures and stroke (Table 3).

The CFS score on admission was a median of 6, indicating moderate frailty. Breakdown by CFS band indicated 4.1% with CFS 1–3 (very fit, managing well); 37.3% with CFS 4–5 (vulnerable, mildly frail); 39.5% with CFS 6 (moderate frailty), 18.8% with CFS 7–8 (severely to very severely frail), and 0.2% scored 9 (terminally ill). MST score on admission was 2 or more for 39.4%, indicating a risk of malnutrition (average 1, median 0, range 0–5).

Efficacy/effectiveness

Table 4 lists LoS, re-admission and discharge destination by SNAP class. Mean GEMRHITH LoS was 6.1 days (range, less than 1–33 days), with the most common being 0–5 days (47.2%). Discharge destination was predominantly their own home (n = 598, 97%) either alone (n = 167) or with support from family/carer (n = 431), with 12 (1.9%) being discharged to a family member's home, and four (0.6%) transferred to an aged care facility. One person died during their GEMRHITH admission, however, this death was attributed to underlying ischaemic heart disease and metastatic lung cancer, and medical assessment determined that it was not an adverse event attributable to the service model.

According to the safety data, 3.6% of patients (n = 22) were transferred back into the ED with the same DRG during their GEMRHITH admission. The re-admission rate to hospital within 7 days of discharge, with the same DRG, was 1.6% (n = 10), and that within 7–28 days of discharge was 2.3% (n = 14). According to standard practice, every month the service Nurse Unit Manager or medical officer reviews all re-admissions to determine whether it was a similar or related DRG or a new admission for an unrelated condition (DRG). A further 45 patients (7.3%) presented to the ED while admitted to GEMRHITH, and 51 patients (8.3%) were re-admitted within 28 days of discharge for conditions not related to their admission DRG.

Eight hospital-acquired complications were reported in seven patients: one pressure injury, five infections (four urinary tract and one surgical site), one hypoglycaemic episode, and one case of delirium.

Table 5 lists the mean admission and discharge differences in clinical/functional outcomes. There were statistically significant improvements for the total and sub-scale FIM scores. The average MRS showed a non-significant trend towards improvement (P = 0.06).

Adoption

Fig. 2 shows the monthly occupancy, unused capacity and separations during GEMRHITH’s first 2 years of operation. Occupancy was measured using the average daily occupied bed days (OBD) in each month, and unused capacity is the gap between OBD and service capacity (eight virtual beds). Average bed occupancy for the period was 5.3 virtual beds, below the 85% target (6.8 beds). GEMRHITH averaged 26 discharges per month (range, 16–36) over the 2 years.

Fig. 2.

GEMRHITH OBD (occupancy) and separations by month. LHS, left-hand side; RHS, right-hand side.

A total of 1135 occasions of service (mean, 14 per patient) were documented from the chart review (Table 6). A total of 459 home visits occurred, with two or more health practitioners visiting the patient together on 115 occasions, generating separate occasions of service for each discipline. In the sample audited by the chart review, the GEMRHITH after hours service was called eight times.

Table 6.Occasions of service by practitioner discipline for a random sample of 80 GEMRHITH patients during their admission.

Service type (± location)		Geriatrician	Medical registrar	Nurse	Pharmacy	Physiotherapy	Occupational therapy	Dietician	Speech pathology	Social work	Allied health assistant	Psychology	Total (%)
		0.5 FTE	0.7 FTE	2.8 FTE	0.6 FTE	1.0 FTE	1.0 FTE	0.4 FTE	0.4 FTE	0.5 FTE	0.6 FTE	0.2 FTE
Chart review		10	11	9	34	3	4	12	8	15	0	0	106	(9)
HITH clinic		12	9	14	1	0	0	0	0	0	0	0	36	(3)
Out-of-hours phone call		3	0	5	0	0	0	0	0	0	0	0	8	(1)
Initial assessment	Ward	73	37	74	38	0	0	0	0	1	0	0	223	(20)
	Phone call	0	0	0	17	0	0	20	4	18	0	0	59	(5)
	Home visit	0	0	5	0	73	72	9	10	7	0	6	182	(16)
	Assessment sub-total	73	37	79	55	73	72	29	14	26	0	6		464
Review	Ward	17	14	3	2	0	1	0	0	0	0	1	38	(3)
	Phone call	23	5	13	18	8	5	8	4	14	0	0	98	(9)
	Home visit	40	1	190	1	45	31	1	6	3	67	0	385	(34)
	Review sub-total	80	20	206	21	53	37	9	10	17	67	1		521
Total (%)		178	77	313	111	129	113	50	32	58	67	7	1135
		(16)	(7)	(28)	(10)	(11)	(10)	(4)	(3)	(5)	(6)	(1)

FTE, full-time equivalent.

Reach

In the first 2 years of operation, GEMRHITH admitted 617 patients to be supported in their home environment, with 82 (13.3%) admitted directly to GEMRHITH from the ED, thus avoiding the need to occupy a physical hospital bed. It is unlikely that the GEMRHITH patient cohort would have returned to their home environment in the same timeframe under a traditional acute HITH model without the additional multidisciplinary team support.

This GEMRHITH service model enabled a wider group of patients with a variety of conditions and broad age range to access otherwise inaccessible acute HITH services offered by the hospital. The mean age of patients admitted to this GEMRHITH service was older (82 years) than in most other HITH studies. Page et al.⁷ reported a median age range of 51–55 years, and Montalto et al.¹ reported that 53.2% were in the 50–79 year age bracket and 54.4% were men. Singh et al. reported a mean age of 83 years but included only patients over 65 years,¹⁸ unlike the current GEMRHITH, which had a minimum age of 18 years.

Patients admitted to GEMRHITH were varied, but were predominantly admitted for GEM, deconditioning and post-orthopaedic rehabilitation (Table 4). Having the availability of various allied health professions working under the guidance of a geriatrician in the service enabled more comprehensive care for such populations, with promising outcomes. Patients admitted to GEMRHITH benefited from accessing multidisciplinary team services to support their holistic care needs, enabling multiple factors to be addressed in their own home, which is not usually observed in acute HITH services.^21,22 Medical oversight of patients was maintained at twice daily patient handover sessions.

The average CFS score on admission was 6 (moderately frail), indicating that a large portion of patients admitted into the service required a level of support and care for basic personal activities of daily living (such as hygiene) and required assistance with activities of daily living such as shopping, cleaning and medication management.³⁰ One-fifth of patients were admitted with a CFS showing severely to very severely frail, indicating that patients with a high degree of care needs were able to be managed in the service. Ready access to geriatric and specialist geriatric and rehabilitation allied health input and personal care workers enabled the provision of the additional support required for this more physically and cognitively vulnerable patient group to facilitate earlier and safer discharges, potentially avoiding longer inpatient admissions or transfers to an aged care facility. There were a high proportion (17.9%) of patients who required an interpreter, which is encouraging given that people from CALD backgrounds may be less likely to participate in HITH.⁷

Efficacy/effectiveness and implementation

The results showed a high degree of adoption, with positive patient clinical and safety outcomes. The occupancy rate was high in comparison with other HITH service benchmarks reported in the literature and in Queensland Health.³¹

It was not possible to make direct comparisons to other studies regarding re-admission rates due to variability in reporting and the definition of re-admission in the literature, however, it appears that GEMRHITH had comparable or lower re-admission rates to those reported. Lim et al.³² reported an 11.9% re-admission rate for a large acute HITH service in Victoria while patients were still on the HITH service before discharge. Sriskandarajah et al.³³ audited a South Australian HITH program and showed that 15% of patients required hospital re-admission. A large-scale Australian study for acute HITH services had a re-admission rate of 2.3% within 28 days, which was lower compared with in-hospital care, and similar to GEMRHITH (1.6% at 7 days and 2.3% for re-admissions at 7–28 days). Levine et al.³⁴ hypothesised that re-admission rates were potentially lower for HITH models for older populations due to improved discharge planning. This was also likely to be the case with GEMRHITH, given that staff caring for the patient were able to see the patient in hospital, provide daily interventions, liaise with caregivers, organise for required equipment or services to be in place prior to transfer and, if required, complete a same-day home visit to the patient immediately after GEMRHITH transfer, facilitating smoother transitions to home.

Given the short LoS (mean, 6 days), large clinical improvements were not expected, however, a statistically, although not clinically, significant improvement in FIM scores was noted, as well as a positive trend towards improvement in the MRS. Improvements were largely noted in FIM motor scores, which is indicative of the cohort of rehabilitation patients from medical, orthopaedic and neurological backgrounds receiving allied health input for rehabilitation. This shows that patients can improve at home in a supported rehabilitation HITH model, while also reducing the demand on physical hospital beds. It was not anticipated that cognitive FIM scores would increase dramatically due to the number of GEM patients admitted to the service. The mild increase in cognitive FIM scores may indicate a small cognitive improvement from being in the home environment.³⁵

Limitations

A limitation to this study was that data were obtained from only one metropolitan Australian hospital and hence the results may not be directly generalisable. There was also an increased risk of bias because this was a non-randomised cohort study. This was a pragmatic report of patient characteristics and outcomes over the first 2 years of GEMRHITH’s operation and hence does not include a control or comparator group, nor does it include measures that would have unduly impacted the day-to-day operations of the hospital staff. Further work could be performed to compare the outcomes of GEMRHITH patients with a control group to determine superiority or non-inferiority of this care model, including for markers such as LoS, a main driver of HITH services.

Patient outcomes after discharge from GEMRHITH are also not known because this was not investigated; this could also be investigated in the future to demonstrate any benefit or otherwise of GEMRHITH compared with being in a physical hospital bed for the same care.

Robust data detailing the number, type and profession spread of GEMRHITH home and other occasions of service were not available. Given that the service was new and was run across two organisational directorates (hospital and community) there was no centralised system with consistent business rules for collecting accurate occasion of service data. Future research could develop a robust and minimally intrusive mechanism to log staff occasions of service and time.